Handling Qualities Implications for Crewed Spacecraft Operations

Handling Qualities Implications for

Crewed Spacecraft Operations

Randall E. Bailey E. Bruce Jackson J.J. (Trey) Arthur NASA Langley Research Center NASA Langley Research Center NASA Langley Research Center 24 W. Taylor St 24 W. Taylor St 24 W. Taylor St Hampton, VA Hampton, VA Hampton, VA 757-864-8682 757-864-4060 757-864-6609 randall.e.bailey@nasa.gov bruce.jackson@nasa.gov trey.arthur@nasa.gov Abstract —Handling qualities embody those qualities or pilot is able to perform a required task or operation [1].

characteristics of an aircraft that govern the ease and precision These same qualities are as critical, if not more so, in the with which a pilot is able to perform the tasks required in operation of spacecraft [2].

support of an aircraft role. These same qualities are as critical, if not more so, in the operation of spacecraft.

The requirement that manual control be provisioned is codified in NASA’s Human-Rating Requirements for Space A research, development, test, and evaluation process was put Systems [3]. Under Item 3.4.1, “The crewed space system into effect to identify, understand, and interpret the shall provide the capability for the crew to manually control engineering and human factors principles which govern the the flight path and attitude of their spacecraft...” This pilot-vehicle dynamic system as they pertain to space exploration missions and tasks. Toward this objective, piloted requirement was established as “a fundamental element of simulations were conducted at the NASA Langley Research crew survival. Manual control means that the crew can Center and Ames Research Center for earth-orbit proximity bypass the automated guidance of the vehicle to interface operations and docking and lunar landing.

directly with the flight control system to effect any flight path within the capability of the flight control system.” These works provide broad guidelines for the design of spacecraft to exhibit excellent handling characteristics. In Further, under Item 3.4.2, “the crewed spacecraft shall particular, this work demonstrates how handling qualities exhibit Level 1 handling qualities (Handling Qualities include much more than just stability and control Rating (HQR) 1, 2 and 3), as defined by the Cooper-Harper characteristics of a spacecraft or aircraft. Handling qualities Rating Scale, during manual control of the spacecraft's are affected by all aspects of the “pilot-vehicle dynamic flight path and attitude.” The primary rationale is that Level system,” including the motion, visual and aural cues of the vehicle response as the pilot performs the required operation 1 handling qualities will allow the crew to effectively or task. A holistic approach to spacecraft design, including the control the spacecraft when necessary for mission use of manual control, automatic control, and pilot completion or to prevent a catastrophic event.

intervention/supervision is described. The handling qualities implications of design decisions are demonstrated using these These requirements are sufficient motivation to properly pilot-in-the-loop evaluations of docking operations and lunar attend to the design of inherently good spacecraft handling landings.

qualities. However, the depth and breadth implicit in these requirements is not often appreciated.

T ABLE OF C ONTENTS Handling qualities are not defined only by the control system design characteristics of the vehicle; they also include the motion, visual and aural cues of the vehicle response. This influence spans across the design of the manual and automatic control modes, cockpit displays, control inceptors (e.g., tactile cueing), crew station design, crew resources management (including mission control) and mission operations.

A holistic approach to spacecraft design, including the use B IOGRAPHIES ......................................................19 of manual control, automatic control, and pilot intervention is outlined to ensure excellent handling qualities. Spacecraft handling qualities research is described which highlight the 1. I NTRODUCTION implications of handling qualities on spacecraft design and operations and capture pertinent lessons-learned for future Handling qualities embody those qualities or characteristics vehicles, including commercial crew vehicles.

of a vehicle that govern the ease and precision with which a U.S. Government work not protected by U.S. copyright platforms as well as all future NASA and other government 2. B ACKGROUND agency and commercial spacecraft.

Handling qualities are illustrated using the concept of the “pilot-vehicle dynamic system” (Figure 1); the elements of This work was spurred on, in part, by the legacy established which form a closed-loop system, driven by a piloting task by Gemini and Apollo where significant research was or mission objective. The pilot acts as the system controller.

conducted for the parametric investigation of what drives The pilot’s role is to serve as “the decision-maker of what is spacecraft handling qualities. The Apollo and Gemini work to be done, the comparator of what’s happening vs. what he investigated, for example, the handling qualities influence of wants to happen, and the supplier of corrective inputs to the such parameters as: aircraft controls to achieve what he desires” [4].

• Spacecraft attitude control mode, control power, target External Disturbances - Aural Stress lighting and target oscillatory motion [6, 7]; Turbulence Cues Wind / X-Wind • “Remote” docking using closed-circuit television [8]; Control • Visual aids, in day and night conditions, to align to a Aircraft Control Force Cockpit Response Task Command Augmented docking target [9]; Feel Pilot Aircraft System • Hand controllers, instruments, and control modes [10]; • Visual simulation compared to “full-size” docking [11]; Visual Motion Cues Cues • Visual aids and attitude control modes in lunar orbit [12]; • Handling qualities design requirements and control law types for lunar landing [13,14].

Figure 1: Pilot-Vehicle Dynamic System [4] For Apollo and Gemini, this work was critical since Handling qualities reflect the precision with which the pilot experience for the required spacecraft operations was non- can accomplish the given task (as the controller of the existent.

closed-loop system) and the associated pilot workload and compensation to meet this level of performance. Although One could argue that a wealth of data and experience are at the “augmented aircraft” – the vehicle’s dynamic response hand today so new work in spacecraft handling qualities is characteristics augmented by its flight control system – not needed. The counter argument is that there is not a would logically be the primary determinant of handling sufficient basis from which to extrapolate to new vehicles, qualities, the other system factors may be equally operations, and operating conditions and mission plans. As influential. These cues (such as the presence or absence of an example, when the Space Shuttle was being developed, motion cues and the environment, e.g., external visibility, proximity operations and docking issues were assumed to be control upsets, aural cues, and pilot stressors) are the a lower priority (i.e., had less technical risk) compared to feedbacks that the pilot uses to perceive the vehicle response other system development tasks because of the Apollo and and meet the demands of the task.

Gemini legacy [15]. This lack of initial concern for proximity operations and docking resulted in “complex Modifications or changes in individual elements within this operational work-arounds over the life of the [Shuttle] closed-loop system may be compensated by the adaptive program.” Stepping back and fully considering the past is pilot, but possibly at a cost of pilot workload or changes in the only way to avoid reliving history.

task performance. These effects cannot be segregated; thus, handling qualities must really be evaluated in the aggregate To avoid reliving this past, a research, development, test, [4,5]. Empirical and historical data can provide perspective and evaluation process was initiated to identify, understand, and estimates on the effects of changing elements within the and interpret the engineering and human factors principles system but the only truly accurate measure is to evaluate the which govern the pilot-vehicle dynamic system (i.e., aggregate closed-loop system.

handling qualities) as they pertain to space exploration missions and tasks for Constellation and all future NASA Spacecraft Handling Qualities and other government agency and commercial spacecraft.

Our foundational understanding of handling qualities (HQ) and the pilot-vehicle dynamic system is rooted in the 3. P ROXIMITY O PERATIONS AND D OCKING aeronautics domain. Although the physics of the systems are The Spacecraft Handling Qualities work initially obviously different, the fundamentals are no different and of investigated proximity operations and docking issues, no less importance for spacecraft.

principally focused around two human-in-the-loop The National Aeronautics and Space Administration experiments each, conducted at both the Ames Research (NASA) initiated a multi-centered Spacecraft Handling Center (ARC) and Langley Research Center (LaRC).

Qualities project in 2007. This project was designed to Subsequent to this work, additional activity was conducted provide data and guidelines for the design, development, at NASA Johnson Manned Space Flight Center, focused test, and evaluation of NASA’s Constellation Program specifically on Orion/Crew Exploration Vehicle handling Desired and Adequate Performance Standards qualities investigations [16].

Equally critical to the task definition are the required task performance standards. The performance standards must be: These works revealed numerous design implications and a) germane to the required operation or task as they apply to design guidance which capture the influence of handling the actual mission; b) include variables or outcomes qualities on a spacecraft design for proximity operations and controllable by the pilot; c) observable to the pilot; and, d) docking.

sufficiently demanding that high closed-loop pilot-vehicle “task bandwidth” is required to aptly stress and test the Piloting Task handling qualities characteristics [2].

A complete task definition is required, not just of the primary control task, but also of any critical auxiliary tasks.

Once the task is defined, ensuring that closed-loop control For proximity operations and docking, it was repeatedly by the pilot is required, the “task bandwidth” is modulated stressed by Shuttle commanders that the docking task is a by: 1) the precision that is demanded of the pilot-vehicle crew effort. Delegation of tasks across the crew was performance; and, 2) the time to complete the task.

required. In lieu of this - for our single pilot evaluations - automation was added to replicate some of these crew 1) Task Performance —The task precision is dictated by the functions, such as using a “range to docking” automated mission. For proximity operations and docking, desired and call-out. Task analysis should be conducted to appropriately adequate docking performance standards were established define the “complete” piloting task; otherwise, false or based on the anticipated docking hardware, as shown in misleading assessments may result. Table 1. The adequate docking standard was defined by the successful docking performance limits. Adequate The concept of operations for rendezvous, proximity performance means that the task can still be accomplished; operations and docking, including docking mechanisms, are albeit, without significant margin for error. On the other first-order influences on handling qualities and the resultant hand, desired performance means that the task can be on-orbit control design. This fact is self-evident because an accomplished with significant margin. Changing the evaluation pilot is rating the vehicle’s handling docking method, changes the task and hence, the handling characteristics for a given task or mission task element.

qualities issues and design challenges.

