Performance of an Autonomous Guided Airdrop System with Glide Slope Control

Abstract

One of the primary difficulties in obtaining high landing accuracies with guided airdrop systems is the lack of effective longitudinal control. Except near canopy stall, symmetric brake deflection causes a change in flight speed with limited change in glide slope angle. Previous efforts have experimentally demonstrated direct control of glide slope by dynamically varying canopy incidence in flight. The current work explores the performance implications of incorporating this type of glide slope control into an autonomous guided airdrop system. Basic glide slope performance of the control mechanism is investigated with a six degree of freedom parafoil and payload simulation incorporating realistic wind and sensor models. It is shown that a simple PID control approach using the incidence angle to maintain an intercept path with the target can be very effective in light winds, providing a factor of three improvement in landing accuracy over current systems with no means of glide slope control. In strong winds, a more advanced controller utilizing incidence angle and symmetric brake to simultaneously control airspeed and glide slope is required to utilize the full benefit of the glide slope control mechanism. I. Introduction irdrop systems provide a unique capability of delivering large payloads to undeveloped and inaccessible locations. Traditionally, these systems have been unguided and, consequently, either a large landing zone is required or a high probability of losing individual payloads must be accepted. Beginning in the 1990’s, autonomous guided airdrop systems based on steerable, ram-air parafoils were developed with the goal of improving the precision and accuracy of air-dropped payload delivery. In practice, the gliding ability of the ram-air canopies can actually create major problems for airdrop systems by making them more susceptible to winds and allowing them to achieve far greater miss distances than were previously possible. Parafoil control mechanisms have not changed since the invention of the ram-air canopy in the 1960’s. The trailing edge of the canopy is deflected downward asymmetrically to turn and symmetrically to change speed or to flare while landing. Symmetric deflection of the trailing edge brakes produces an increase in both lift and drag; this provides a reduction in speed but little change in glide angle until stall. An alternative control mechanism is to change the rigging incidence angle in flight. By varying the length of the forward risers in concert with the brakes, the canopy can be rotated longitudinally to control the trim angle of attack directly in flight. The ability to change glide slope in flight in this manner was demonstrated on an airdrop system by Slegers, Beyer, and Costello [1]. Recent flight tests have extended this work by studying the influence of canopy aspect ratio on the effectiveness of glide slope control, the relationship between incidence angle control and symmetric brake deflection, and the transient response to rapid changes in incidence angle [2]. The current work seeks to use this body of flight test knowledge to develop autonomous glide slope control algorithms and explore the impact of glide slope control on landing accuracy in simulation. To this end, a six degree of freedom model was created to match the latest available flight test data using the methods described in Ref [3]. The simulation environment incorporates three dimensional, time varying wind fields generated using the Dryden turbulence spectrum. The simulation uses sensor error models to generate synthetic GPS and altimeter data for the guidance, navigation, and control (GNC) algorithm. The goals of the simulations are (1) to study the dynamics of the glide slope control mechanism and the coupling of the incidence angle control to the rest of the flight dynamics and (2) to generate a robust method for autonomous parafoil GNC incorporating glide slope control. 1