Cornell University LIMS website

    Perching aircraft project

Objectives:

• • To enable vertical or very short landing in high efficiency (low thrust-to-weight ratio) reconnaissance aircraft through shape reconfiguration. Using aerodynamics for perching avoids the weight penalty of high thrust-generating devices, such as those employed by helicopters and jump jets.
• • To model the unsteady aerodynamics of perching through analytical methods and wind tunnel experimentation. Some of these effects include dynamic stall, post-stall vortex shedding, and wake interference on the tail.
• • To optimize a perching trajectory given the aerodynamic constraints of the vehicle. Costs to minimize include time to landing, control effort during the maneuver, undershoot of the landing target (if allowed), and time spent under a specified velocity
• • To develop a controller to reject disturbances and track the perching trajectory. This controller has time-varying gains dependent on the morphological state of the vehicle

Overview of the Aircraft and its Aerodynamics

The perching aircraft concept is a blended wing-body with an inverted V-tail attached by two booms. This design is based off of the Mars ARES scout aircraft, a vehicle that would unfold from a re-entry capsule and fly over the planet's surface, gathering atmospheric and geological data. The idea to perch this airframe grew from the problem of trying to save the ARES from a crash landing without adding too much complexity to the system; however, the research we are undertaking has expanded into more terrestrial applications such as landing in an urban environment. Three additional degrees of freedom are added to our perching aircraft prototype, as depicted in Figure 4. These additional actuators allow the aircraft to morph in flight in order to perch, which requires a large amount of lift and drag. The fuselage is pitched up past stall in order to generate a large amount of drag. The wings rotate downward in order to remain horizontal, and the tail rotates out of the unsteady wake behind the fuselage. The attached flow over the wings and tail generates lift and maintains controllability about all 3 axes throughout the maneuver.

The aerodynamics involved in this maneuver are highly nonlinear and unsteady. Although they can be approximated by empirical relations, one aspect of our project is to characterize the aerodynamics from wind tunnel experimental data. Figure 5 shows the 2' wingspan model in the tunnel. This model is rapid prototyped from ABS plastic and is fully articulated with standard aircraft servos and additional DC motors for morphing. A 6 axis load cell records aerodynamic forces and moments real-time while the aircraft morphs in the flow. This enables the nonlinear aerodynamics of morphing and post-stall flow to be studied and modeled for flight simulation and control design.

Trajectory Optimization

The trajectory optimization problem is formulated as a two-point boundary value problem between a cruising state and a perching state at a specified point. For this study, the spatial bounds of the trajectory are to be minimized. With a thrust-to-weight ratio less than one, the trajectory invariably "undershoots" the landing site, since an increase in potential energy is required to reduce the speed to zero. Thus, this study addresses two major concerns: minimizing the undershoot and minimizing the distance from the landing site required to start the maneuver. Both goals have great practical value. Minimizing the undershoot is important due to spatial limitations, since the landing site may be close to or at ground level or may be obstructed by objects in the environment. Minimizing the required distance to start the maneuver is important because on-board sensors – such as CCD cameras, radar, or infrared range finders – have a finite range at which they can accurately identify and track the landing site. Increasing these sensors’ ranges generally means increasing their size and weight, thereby increasing the demands on the aircraft design.

The persistent undershoot in the trajectories at low thrust-to-weight ratios motivates the division of the problem into two halves. For this study, the goal of minimizing undershoot is given higher priority; thus, the problem can be divided and solved sequentially. These two halves, called the dive phase and the climb phase, are depicted in Figure 6. The climb phase can be solved first for minimum undershoot because the global (over all possible trajectories) minimum undershoot is a function only of the climb phase since only the dynamics here determine how quickly the aircraft can pull up to the landing site with the specified final speed. The dive phase then connects the initial condition to the starting point of the climb phase. This proposition assumes that the initial condition of the climb phase is a reachable end condition of the dive phase. The objective of the dive phase is to minimize the starting distance required to attain a final condition that matches the initial condition of the climb phase, which is stipulated to be the aircraft’s straight and level trim point for maximum endurance. The requirement to minimize maximum undershoot determines the optimal trajectory for the climb phase, which in turn determines the end condition of the dive phase. Thus, the optimal solution for the dive phase is only optimal among the set of trajectories that match up with the optimal trajectory for the climb phase. Although the solution to the dive phase may be suboptimal over all possible trajectories, it is optimal given the constraint that minimizing the maximum undershoot is the highest priority. Thus, the computed optimal trajectory minimizes the undershoot over all possible trajectories and minimizes the starting distance over the subset of trajectories that minimize the undershoot.

In-flight morphing allows the aircraft to land over a much shorter distance by enabling much larger possible pitching moments. This is because the morphing actuator ranges are much larger than the elevator’s; therefore, the lifting surfaces are able to rotate to higher angles of attack in order to generate larger pitching moments. By comparing the trajectories in Figure 7, the direct results of these greater pitching moments can be seen by noting the reduced undershoot and starting distance in the morphing case.

Publications

• Wickenheiser, A. and Garcia, E. "Optimization of Perching Maneuvers Through Vehicle Morphing", Journal of Guidance, Control, and Dynamics, Vol. 31, No. 4, 2008, pp. 815-823. (PDF)
• Garcia, E., Wickenheiser, A., Dietl, J. and Manzo, J. "Morphing Aircraft in Perching Maneuvers", 3rd International Conference on Smart Materials, Structures, and Systems, June 8-13, Acireale, Sicily, 2008.
• Wickenheiser, A. and Garcia, E. "Dynamic Wind Tunnel Testing of Perching Maneuvers", 18th International Conference on Adaptive Structures and Technologies, October 3-5, Ottawa, ON, 2007. (PDF)
• Andrews, J. J. and Garcia, E. "Experimental Analysis of a Perching Aircraft Using Dynamic Testing", AIAA Region I-NE Student Conference, April 27-28, Cambridge, MA, 2007. (PDF)
• Wickenheiser, A. and Garcia, E. "Perching Aerodynamics and Trajectory Optimization", Smart Structures and Materials 2007: Active and Passive Smart Structures and Integrated Systems, March 18-22, San Diego, CA. published in: Proc. SPIE Vol. 6525, 65250O, 2007. (PDF)
• Wickenheiser, A. and Garcia, E. "Perching Trajectory Optimization Through Aircraft Morphing", CanSmart International Workshop on Smart Materials and Smart Structures, October 12-13, Toronto, Ontario, Canada, 2006. (PDF)
• Wickenheiser, A. and Garcia, E. "Longitudinal Dynamics of a Perching Aircraft", Journal of Aircraft, Vol. 43, No. 5, 2006, pp. 1386-1392. (PDF)
• Wickenheiser, A., Garcia, E., and Waszak, M. "Longitudinal dynamics of a perching aircraft concept", Smart Structures and Materials 2005: Smart Structures and Integrated Systems, March 6-10, San Diego, CA. published in: Proc. SPIE Vol. 5764, pp. 192-202, 2005. (PDF)