In conducting handling qualities evaluations, the criticality 2) Time Constraint —A time constraint may be inherent for of the designing appropriate tasks cannot be over- certain tasks. If it is not naturally part of a task, it must be emphasized. History has shown that latent handling qualities imposed; otherwise, this becomes an uncontrolled variable deficiencies are exposed by using piloting tasks that create in the task that can result in potentially adverse “stress cases” – those conditions that necessitate closed-loop consequences.

control of the vehicle to achieve the task objective. The task should not be unreasonable, but it is usually one which For proximity operations and docking, a time constraint was would not often occur. An example in the aviation domain is established by specifying an approach and docking closure the offset landing task. This task is within the capabilities of rate (e.g., 0.1 ft/sec). This closure rate was included as part the aircraft and is “legal” to perform, but operationally, of the desired and adequate performance standards. The pilots will usually perform a go-around rather than desired performance standard was met if the closure rate for correcting back to the centerline of the runway in an attempt the duration of the task was held within ±25% of the target to land. Using this task to evaluate handling qualities value. The adequate performance standard was within the exposes the pilot-vehicle dynamic system to a high expected performance limits of the docking mechanism.

bandwidth task for the purposes of exposing potentially Table 1: Docking Task Performance Standards latent deficiencies.

In proximity operations and docking, this same approach Desired Adequate was used. The initial condition on a +V-bar docking Performance Performance approach to the International Space Station (ISS) was offset Radial Offset ±1.5 in ±1.5 to ±3.2 in by 2.5 ft from the docking axis [17,18]. This offset required the evaluation pilot to laterally and vertically reacquire the Roll/Pitch/Yaw ±2.0 deg ±2.0 to ±3.0 deg Angle docking axis. This “stress case” was validated as a good HQ task by experienced Shuttle commanders as one that Axial Closure 0.075 to 0.125 ft/s 0 to 0.075 ft/s OR stressed closed-loop control without completely diverging Rate (0.1 ft/s) 0.125 to 0.15 ft/s from the concept of operations. Operationally, this task Radial (Linear) ±0.0325 ft/s ±0.0325 to mirrors the aviation-domain offset landing task where the Rate ±0.1125 ft/s offset is at the edge of the allowable ISS docking corridor that would likely, but not necessarily, require the vehicle to Roll/Pitch/Yaw ±0.05 deg/sec ±0.05 to stop the approach and reposition before attempting docking. (Angular) Rate ±0.15 deg/sec The required HQs are thus established by the precision dispelled any movement toward eliminating this capability.

within which the task must be completed for mission Their experience showed that, while not often used, manual success. These standards were based on data relative to the rotational control enabled many orbital Shuttle missions.

anticipated docking mechanisms for Constellation and those Three Degree-of-Freedom (3 DOF) Control HQs anticipated for the commercial off-the-shelf platforms [19].

If the docking mechanism or operational concepts are The introduction of rate command/attitude hold (RCAH) changed, the HQ standards and design challenges will also control laws to maintain spacecraft attitude (allowing the change.

pilot to perform translational inputs only) can significantly, but not completely obviate rotational coupling effects.

Handling Qualities Influences Coupling effects may still be apparent as the vehicle slices One predominate handling qualities factor in proximity through the selected attitude hold deadbands [18,19].

operations and docking is the rotational coupling of the As one pilot observed, the “vehicle is … going to hold vehicle due to translational inputs [17,18]. This effect is attitude plus or minus the deadbands, but I don’t feel like illustrated in Figure 2 using a simplistic example where the I’m really in control.” Pilot compensation was either to: a) relationship between reaction control system (RCS) jet ignore the deadbands (there were cockpit indications of longitudinal location varies with respect to the vehicle imminent deadband-limit RCS firing); b) offset the docking center-of-mass (CM). If the RCS jet is at the same x-axis alignment in anticipation of a deadband firing; or, c) correct (body) position, translational RCS jet firings pass through for them immediately when they occur. In either case, a the CM and do not introduce spacecraft rotation changes handling qualities penalty is incurred, resulting in a net (i.e., uncoupled or “neutral coupling”). The left- and right- improvement in HQRs for 3 DOF control, but not “night- hand sketches in Figure 2 place the RCS jets forward or aft and-day” differences [17].

of the CM x-axis (body) position so translational RCS jet pair firings introduce either “adverse” or “proverse” Feedforward Compensation spacecraft rotation changes, respectively. For a configuration with adverse coupling, a command to translate Pulse-width-modulation-type feedforward compensation can the vehicle right would cause an initial yaw to the left (as actuate RCS firing durations and combinations to reduce viewed through the centerline camera), seemingly counter to residual RCS translation-to-rotational coupling effects the translation command. Conversely, in a case with [20,21].

proverse coupling, a command to translate right would effect a yaw rotation in the right direction (as viewed Pilot-in-the-loop testing vividly demonstrated the benefits of through the centerline camera), but seemingly additive to these systems. Feedforward compensation provided Level 1 the translation command. This coupling effect may also HQRs by virtually eliminating translational-rotational occur in the lateral translation/yaw rotational axes as well.

coupling effects [20].

Six Degree-of-Freedom (6 DOF) Control HQs Handling qualities tests were performed with feedforward compensation in both nominal and off-nominal conditions Rotation coupling effects were investigated during Apollo such as with intentionally introduced inaccuracies or and Gemini and careful RCS design enabled successful impurities associated with the controller (e.g., by offsetting manual control of proximity operations and docking.

the CM from the design position, varying the thruster output from nominal value, etc.). In all cases tested, the These issues were reevaluated with representative current- feedforward compensator handling qualities were largely day RCS thruster sizes [17]. The HQ data show that desired unaffected [21].

performance can generally be obtained without coupling, but six degree-of-freedom (6 DOF) control is higher Further failure mode effects testing is needed, however. The workload and thus, not satisfactory without improvement concern is the handling qualities “cliff” that may be lurking.

(not Level 1 HQRs). Adequate HQ performance (Level 2 Particular emphasize should include the real-world effects HQRs) is generally possible within the range of coupling of thruster size and cycle times and how these issues effects evaluated. Large pulse widths (minimum RCS jet manifest themselves into residues from perfect firing times) and large thruster sizes, coupled with adverse compensation and latencies caused by duration times to rotational coupling can create Level 3 HQRs. Coupling was complete the pulse width commands. The hope is that the correlated (statistically significant) to fuel usage; lower control law compensation will be robust and not prone to coupling resulted in better HQRs and less fuel being any catastrophic degradation of handling qualities due to expended in the operation.

these effects or others.

For instance, on occasion, the desirability or need for manual rotational (6 DOF) control was questioned; however, the more senior Shuttle commanders quickly Figure 2: Illustration of Adverse, Neutral, and Proverse Translational-into-Rotational Coupling Fuel Consumption But by far, the preeminent display factor influencing these evaluations was the docking camera. Evaluations were not The rotational-translational coupling effects influence fuel flown without the docking camera or with boresight or consumption in two ways. First, fuel consumption is driven parallax issues, but its critical role in the task/mission was by the design of a control law compensator that might apparent. The importance of visual aids for docking minimize the coupling effects (e.g., RCAH vs. feed-forward alignment was established in Gemini/Apollo and design controller). Second, the rotational coupling (i.e., handling provisions were made accordingly [9]. The appropriate qualities) affects the number of translational inputs required provisions must be ensured in future spacecraft designs for the task [20,21]. Larger attitude deadbands for an RCAH either by redundancy in the event of a camera failure or controller uses less fuel in compensating for rotational- alternate positioning aids.

translational coupling due to translational inputs than smaller deadbands. Analogously, the feedforward Task Influences compensator uses more fuel than a typical RCAH controller.

Apollo and Gemini docking used much higher closure rates However, testing showed that during the task, a properly than the 0.1 ft/sec standard used by Shuttle (nearly an order designed feedforward controller used less overall fuel than of magnitude). The reasons are many (different docking an RCAH design because the pilots used fewer translational mechanisms and an almost order of magnitude mass inputs. The predictability of the system and fewer deadband difference between Gemini/Apollo and Shuttle), but they firings resulted in overall less fuel expenditure during the may not hold true in future operations.

task [20]. The net result is that better handling qualities corresponded to less propellant consumed.

An evaluation of docking closure rate was conducted using 0.5 ft/sec closure starting at 50 ft from the docking port with Display / Aiding Information an initial 2.5 ft offset from the approach corridor [17]. The Various display information methods have been attempted handling qualities effects are illustrated in Figure 3, using a to compensate for rotational-translational coupling effects “bubble” chart where each HQR is shown and the size of the [20, 21]. Each have varying degrees of success, none major HQR bubble is proportional to the number of ratings at that and few with consensus of opinion. The most successful of level.

these are the deadband indicators – explicitly showing the crew the attitude (and rate) at which a deadband in the With a RCAH holding the docking attitude, the higher RCAH controller will fire. Translational rate and closure rate task produced generally worse HQRs, but not positioning information with respect to the docking Local dramatically so. Adequate, if not desired docking Vertical Local Horizontal (LVLH) axis was also put to performance was the norm. However, with 6 DOF control excellent use.

and rotational due to translational coupling, ratings of 10 were generally given. At these higher closure rates, “you don’t have enough time to be messing around with all these kind of control inputs that you have to do.” Potentially • Missions were slated toward the scientifically catastrophic consequences could result. interesting lunar polar regions, unlike the equatorial landing sites of Apollo.

Rotating Targets • Missions were to be flown “any time.” Current low earth operations to the ISS provide stationary or • A lunar habitat and a sustainable infrastructure would quasi-stationary docking targets. Future docking be developed for an enduring scientific presence.

requirements may involve rotating targets, especially in case These mission objectives created several design constraints of attitude control failures resulting in an uncooperative relevant to the handling qualities, and ultimately, the target vehicle.

vehicle’s design: Evaluations of docking attempts with rotating targets 1) In the polar region, the very low sun angles create showed this was possible, but the pilot workload was not visually powerful shadow effects that cause the loss of tolerable [21].

natural vision depth cues and impair translational The pilots demonstrated amazing adaptability to the various velocity and landing zone awareness.

configuration effects and targets but an overwhelming need 2) Operations would be independent of the landing zone for reconfigurable digital autopilot control (DAP) was lighting conditions (unlike Apollo, where “optimal” voiced. The ability to tailor the DAP gains and modes, in lighting was designed by the timing of the launch).

terms of pulse size, direct vs. pulse, and segregated, 3) Habitat build-up constrains the landing approach selectable modes per axis (rotational and translational), were procedures and dictates much higher landing accuracy critical to success.

than Apollo.

These design constraints further define the operational task or mission task elements which the vehicle’s handling qualities and consequently, the spacecraft design: 9 a) Landing precision is dictated by the proximity to a Level 3 habitat build-up; no “overflight” is permitted.

b) Minimum fuel consumption is preeminent.

c) Natural vision may or may not be useful to the Level 2 pilot/crew.

3 1 Rating 2 Ratings d) Approaches to the landing site will need an out-of- 2 3 Ratings Cooper-Harper Pilot Rating plane maneuver (such as a “dogleg” approach) to 4 Ratings Level 1 minimize collateral damage to the habitat from the 10 ft IC 20 ft IC 50 ft IC descent module in the event of a landing abort Pulse RCAH Pulse RCAH Pulse RCAH (with subsequent staging of any descent hardware).

Initial Condition and Control Mode The lack of natural vision references for the pilot/crew is Figure 3: Effect of Tasking Differences on HQRs both a blessing and a curse. During Apollo, the constraints placed by the design of the Lunar Module (LM) window for crew visibility and landing trajectory were “a major 4. L UNAR L ANDING HQ I MPLICATIONS problem” [22]. Apollo used a “pitch-up” maneuver in the final approach descent and landing phase and a judiciously The holistic design (and impact) of handling qualities on a designed window configuration to enable the flight crew to spacecraft design is illustrated in the following, using the have approximately two minutes of viewing of the planned design and results of a piloted simulation of a lunar landing landing site. The pitch-up sacrificed fuel usage by rotating vehicle.

the vehicle to a near vertical orientation, so the crew could see and verify the integrity of the landing site (e.g., free of Lunar Lander Handling Qualities Design Implications rocks and debris, terrain slope within design allowances) As NASA looked to return to the lunar surface, the design through a window that provided almost 70 deg nose-low issues and lessons-learned of Apollo were closely studied as visibility. Optimal visibility conditions were provided by they are still relevant today for practical planetary and lunar operational procedures and constraints, including launch lander vehicle designs and missions [22-27]. From this times, dates, and landing locations.

baseline, additional challenges for the vehicle design were levied to improve the scientific return for manned lunar Automated landing of Apollo was possible, but all landings missions.

were manually flown with the crew taking over control between 550 and 240 ft above the touchdown zone elevation [26]. After Apollo 11, each crew performed at least 1, and as many as 18, landing zone re-designations during the method requires very large attitudes for rate or pitch-up phase [27]. position hold functions. Elaborate display and alerting functions would be needed for the crew to Technologies developed since Apollo offer potential ways appropriately monitor normal operations and provide to break some of the aforementioned Apollo operational and timely intervention and recovery in event of failures.

design constraints as well as meeting the new Exploration challenges. For instance, Synthetic and Enhanced Vision Control Power —“Satisfactory” and “Acceptable for (SV/EV) technologies, conceived and developed in the Emergency Operations Only” handling characteristics aviation domain, may eliminate the challenges imposed by a guidelines for rotational control effectiveness, using the natural vision constraint and thus, enable additional benefits Cooper rating scale [35] emerged from the Apollo data [36].

for safer, more efficient operations [28]. These technologies The criterion, for a RCAH control law, boils down to: 1) the are being used, in whole or in part, under the Autonomous maximum rotational acceleration authority; and, 2) the Landing and Hazard Avoidance Technology (ALHAT) [29]. chosen maximum rate command provided to the pilot.

SV/EV technologies offer the possibility that fuel-expensive The maximum rotational acceleration authority is Apollo-like pitch-up maneuvers would be unnecessary using approximately the RCS jet size, multiplied by the number of “equivalent vision” instead [30]. SV/EV technologies can RCS jets and their moment arm and divided by the provide equivalent visual information sufficient for the crew rotational inertia. Part of the HQ design trade-space can be to obviate the impact of landing zone obscurants and be illustrated using the Apollo LM experience. For Apollo, the independent of window sizes and placements, visual criteria were met when four RCS jets were used. Nominally, illusions, or lighting/depth perception conditions. however, only two jets were fired to conserve RCS propellant but for circumstances that dictate higher These challenges exist within the broader context of manual rotational control (e.g., excessive rate error in the RCAH control, handling qualities, and the engineering design trade- control law), four jets were available [13]. Apollo also space for a Lunar Landing vehicle.

placed the RCS jets on the ascent module. This placement increased the moment arm (increased rotational Lunar Lander Handling Qualities Design effectiveness). It also increased the LM ascent mass. The RCS jets were also used for on-orbit attitude control. This A piloted simulation study investigated the handling obviated the need to have two sets of RCS jets (one for qualities of a lunar landing vehicle and the influence of descent; one for ascent and on-orbit) but the LM RCS jets, critical design issues with operational employment sized for the landing mission, were found to be too large for consequences. The focus was on visual cueing and control on-orbit operations, causing proximity and operations HQ effector sizing during manual control of the approach and issues [37].

landing of a lunar landing vehicle, enabling a quantification of the trade-space for the vehicle design and operational The present-day concern is that the rotational control deployment. Automatic control modes of the vehicle were a effectiveness required to meet the Apollo guidelines are recognized design feature of future lunar lander vehicles; significantly larger than what current designs may the implications of which are discussed in Section 5.

practically provide. Validation of the guidelines is needed and possible alternative methods to satisfy handling Control Laws —RCAH control laws with a phase-plane qualities should be explored.

controller were evaluated. Other control law methods were rejected for evaluation: NASA Ames has recently conducted two Vertical Motion Simulator tests evaluating the validity of these rotational  Translational control using the RCS, while precision control effectiveness criteria from Apollo [31,38].

control is offered, will typically be underpowered and Evaluations were conducted using a 15 degree (deg) thus, not conducive to minimum fuel/minimum fuel glideslope trajectory to a pitch-up for vertical landing, objectives [31].

mirroring the initial Apollo lunar landings. The term  Thrust vectoring control (TVC) of the descent engine “glideslope” is used throughout as an analogy to the [32] offers excellent control power, but is critically aeronautics domain, where a constant approach angle to flawed due to potential failure modes. Apollo landing is flown. In this application, a constant angle evaluated this concept but rejected it, limiting TVC approach is only approximate; the trajectory curves from the rates to provide trim control to avoid catastrophic TVC initial approach angle to smoothly transition to a hover point actuator failure consequences [33].

above the landing zone, transitioning to a vertical descent to  More sophisticated control methods such as landing.

translational rate command-position hold functions were dismissed because, while excellent performance Testing was conducted with and without explicit pitch, roll, has been demonstrated in rotorcraft applications and and rate-of-descent guidance commands shown on head- elsewhere [34], they lack in simplicity of design. They down displays. Testing was planned to vary the evaluation also can suffer from the human-centered design task, flying with and without a 250 ft lateral offset/dog-leg principles since, with only one-sixth gravity, this approach from the initial trajectory to the vertical landing.

However, pre-test work showed that, without explicit angle was flown for the initial Apollo LM missions, but the guidance, the evaluation pilots could not reliably complete later so-called “J” missions used trajectories approaching 25 the task with a lateral offset. For the no-guidance test cases, deg glideslope. Future vehicles may find it advantageous to only a straight-in task was flown. use higher glideslope angles to landing [39]. Trajectory design is part of the trade-off in fuel consumption and For the lateral offset task with guidance, the Apollo design natural vision viewing of the landing site.

guidelines were roughly supported. But for a straight-in task, without an offset, the degradation of handling qualities Second, the previous work provided for and allowed the use with decreasing control authority was not strongly of direct translation control by the pilot using a translational supported. hand controller. Although the control authority was minimal, it was permitted and introduced a degree-of- In the second test, the characteristics of the Lander Vehicle freedom which differed from Apollo. In Apollo, only were changed, but the major elements of the test were rotational attitude control to effect translational positioning replicated.

was used because of the significant amount of rotational coupling introduced by translational control inputs. The A summary plot of the data against the guidelines is shown RCS jets were located on the Ascent Module for maximum in Figure 4. The plot uses median HQRs for the twelve rotational control authority. Evaluations were conducted evaluation pilots. The trends generally follow the first without allowing direct translational control for comparison.

simulation test results. At the highest levels of control power tested, median Level 1 pilot ratings were generally Finally, the critical influence of visual cues was considered.

given. The individual rating data, however, exhibits As the trajectory approach angle is changed, the pitch substantial scatter. Some Level 3 or nearly Level 3 ratings attitude will vary; thus, affecting the amount of time that the were given for those configurations with median Level 1 crew can use for natural vision viewing of the landing site.

ratings. The rating scatter warrants concerns since this The previous work showed that manual flight to an offset suggests the possibility of lurking cliff-like HQ landing zone was not possible. One might consider that this characteristics where Level 1 handling characteristics can result was because the out-the-window simulated visual quickly degrade due to the presence of system cues were insufficient to support the task (as discussed nonlinearities, latency, or handling characteristics later). On the other hand, the visual cues might have been inappropriately sensitive to pilot techniques [38].

sufficient but the control characteristics were deficient. In any event, the influence of natural vision and the potential Cheatham and Hacker Lunar Lander Handling Qualities of “equivalent vision” technologies warranted exploration.

Equivalent vision may represent a new tool in the manned 90 Level 1 - Median planetary exploration designers’ “tool-bag.” Level 2 - Median 80 Level 3 - Median These experimental objectives were evaluated in the context 70 Satisfactory of variations in the control authority of a lunar landing vehicle which uses a RCAH control law, analogous to the guidelines developed under Apollo.

Acceptable for Emergency Simulated Spacecraft 30 Ops Only Max. Rate Command (deg/sec) A simulation model (Figure 5) was built based on evolving lunar lander designs [32]. 10 0 2 4 6 8 10 12 14 16 18 20 The simulation model was built using object-oriented Time Constant (sec) programming techniques within the Langley Standard Realtime Simulation in C++ (LaSRS++) software Figure 4: HQR Level Data [38] vs. Guidelines [13] framework [40]. A generic-spacecraft simulation within this framework served as a test bed for modeling spacecraft Test Description dynamics, propulsion, control systems, guidance, and A pilot-in-the-loop simulation was conducted to explore the displays.

control and display interactions on the handling qualities of The vehicle model was composed of interconnected stages, a lunar lander vehicle.

simulated using a parent/child relationship. The ascent stage This work built upon the previous work [31,38] but was modeled as the parent with the descent stage as a child.

explored the control and display interactions in a lunar The descent stage contained separate mass models for the lander design. main engine fuel, RCS fuel, and the airlock. Force models were provided for the main engine, each of the sixteen RCS First, the previous work evaluated only shallow 15 deg engines, and the four landing gear models. For simplicity, glideslope trajectories to the landing zone. This trajectory the throttle-able, thrust output of the main engine force model was placed close to the lateral and horizontal axes of the stacked center-of-mass. Active gimbaling of the main engine was provided to trim the thrust vector through the CM. All forces and moments were applied to the stacked vehicle’s center of mass. Fuel slosh was not modeled.

The pilot, as defined by the design eye point reference (DERP), was located 28.1 ft above the extended landing gear position. The center of mass at the initial starting condition was 6.6 ft below the DERP. The four RCS jets were placed in a ring around the descent stage using an orthogonal quadrant of four individual thrusters for attitude control. The RCS thruster size was experimentally varied.

The RCS jets were located at a 14.8 ft radial from the centerline of the vehicle (i.e., the moment arm), approximately 7.7 ft below the DERP.

Figure 6: Descent Module (top view) RCS Locations – Quads 1-4 and Jets 1-4 on each Quad The RCS jet size was experimentally varied to create various angular acceleration control authorities.

An autothrottle was mechanized to hold a pilot-commanded vertical descent rate. Translational Hand Controller (THC) pilot inputs commanded discrete one ft/s increments to the commanded vertical descent rate. Proportional and integral control modulated the descent engine thrust. The engine throttle response rate was not limited.

Simulator The experiment was conducted in the fixed-base Lunar Flight Deck simulation facility at NASA Langley Research Center (LaRC). The Lunar Flight Deck used multiple o projectors and a dome screen to provide a 135 horizontal Figure 5: Simulated Lunar Landing Vehicle o (H) x 67.5 vertical (V) field-of-view. The screen was located 10.6 ft from the DERP. The dome and projection Control Laws o o system were biased to provide 22.5 up and 45.0 down The guidance and control laws were implemented in viewing angle from the DERP. The multiple image Simulink/Stateflow and autocoded for use in the real-time generators and projectors are warped and blended to provide simulation shell (LaSRS++).

uniform and seamless ~30 pixels per deg resolution across the field-of-view, referenced to the DERP.

RCAH control laws (in body-axis pitch, roll and yaw axes) included an RCS mixer which determined the jets to fire to A lunar polar landing site was preferred for the evaluations provide the commanded accelerations. The rate deadband but terrain data of reasonable resolution was not available.

for the phase plane controller was 2.0 deg/sec and the Instead, the Apollo 15 landing site was utilized because of attitude deadband was 0.5 deg.

the availability of higher resolution data and the interesting terrain features it provided. The out-the-window (OTW) Sixteen RCS jets were arranged in four quads, with four lunar database was created using lunar topology and orthogonal jets in each quad (Figure 6). A simplified imagery data collected by the Clementine project [41].

thruster mapping dedicated RCS thruster firings to each Medium resolution data was used in the general area of the moment command. Because of the simplified logic, when Apollo 15 landing site, approximately 500 meters per simultaneous pitch and roll inputs are commanded, the elevation post, draped with 125 meter per pixel imagery.

control authorities are effectively halved in each axis.

Higher resolution data was inset in the immediate vicinity of Various jet combination mappings were considered but the Apollo 15 landing site (15 meters per elevation post- rejected for simplicity of design.

spacing, draped with 4 meter per pixel imagery).

In the landing area, the resolution of the OTW database was good, but not to levels approaching the real-world. The OTW presentation was augmented with boulder and habitat The resulting parabolic shaping mimics that used in Apollo.

models to increase the visual density. The shaped roll and pitch inputs were multiplied by 12 deg/sec to create the pilot commanded roll and pitch RCAH The evaluation pilot (EP) conducted the task in a standing control law inputs.

position, analogous to the Apollo Lunar Module design. The crew station is shown in Figure 7. A window (20 inch H x The left-hand THC was provisioned for 3-axis translational 14 inch V) approximated the field-of-view anticipated for a control commands but only up/down vertical commands Lunar Lander vehicle. Two head-down (13.25 inch H x 10.5 were active. Up/down pilot inputs of the THC were inch V) displays, using 1280 x 1024 pixel resolution, converted into vertical descent rate command increments.

provided primary flight, navigation, and systems awareness Each discrete THC input would increment the value of the for the EP. commanded descent rate by one ft/sec.

Aural call-outs of the altitude above the terrain (extended gear height) were provided during the landing phase in 10 ft increments starting at 150 ft and continuing until touchdown.

Head-Down Displays The two-color head-down displays showed four primary display formats: a) Attitude Direction Indicator (ADI); b) Navigation Display (ND); c) Vertical Situation Display (VSD); and Auxiliary Display.

The ADI (Fig. 8) served as the primary flight display. The ADI showed lunar attitude reference using a roll/pitch/yaw “eight-ball.” The ADI included digital readouts of ground speed (in knots), altitude (in ft, height above the landing Figure 7: Lunar Landing Flight Deck zone), radar altitude (in ft, height terrain below extended gear position), and vertical descent rate (in ft/sec). On the Control Inceptors vertical descent rate display, a magenta circle indicated the guidance-commanded descent rate. Yellow pitch, roll, and The EPs used an Apollo-vintage Rotational Hand Controller yaw “flight director” needles were also displayed depending (RHC) and Translational Hand Controller (THC). The upon the experimental condition. The needles provided “fly- controllers were mechanical and their characteristics fixed.

to” pitch, roll, and yaw attitude guidance to fly the The right-hand RHC provided 3-axis rotational control trajectories to landing. Additional, less-critical status mode commands, using a base pivot for roll, palm-pivot for pitch, information was shown in periphery.

and twist-pivot for yaw. The mechanical characteristics of the RHC provided +/- 27.5 deg total deflection in roll using a maximum force of 3.5 lb applied at the roll pivot point (five inches above the roll rotation point, approximately mid-grip). There was approximately 0.5 lb of hysteresis with 1.0 lb of break-out force.

The mechanical characteristics of the RHC provided +/-15.0 deg total deflection in pitch with maximum force of 3.5 lb applied at the pitch pivot point (~2.625 inches above the rotation point, near top of grip). There was approximately 0.75 lb of hysteresis with 0.25 lb of break-out force.

A deadband function using 15% of the full-throw deflections was applied to the pitch and roll signal and scaled to create a +/-1.0 command signal. The scaled roll and pitch inputs (  ) were parabolically shaped to create the pitch rate and roll rate commands ( q and p , respectively) s s כ ߜ כ ݁݌ ݄ܽܵ ൅ ሻ using the equation ݌ ݍ ݁݌ ݄ܽܵെ ሺ1 כ ߜ ൌ ௦, ௦ |ߜ| where the term ݄ܽܵ ݁݌ was equal to 0.90. The shaping Figure 8: ADI function appropriately tailored the response sensitivity for the sidestick controller.

The ND provided a top-down, “God’s eye” view of ownship position and the planned landing zone (Fig. 9). The ND was always drawn with ownship in the center, heading up, and guidance provided corrections to pitch angle, roll angle the range scale automatically adjusted to keep the landing and thrust level of the vehicle to correct for any zone on the display. Synthetic imagery was used in the deviations from the desired trajectory.

background. The ND transitioned from a flight-mode to a • The guidance mode switched to a hover phase at a hover mode once the ground speed was less than 38 knots.

fixed-time prior to the end of the straight-line approach phase trajectory. In this phase, pitch angle, roll angle • In the flight-mode, the landing zone was depicted by a and thrust level commands would bring the vehicle to a magenta-colored hexagon symbol.

hover over the landing target at ~100 ft height.

• In hover mode, the landing zone was depicted by two • The final descent phase was initiated once the pitch, octagons. The octagons were sized to accurately roll, velocity and horizontal position of the vehicle was represent the size of the desired and adequate landing within parameters over the designated landing area.

performance standards. In addition, digital display of the body-axis forward and side velocities were Hover Cue Guidance presented with a graphical representation of this same An alternative form of flight guidance was mechanized for information, indicated by a white line and cross symbol experimental variation as a possibly simpler means of emanating from the ownship position.

enabling lunar landings and tailored toward manual control • In the hover mode, a hover cue symbol was also considerations. In this case, the guidance was in the form of displayed, depending upon the experimental condition.

a hover cue. The hover cue is part of an integrated symbology set, including the ownship position symbol, velocity vector, and conformal landing zone designator.

The hover cue was mechanized as a “fly-to” symbology element where the positioning and dynamics of the cue provided manual control guidance for the pilot to smoothly approach a desired landing position and achieve zero horizontal velocity (i.e., a hover condition). The pilot’s task is to use the pitch and roll inceptor to place the hover cue and hold its position over the desired landing site (i.e., “put the ball in the octagon”).

The hover cue closely follows from rotorcraft work [42] with adaptation for a non-atmospheric flying vehicle as described in the following. The cue is, in essence, a flight director. The motivation for this “guidance” element is that, if successful, a hover cue would be ideally suited for a manually-controlled re-designated landing task, especially when coupled with a head-up display or head-worn display projecting this cue against conformal imagery or the actual Figure 9: ND – Hover Mode lunar terrain. In these cases, the hover cue would be controlled by the pilot, superimposed on a pilot-selected The VSD provided a profile/side view of ownship position landing site; the hover cue “guidance” provides a simple and the landing zone. The auxiliary display provided a means of achieving a hover over a pilot-designated point.

simulated forward looking infrared (FLIR) image of the o o outside world from 45 to 75 nose-low from the DERP.

For this work, the conformal landing zone designator was This camera view supplemented the EP’s visibility below the desired touchdown/landing position. The velocity vector o the 45 cut-off of the simulated window.

provided a graphical depiction of the horizontal velocity (digital readouts of forward and side velocity were also Command Guidance provided). The hover cue was shown on either the ND or on The ADI flight director bars were driven to provide explicit a Head-Worn Display (HWD) [41].

pitch, roll, and yaw command guidance to fly an approach path, transition to a vertical descent, and conduct a vertical Ownship position was fixed in the center of the display and landing. The guidance was defined in a series of phases, automatic range settings were applied to keep the landing depending on distance from the landing site. zone at a reasonable range scale on the displays.

The command guidance used three phases: Earth-bound vehicle applications of the hover cue guidance have been implemented, but to the authors’ knowledge, this • The approach phase used a constant deceleration profile was the first lunar lander application.

which allowed for a near-constant pitch (deck) angle, flight path angle and thrust-to-weight ratio. The In the design of the hover cue dynamics, it has been Table 2 Values for RCAH Time Constant / Hover Cue assumed that three coincident real-axis roots between aircraft velocity and stick position produced desirable RCS Jet Size Control Power   dynamics [42]. The positioning of the hover cue followed this assumption where the pilot applies compensation to 68 lbf 1.1 deg/sec 7.0 sec control the hover cue in a position (x) closure task (Figure 2 100 lbf 1.6 deg/sec 5.0 sec 10).

180 lbf 2.9 deg/sec 2.75 sec With this assumption, the longitudinal position of the hover 270 lbf 4.3 deg/sec 1.75 sec cue, A is derived. The lateral position, A , was directly x y 540 lbf 8.6 deg/sec 1.0 sec analogous, but not shown here.

The longitudinal position was computed from: Navigation Display ߱ଷ ߱൅ ݏሺ ሻ ଴ ܣ ܭ ൌ ሶݔ ௫ ௫ሶ Conformal Landing ଷ ଴ Zone Designator Hover Cue where ܭ is a display gain, deg/ft/sec.

௫ሶ (Kx*xc, Ky*yc) ߱ is a bandwidth parameter, rad/sec ௢ (Ax, Ay) ሶݔ is the body-axis forward velocity, ft/sec (Ku*u, Kv*v) This equation is algebraically expanded to yield: ܭ ܭ ܭ ߱௫ሶ ߱௫ሶ ߱௫ሶ ଶ ܣ ൌ ݏሷݔ ൅ 3 ݔሷ ݏ൅ 3 ܭ ൅ ሷݔ ሶݔ ௫ ௫ሶ ଷ ଶ Velocity Vector ଴ ௢ ଴ The term, ܭ ሶݔ positions the reference for the hover cue at ௫ሶ the same location as the velocity vector, where ܭ ݑܭ ൌ . in Ownship Position ௫ሶ Figure 10. For our lunar lander application, a small angle ൗ ݉ܶെ approximation was invoked so ݔሷ ൌ ߠ where T is the vehicle thrust, m is the mass, and  is the pitch attitude.

Thrust and mass were real-time variables in the guidance ൗ ݉ܶെ calculation. It then follows that ൌ ݏሷݔ ݍ , where q is the vehicle pitch rate.

ଶ The first term, ݏሷݔ was derived using the RHC stick position (  ). The vehicle control laws were RCAH-type, but manifested as a discrete, phase-plane controller without Figure 10: Illustration of Hover Symbology aerodynamic damping effects. Using a loose approximation ݍ ܭ ோ஼஺ு ଶ for an RCAH control law, ൌ , the ݏሷݔ ൗ ൗ The sensitivity or scaling of the hover cue critically ߜ ሺ ݏ߬൅ 1ሻ influences its utility and usability. The selections and ݏ ൗ ݉ܶെ term was approximated by ൫ ܭ൯ ቀ ߜ ቁ .

ൗ ோ஼஺ு ሺ ݏ߬൅ 1ሻ rationale were as follows: By this assumption, the hover cue effectively uses a • The ND range was automatically adjusted to keep the washed-out RHC input (  ) where the time constant,  , is landing zone conformally on the screen. The visual dependent upon the responsiveness of the pitch rate angle subtended by the ND was approximately 20 deg.

controller. The time constant (  ) was selected to Therefore, the display unit scaling, K , was equal to 20 x approximate the pitch rate response of the phase-plane deg/ND , where ND was the automatically set range range controllers and the gain, K was equal to the maximum, RCAH range scale in ft. The minimum navigation scale range full stick deflection rate command (12 deg/sec). Since the (ND ) was 200 ft.

range angular acceleration was experimentally varied, the selected • Pre-test experimentation indicated that a bandwidth values for  (shown in Table 2) were varied accordingly.

selection (  These  values were defined by the approximate time to 0) of 0.5 rad/sec was a reasonable selection considering the piloting demands for this task and the reach 63% of the spacecraft’s rate response for a full-stick vehicle dynamics being considered as “representative” input command.

for a lunar landing mission.

• As detailed in [42], the relationship between the velocity scaling and the display scaling (in addition of the bandwidth parameter) effectively specify the speed of the position loop closure response. For the selected disconnect alert was triggered. Concurrently, the landing bandwidth, root locus calculations for the assumed loop zone was automatically re-designated left or right by 250 ft closure task dictated that a well-damped loop gain, from its original position. The EP was then tasked to fly and ܭ land on this re-designated landing zone using either the ௫ ൗ should be 0.065 (which provided a damping ܭ ௫ሶ hover cue (hereafter referred to as the “aided” condition) or ratio of 0.85). The velocity display scaling ܭ falls out.

௫ሶ explicit pitch/roll guidance (hereafter referred to as the “guided” condition).

Smoothing was applied to the input signals to temper (somewhat) the jerkiness of the symbol movement due to The 250 ft offset landing zone mimicked the scenario where the discrete nature of the RCS firings. First-order, low-pass a re-designated landing zone had been executed and the filters were applied to the attitudes, and attitude rates with pilot must now fly to the new landing area. The scenario time constants of 0.2 sec and 0.1 sec, respectively. First- was also analogous to the offset landing task from the order, low-pass filters were also applied to the RHC inputs aeronautics domain. The 250 ft offset was chosen to match to eliminate high-frequency noise in the signal, using a time previously conducted work [31,38].

constant of 0.2 sec.

The pilot’s task was to control the vehicle’s trajectory, fly to Experiment a hover transition, and execute a vertical descent to land to achieve prescribed landing performance standards using The primary experimental goal was to evaluate the either explicit guidance or a hover cue in addition to interaction of controls and displays on the handling qualities available out-the-window visual cues.

of a lunar landing vehicle. Variations in approach trajectory were experimentally tested because they interact with fuel In both conditions, the desired descent rate to fly to the and natural vision viewability of the landing zone.

hover transition point, developed from the guidance concept described above, was provided. Vertical descent rate was A full-factorial matrix was developed, consisting of 3 controlled by discrete THC inputs.

approach trajectories (15, 30, and 45 deg), five control powers (Table 2), and 2 guidance conditions (head-down The desired and adequate performance standards are shown guidance on ADI and hover cue flight director). The hover in Table 3. These parameters only applied to the landing. At cue was also evaluated as part of a “virtual HUD” concept the end of each run, a “scorecard” was shown on their head- using a HWD [41]. Due to time limitations, a subset of this down display given them their performance against each of matrix was identified as the primary matrix which the these standards.

majority of EPs were to evaluate; the remainder of the matrix was tested as time permitted.

Table 3 Task Performance Standards Each run of the experiment began 1000 ft above the landing target. Three different approach angles were experimentally Parameter Desired Adequate varied: 15 deg; 30 deg; and 45 deg. The initial distance up- Performance Performance range from the landing site was varied accordingly. The Range At < 15 ft < 25 ft initial velocity varied as well to provide approximately the Touchdown same duration of the approach from start to finish.

Sink Rate < 5 ft/sec < 7 ft/sec Evaluation Tasks —Each run started at approximately 1000 Forward/Side < 2 ft/sec < 4 ft/sec ft above the landing zone elevation flying on auto-pilot Velocity following the constant deceleration trajectory profile, to a Pitch/Roll Angle < ± 3 deg < ± 6 deg designated landing zone directly forward of the vehicle’s Pitch/Roll Rate < ± 3 deg/sec < ± 6 deg/sec track. This period of auto-flight provided two benefits: 1) it Yaw Rate < ± 1.0 deg/sec < ± 1.5 ft/sec allowed the pilots to prepare for the next phase (manual flight); and, 2) it was representative of an operational profile, where the auto-flight system will fly to a certain No performance standards were enforced during the altitude, but the crew will take-over manually for the final approach phase with the exception of a vertical descent rate landing. This procedure was done for all Apollo flights [27] and will likely occur for other crewed missions. An performance standard. The pilots were instructed to follow the vertical descent rate guidance, but the pilots were told additional motivation is that this auto-flight period might be that if they stopped their descent, their task performance representative of a re-designation period where the crew could be no better than “adequate”. If they climbed, their might be evaluating real-time sensor data and navigation task performance was considered “not adequate.” Fuel systems data to decide if the planned landing zone is safe usage (propellant consumption) was not a parameter in the and appropriate or if a landing zone re-designation is task performance standards, because this test was designed necessary [29].

as an evaluation of handling qualities. Fuel consumption, of At 500 ft above the touchdown zone, the auto-flight system course, is critical for a spacecraft design and it is analyzed was disconnected automatically. An aural auto-flight post-test because of its criticality. To use it as a Cooper- reasonable piloting task. The flight crew had sufficient time Harper task performance standard would skew the piloting to effect a landing.

technique to minimize propellant usage. The vertical At the hover transition point, the constant deceleration descent rate requirement was the one concession to profile could create a significant pitch task. This task was criticality of fuel. The descent rate criteria essentially most demanding for the lower glideslope angles. For required that the vehicle was continually descending toward instance, the 15 deg glideslope approach used a fairly large the landing zone. If a hover condition, or worse yet, a climb o pitch attitude (~20 pitch attitude) that had to be removed is performed, fuel consumption will be unreasonable given very quickly and precisely as the hover transition point was that this type of vehicle will undoubtedly be “fuel-critical.” reached. If not, the vehicle would undershoot or overshoot While a hover or climb may be technically feasible, it would the landing zone and then, low altitude re-corrections were have to be a last-ditch maneuver. A hover or climb necessary. In contrast, the 45 deg glideslope approach essentially buys the crew more time to control the vehicle.

involved very little pitch attitude change. The ground speed This is a sign of poor handling qualities since the landing approaching the landing zone was considerably less. As site is assumed to be clear of obstructions. Therefore, task such, the hover transition was neither as large nor did it performance penalties were enforced.

require as high a degree of precision. The 45 deg approach Other Elements task was more of a lateral tracking task, whose duration consists of the entire approach. The 15 deg approach task Experiment Protocol —Prior to the start of data collection, was much more demanding, involving both the lateral offset the EPs were briefed on the purpose of the experiment and and the abrupt pitch-over near the hover transition point.

flew numerous practice trials following the experiment protocol with a variety of experimental configurations until These observations were quantified by the HQRs as shown they reached an acceptable, consistent level of proficiency in Figure 11. The pilot rating data is plotted for the highest in the task. Upon the start of formal data collection, for each 2 control power condition (8.6 deg/s ) as the task and configuration, the EPs first flew one approach and landing guidance method was varied. For the shallow 15 deg for practice/familiarity, followed by a minimum of two runs glideslope, some Level 1 and some Level 3 ratings were for “data.” The EP had the option for a third run if they felt given. These significant rating discrepancies reflect subtle that the two runs exhibited very different characteristics that timing and piloting differences in transitioning to the hover potentially hampered their handling qualities assessment.

phase can create significant performance differences.

Conversely, predominately Level 1 ratings were given in the Following the tasks, the EPs assigned a HQR using the 45 deg glideslope, reflecting an easier piloting.

Cooper-Harper rating scale. Pilot comments were then given, generally prompted via a comment card. Finally, Apollo used more of a constant jerk profile (first derivative NASA Task Load Index (TLX) workload ratings were of acceleration) [43] with a tailored pitch-up prior to the given.

hover transition. Apollo experience showed that a constant deceleration profile was not preferred because of the abrupt Evaluation Pilots —Eight EPs flew this experiment, pitch change and the need for a natural vision viewing completed in 2009. All subjects had graduated as pilots opportunity. These data show that a constant deceleration from the U.S. Air Force or Navy Test Pilot Schools and profile is do-able, especially for higher glideslopes.

were experienced in aircraft handling qualities evaluations.

Needless to say, none were experienced in lunar landings.

Some had considerable experience in rotary wing vehicles.

Two of the EPs were current or former Astronauts.

9 1 Rating 2 Ratings Holistic Impact of Design/HQs 8 3 Ratings 4 Ratings Level 3 The influence – and interdependency – of manual control handling qualities and the vehicle design considerations are described using the handling qualities results.

Level 2 Effect of Glideslope Variations – Piloting Task Influence — By the nature of the constant deceleration approach profile, the approach phase was essentially a pitch attitude Cooper-Harper Pilot Rating maintenance task. Very few pitch inputs were required to Level 1 maintain the approach path once manual control was 15 deg 30 deg 45 deg required. So the initial task was primarily a roll task to start Guide Aided Guide Aided Guide Aided tracking toward the new landing zone. The pilot had from Glideslope the 500 ft altitude point to the hover transition to correct for the lateral offset. For all glideslopes flown, the 250 ft offset Figure 11: HQRs of Glideslope and Guidance Variation approach (simulating a landing point redesignation) was a 2 (8.6 deg/s control power) Guidance Method —Two concepts for guidance were tested, The pilot rating and comment data also indicated areas Guided (using ADI needles) vs. Aided (using a hover cue on where the hover cue needs to be improved as well. The the navigation display). discrete nature of the control law induced some substantial “jumpiness” in the position of the symbol. Throttle changes The data generally showed that explicit attitude guidance caused by descent rate control inputs and RCS deadband was good in that it provide simple and straight-forward firings were most problematic. Pilots often learned to command information to enable the transition from the “filter” this out, but they desired more predictable response approach to a hover over the landing area. The command behavior. Also, pilots could get behind the hover cue and information provided guidance as to when and what enter into overcontrol / pilot-induced oscillations for the low magnitude of pitch and roll attitude changes were required control power configurations. Vehicle attitude still needed to establish a hover over the landing zone. The display was to be monitored and this information was not in immediate unambiguous and clear and all critical flight information spatial proximity. Methods to ameliorate these (attitude, guidance, speeds, and descent rate information) characteristics were identified but not implemented at this was in close visual proximity. Especially for low control time. The aided condition was clearly preferred especially power configurations, vehicle attitude was critical “lead” since it broke the criticality of timing in successfully information required to keep the vehicle under control and completing the task. The pilots were more directly in control to anticipate the size and timing of control inputs to get into of the hover transition task.

the hover position.

Head-Down vs. Head-Up Information —The various Unfortunately, the pitch attitude change, especially for the trajectories, in addition to piloting task differences, 15 deg approaches near the hover transition point, required modulated the amount of time that the EPs could use the close attention by the EP. It also required precision on the OTW visual cues for the task.

part of the EP to track the attitude to achieve desired performance. If neither of these conditions were met, the • The task was flown almost completely head-down. All vehicle could significantly overshoot or undershoot the information required to conduct the task was generating low altitude corrections and workload on the part available on the ADI and ND.

of the pilot to use the guidance and the ND to re-correct to • However, almost all EPs noted the importance out-the- the landing zone.

window visual cues for verification and validation of the head-down information. This was not part of the test This control behavior is most apparent in the HQR data for but it should be evaluated in the future. For this test, the the lowest control power test (1.1 deg/s ) as shown in Figure navigation solution was perfect (no errors) so the ND 12. Clear Level 3 handling qualities were indicated for 15 presentation was a completely faithful representation of deg glideslope landing tasks using ADI guidance. As the the outside landing zone.

task got easier, the impact was not as severe (15 to 45 deg • The OTW visual cues were insufficient alone for the glideslope).

EPs, in general, to control the vehicle – at least to the degree that they could in rotary wing aircraft. Higher resolution terrain data is required to test and possibly, achieve this capability.

• Head-up information, conformal information was desired/preferred. The ability to use the SV/EV Level 3 information and out-the-window was expressed.

• It was also noted by the astronauts that the landing task was going to be a “crew effort.” The out-the-window Level 2 visual cues may not be critical for the commander (i.e., pilot-flying), because they will be part of the crew’s 3 1 Rating 2 Ratings tasking, where the pilot (i.e., pilot-not-flying), might 2 3 Ratings Cooper-Harper Pilot Rating visually designate the landing zone and verify/validate 4 Ratings Level 1 the navigation position accuracy.

15 deg 30 deg 45 deg Guide Aided Guide Aided Guide Aided The HWD showed significant potential in providing head- Glideslope up information while simultaneously showing hover cue Figure 12: HQRs of Glideslope and Guidance Variation (aided) guidance [41]. Those with previous familiarity with (1.1 deg/s control power) the hover cue (e.g. helicopter pilots) especially showed proficiency using this combination. The combination of In general, the EPs preferred the aided condition, using the SV/EV information, head-up conformality, and appropriate hover cue. The cue took some training (for those that didn’t crew tasking suggest that a long pitch-up maneuver for have experience with it), but after training, the cue provided natural vision acquisition of the landing zone will not likely some captures of the landing zone.

be required.

Effect of Control Power —The data – the pilot ratings and “manual control” and “automated control” is the physical comments – clearly indicate that increasing control power process of activating the controller (i.e., whether it is a produces better handling qualities. For the Apollo-type human or automation system at the controls). The system configurations (2-jet and 4-jet Apollo RCS with 4.3 and 8.6 performance and workload (i.e., handling qualities) during deg/s angular acceleration), Level 1 ratings were generally this operation are still critically important.

given. On some evaluations, some improvement was still Human-Centered Automation warranted with the higher control power configurations.

These poorer ratings were due primarily from display History has shown that the best automation designs are deficiencies where improvements were desired.

human-centered. They are designed understanding the human needs and limitations in monitoring, intervention, Control powers at 1.6 deg/s and below are borderline Level and adaptation with automation.

2-3 or Level 3. Even though desired performance may be achievable with these configurations, and some “good” Human-centered automation design principles are many and Level 2 ratings were given, cliff-like characteristics are varied (e.g., see [44]). Three key principles for human- apparent in these configurations. If the EP fell behind the automation interaction are directly related to the handling task, made an inadvertent input, or committed a control qualities process and must be addressed: reversal, they could find themselves struggling to get back to the landing zone. Pilot ratings of 8, 9, and 10 were 1) The automation must be observable by its human indicative of these conditions.

operator (i.e., the human operator is appropriately informed).

Approach Phase Piloting Task Influence —Stringent approach performance criteria were not in effect with the 2) The current and near-term future behaviors of the exception of the “don’t hover,” “don’t climb” vertical automation must be comprehensible, understood and descent rate requirement. This procedure may have allowed predictable by the human operator.

some of the “less than desirable” configurations to achieve “reasonable” landing performance which may have biased 3) The automation must be contextually appropriate for the ratings toward better ratings - the EPs were able to solve its application, designed to complement the human the conundrum presented by these sluggish configurations.

operator, and not automated just because it is possible.

The EPs learned that THC control inputs could modulate the amount of time available to them to complete the approach These human-automation interaction issues are mirrored in and landing. If handling issues became apparent, they could the lessons-learned from Apollo and their use of automated use the THC to reduce their rate of descent, in essence, entry guidance and control [45]: buying them more time to sort out the horizontal positioning • “Guidance logic must be simple… the more control problem. In many cases, desired landing performance was achievable, even though the approaches complicated the guidance logic, the more difficult the guidance is to monitor during the mission. The were far from optimal. However, these performance differences were evident in the fuel usage. Better handling monitoring difficulty complicates the development of the monitoring procedures and increased the time configurations used less fuel.

required for flight-crew training.” More work is needed in defining approach standards for the • “The guidance logic should be compatible with a task. It may be necessary and appropriate to establish backup or an alternate trajectory control procedures.

approach corridors and approach speeds. The approach That is, once an anomaly is detected in the trajectory standards might mimic a Landing Signal Officer, where an control of the primary guidance system, an alternative unstabilized or botched approach might warrant a “wave- technique must be available that will allow satisfactory off” (i.e., trigger an abort) to avoid the possibility of damage trajectory control to be implemented so that the to an existing lunar habitat.

spacecraft will land near the originally selected target.” • “The interaction between guidance system performance 5. H ANDLING Q UALITIES FOR A UTOMATION and attitude control system performance must be recognized. Realistic attitude control system response Often, the importance of handling qualities is dismissed requirements must be established, and guidance-logic under the premise that the task/mission objective will be met design must minimize the need for rapid response.” using automated control. In fact, handling qualities or, at least, the principles of the pilot-vehicle dynamic systems HQ Evaluations For Human-Centered Automation analysis are critically important during automatic control.

The handling qualities of automated tasks (i.e., human- Human-automation interface requirements flow from automation interface requirements) should be evaluated in “manual control” handling qualities. Handling qualities are three ways: the design and evaluation of the pilot-vehicle dynamic system (Figure 1); the only fundamental difference between 1) Pilots should conduct handling qualities of all tasks to response as the pilot performs the required operation or task.

be flown by the automation. This is not to imply that these tasks could or should be manually flown. But by The handling qualities implications of design decisions are conducting these evaluations, the pilot gains an demonstrated using these pilot-in-the-loop evaluations of appreciation of what the automation must do to docking operations and lunar landings. A general trend, successfully control the vehicle. This knowledge is noted in both proximity operations and docking and lunar critical for understanding the behavior of the landing, is that superior handling qualities generally results automation and what the pilot must do in the event in lower fuel expenditure during the task. Data to support (likely or unlikely) that they need or want to intervene control and display requirements are given based on task or take-over. This task also defines what information performance and workload.

(e.g., out-the-window visual cues or displays) are needed by the human to monitor, control, or interact The importance of handling qualities is sometimes with the automation. dismissed under the premise that the task/mission objectives will be met using automated control. It is argued that 2) Classic handling qualities evaluations must also be human-automation interface requirements flow from conducted in scenarios where, during the conduct of “manual control” handling qualities. Therefore, handling automated tasks, the pilot takes control of the vehicle qualities are still critically important and these principles and completes the task or temporarily intervenes and should be closed adhered to.

then re-engages the automation. This task evaluates the ability of the crew to interrupt the automation and the R EFERENCES potential for upsets or discontinuities in the automation [1] Cooper , G.E., and Harper Jr., R.P.: “Use of Pilot during this process. It also demonstrates the pilot’s Rating in the Evaluation of Aircraft Handling comprehension and understanding of the automation.

Qualities.” NASA TN D-5153, April 1969.

3) Finally, handling qualities evaluations must be [2] Bailey, R.E., Jackson, E.B., Bilimoria, K., Mueller, E., conducted in scenarios where, during the conduct of Frost, C., and Alderete, T.: “Cooper-Harper automated tasks, the automation unexpectedly fails Experience Report for Spacecraft Handling Qualities (passively or actively) and the pilot must intervene or Applications,” NASA Langley Research Center, take control of the vehicle and complete the task.

Hampton, VA, NASA/TM-2009-215767, May 2009.

There has yet to be realized an automation design or [3] National Aeronautics and Space Administration: implementation which is fool-proof and perfect. Adaptive “Human-Rating Requirements for Space Systems.” automation is a promising, yet immature, technology.

NASA Procedural Requirements (NPR) 8705.2B, May Manual control of spacecraft is now and will be for the 06, 2008.

foreseeable future desired and/or required to “allow the crew to effectively control the spacecraft when necessary [4] Harper, R.P, Jr. and Cooper, G.E.: “Handling Qualities for mission completion or to prevent a catastrophic event” and Pilot Evaluation.” AIAA Journal of Guidance, [2]. A human-centered automation design, using “handling Control, and Dynamics, Vol. 9, September-October qualities” principles, ensures that the greatest adaptive 1986, pp. 515-529.

controller ever designed – the human – can intervene, adapt, and overcome as necessary in the event that the automation [5] Young, L.R.: “On Adaptive Manual Controls.” IEEE is not successful.

Transactions on Man-Machine Systems, Vol. 10 No. 4, pp. 292-331.

6. C ONCLUDING R EMARKS [6] Riley, D.R., Jaquet, B.M., Bardusch, R.E., and Deal, A research, development, test, and evaluation process was P.L.” “A Study of Gemini-Agena Docking Using a put into effect to identify, understand, and interpret the Fixed-Base Simulator Employing a Closed-Circuit engineering and human factors principles which govern the Television System.” NASA TN D-3112, 1965.

pilot-vehicle dynamic system as they pertain to space exploration missions and tasks. Toward this objective, [7] Riley, D.R., Jaquet, B.M., and Cobb, J.B.: “Effect of piloted simulations were conducted at the NASA Langley Target Angular Oscillations on Pilot-Controlled Research Center and Ames Research Center for earth-orbit Gemini-Agena Docking.” NASA TN D-3403, 1966.

proximity operations and docking and lunar landing.

[8] Long, E.R., Jr., Pennington, J.E., and Deal, P.L.: These works provide broad guidelines for the design of “Remote Pilot-Controlled Docking With Television.” spacecraft to exhibit good handling characteristics. In NASA TN D-3044, October 1965.

particular, this work demonstrates how handling qualities include much more than just stability and control [9] Pennington, J.E., Hatch, H.G., Jr., Long, E.R., and characteristics of a spacecraft or aircraft. Handling qualities Cobb, J.B.: “Visual Aspects of a Full-Size Pilot- include the motion, visual and aural cues of the vehicle Controlled Simulation of the Gemini-Agena Docking.” [21] Jackson, E.B., Goodrich, K.H., Bailey, R.E., Barnes, NASA TN D-2632, 1965. J.R., Ragsdale, W.A., and Neuhaus, J.R.: “Investigation of Control System and Display [10] Jaquet, B.M. and Riley, D.R.: “An Evaluation of Variations on Spacecraft Handling Qualities for Gemini Hand Controllers and Instruments for Docking with Stationary and Rotating Targets.” NASA Docking.” NASA TM X-1066, 1965.

Langley Research Center, Hampton, VA, NASA/TM– 2010-216194. Feb 2010.

[11] Riley, D.R., Jaquet, B.M., Pennington, J.E., and Brissenden, R.F.: “Comparison of Results of Two [22] Hackler, C.T., Brickel, J.R., Smith, H.E., and Simulations Employing Full-Size Visual Cues for Cheatham, D.C.: “Lunar Module Pilot Control Pilot-Controlled Gemini-Agena Docking.” NASA TN Considerations.” NASA TN D-4131, February 1968.

D-3687, Nov. 1966.

[23] Mindell, D.A.: “Digital Apollo: Human and Machine [12] Hatch, H.G., Jr., Riley, D.R., and Cobb, J.B.: in Spaceflight.” The MIT Press, Cambridge, MA, “Simulating Gemini-Agena Docking.” Astronautics 2008.

and Aeronautics, Vol. 2, No. 11, Nov. 1964, pp. 74-81.

[24] Major, L.M, Brady, T.M., and Paschall II, S.C.: [13] Cheatham, D.C. and Hackler, C.T.: “Handling “Apollo Looking Forward: Crew Task Challenges.” Qualities For Pilot Control Of Apollo Lunar-Landing Paper presented at the 2009 IEEE Aerospace Spacecraft.” Journal Of Spacecraft And Rockets, Vol.

Conference, 7-14 March 2009.

3, No. 5, 1966, pp. 632-638.

[25] Engle, M.: “Operational Considerations for Manned [14] Stengel, R.F.: “Manual Attitude Control Of The Lunar Lunar Landing Missions – Lessons Learned From Module.” Journal of Spacecraft and Rockets, Vol. 7, Apollo.” Paper presented at the Space 2004 No. 8, 1970, pp. 941-948.

Conference and Exhibit, AIAA Paper No. 2004-6081, San Diego, California, 28-30 September 2004.

[15] Goodman, J.L.: “History of Space Shuttle Rendezvous and Proximity Operations.” Journal of Spacecraft and [26] Cohen, M.M.: “From Apollo LM to Altair: Design, Rockets, Vol. 43, No. 5, September-October 2006.

Environments, Infrastructure, Missions, and Operations.” Paper presented at AIAA SPACE 2009 [16] Stephens, J-P, Vos, G.A., Bilimoria, K.D., Mueller, Conference & Exposition, AIAA Paper No. 2009- E.R., Brazzel, J., and Spehar, P.: “Orion Handling 6404, Pasadena, California, 14-17 September 2009.

Qualities During ISS Proximity Operations and Docking.” AIAA Guidance, Navigation, and Control [27] Major, L.M., Brady, T.M., and Paschall II, S.C.: Conference, AIAA Paper 2011-6261. 8-11 August “Apollo Looking Forward: Crew Task Challenges.” 2011, Portland, Oregon.

Paper presented at the 2009 IEEE Aerospace Conference, Big Sky, MT, 7-14 March 2009.

[17] Bailey, R.E., Jackson, E.B., Goodrich, K.H., Ragsdale, W.A., Neuhaus, J., and Barnes, J.: “Investigation of [28] Prinzel II, L.J., Kramer, L.J., Norman, R.M., Arthur Reaction Control System Design on Spacecraft III, J.J., Williams, S.P., Shelton, K.J., and Bailey, R.

Handling Qualities for Docking.” AIAA Journal of E.: “Synthetic and Enhanced Vision System for Altair Guidance, Control, and Dynamics, Vol. 32, No. 6, Lunar Lander.” Paper presented at the 15th Nov-Dec 2009, pp. 1723-1735.

International Symposium on Aviation Psychology, Dayton, OH, 27-30 April 2009.

[18] Mueller, E., Bilimoria, K., and Frost, C.: “Dynamic Coupling and Control Response Effects on Spacecraft [29] Hirsh, R.L., Chua, Z.K., Heino, T.A., Strahan, A., Handling Qualities During Docking.” Journal of Major, L., and Duda, K.: “Developing a prototype Spacecraft and Rockets. Vol. 46, No. 6, pp. 1288- ALHAT Human System Interface for landing.” Paper 1297. 2009.

presented at the 2011 IEEE Aerospace Conference, Big Sky, MT, 5-12 March 2011.

[19] American Institute for Aeronautics and Astronautics: “Guide for the Serviceable Spacecraft [30] Bailey, R.E., Kramer, L.J., and Prinzel II, L.J.: “Fusion Grasping/Berthing/Docking Interfaces.” AIAA G-056- of Synthetic and Enhanced Vision For All-Weather 1992, American Institute for Aeronautics and Commercial Aviation Operations.” In NATO Human Astronautics, New York, New York, 1992.

Factors and Medicine Symposium on Human Factors and Medical Aspects of Day/Night All Weather [20] Bilimoria, K., Mueller, E., and Frost, C.: “Handling Operations: Current Issues and Future Challenges.

Qualities Evaluation of Piloting Tools for Spacecraft Neuilly-sur-Seine, France: NATO RTO, April 2007, Docking in Earth Orbit.” Journal of Spacecraft and pp. 11-1 – 11-18.

Rockets. Vol. 48, No. 5, September–October 2011.

[31] Mueller, E., Bilimoria, K.D., and Frost, C.: “Effects of Security International Symposium, Orlando, FL, April Control Power and Inceptor Sensitivity on Lunar 2010.

Lander Handling Qualities.” Paper presented at the [42] Schroeder, J.A. and Merrick, V.K.: “Control and AIAA SPACE 2009 Conference & Exposition, AIAA Display Combinations for Blind Vertical Landings.” Paper 2009-6407, Pasadena, CA, 14-17 September Journal of Guidance, Control, and Dynamics, Vol. 15, 2009.

No. 3, May-June 1992, pp. 751-760.

[32] Duda, K.R., Johnson, M.C., and Fill, T.J.: “Design and [43] Bennett, F.V.: “Apollo Experience Report – Mission Analysis of Lunar Lander Manual Control Modes.” Planning for Lunar Module Descent and Ascent.” Paper presented at the 2009 IEEE Aerospace NASA TN D-6846. June 1972.

Conference, 7-14 March 2009.

[44] Billings, C.E.: “Aviation Automation: The Search For [33] Peters, W.H. and Cox, K.J.: “Apollo Experience A Human-Centered Approach.” Lawrence Erlbaum Report Guidance And Control Systems - Digital Associates, Publishers. 1997 Mahwah, New Jersey.

Autopilot Design Development.” NASA TN D-7289, June 1973.

[45] Graves, C.A. and Harpold, J.C.: “Apollo Experience Report – Mission Planning for Apollo Entry.” NASA [34] Cooke, A.K.: “Rotary-Wing Control and Handling TN D-6725. March 1972.

Qualities.” Encyclopedia of Aerospace Engineering, John Wiley and Sons, Inc [35] Cooper, G. E.: “Understanding and Interpreting Pilot A CKNOWLEDGEMENTS Opinion.” Aeronautical Engineering Review, Vol. 16, This work was made possible only by the contributions of No. 3, March 1957, pp. 47-51.

many individuals in addition to the author’s. The contributions of Mr. Ken Goodrich (NASA LaRC), Mr.

[36] Jarvis, C.R.: “Flight-Test Evaluation of an On-Off James Barnes (ARINC), Mr. Jason Neuhaus (NASA LaRC), Rate Command Attitude Control System of a Manned and Mr. Al Ragsdale (Unisys, retired) were especially Lunar-Landing Research Vehicle.” NASA TN D-3903, critical to the success of these works. In addition, the April 1967.

assistance of the Simulation Development and Analysis Branch at NASA Langley, led by Ms. Victoria Chung, was [37] Slayton, D.K., Stafford, T.P., Armstrong, N.A., also gratefully appreciated.

Collins, M., and Cooper, L.G.: “Apollo: Past, Present, and Future.” Paper presented at The Society of th B IOGRAPHIES Experimental Test Pilots 13 Symposium, Beverly Hills, CA, 25-27 September 1969, Randall Bailey received a B.S. in Aerospace Engineering from the [38] Bilimoria, K.D.: “Effects of Control Power and University of Virginia and an M.S. in Guidance Cues on Lunar Lander Handling Qualities.” Mechanical Engineering from the State Paper presented at the AIAA SPACE 2008 Conference University of New York at Buffalo. He and Exposition, AIAA Paper 2008-7799, San Diego, has been with the NASA Langley California, 9-11 September, 2008 Research Center for more than 10 years. He is the technical team lead for [39] Cohanim, B.E., Fill, T.J., Paschall II, S., Major, L.M., flight deck interface technology in the Crew Systems and and Brady, T.: “Approach Phase Δ V Considerations Aviation Operations branch. Prior to joining NASA, Mr.

for Lunar Landing.” 2009 IEEE Aerospace Bailey worked at the Calspan Corporation serving as the Conference, 7-14 March 2009, pp 1-11.

Technical Director for the Flight Research Department.

[40] Leslie, R., Geyer, D., Cunningham, K., Madden, M.F., Bruce Jackson received a MS in Kenny, P.S., and Glaab, P.: “LaSRS++: An Object- Aeronautics and Astronautics from Oriented Framework for Real-Time Simulation of Stanford University in 1988. He has Aircraft.” AIAA Modeling and Simulation been a senior research engineer with Technologies Conference, AIAA Paper 98-4529, NASA Langley’s Dynamic Systems August 1998, Boston, MA.

and Control Branch for 22 years concentrating on simulation, [41] Arthur, III, J.J., Bailey, R.E., Jackson, E.B., Barnes, modeling, control design and flying qualities research; J.R., Williams, S.P., and Kramer, L.J.: “Part-task previously he worked at the Manned Flight Simulator simulation of synthetic and enhanced vision concepts Facility at the U.S. Navy’s Naval Air Test Center in for lunar landing.” Paper presented at 2010 Synthetic Patuxent River, MD.

and Enhanced Vision, SPIE Defense, Sensing and Trey Arthur received a BS in Aerospace Engineering from the North Carolina State University and an MS in Aeronautical Engineering from the George Washington University in 1997. He has been a research engineer in the Crew Systems Branch for 15 years with an emphasis on advanced head-up display technologies; previously, he worked at Computer Science Corporation as a software developer for NASA’s 737 research aircraft.