Minimum complexity guidance, navigation, and control for an autonomous parafoil payload delivery system
Date
Authors
Major Professor
Advisor
Committee Member
Journal Title
Journal ISSN
Volume Title
Publisher
Altmetrics
Abstract
A guided ram-air parachute, or parafoil, offers a lightweight and efficient means for the autonomous placement of payloads to specified ground coordinates. The technology has a wide variety of applications, including advanced precision airdrop operations and the safe return of high altitude balloon or sounding rocket payloads. Current systems rely on complicated and expensive sensor suites, and computationally intensive guidance and control algorithms. In the present work, an integrated guidance, navigation, and control scheme is presented which utilizes inexpensive "off-the-shelf" sensors and computationally simple guidance, navigation, and control (GNC) algorithms. The GNC scheme is intended for implementation on a Peripheral Interface Controller (PIC). The GNC scheme is intended to serve as the primary GNC scheme for a small-scale parachute system, but could equally well serve as a backup for a more complicated system in the event of CPU or sensor failure. Navigation data is collected from a GPS receiver and a 3-axis digital compass, sampled at 1 Hz and 2.5 Hz, respectively. The wind speed and heading are estimated using the supplied navigation data and an estimate of the vehicle airspeed. It is shown that an estimation accuracy of better than 3 ft/s for velocity and 10 degrees for heading can be achieved without additional filtering. A linear proportional feedback controller is designed to control the vehicle course angle. It is shown that the effect of the slow sample rate of the navigation sensor data acts as a filter in the forward loop and contributes non-minimum phase zeros to the closed-loop dynamics. The guidance algorithm generates a series of inertially fixed waypoints consisting of three distinct phases: initial homing, energy management, and final approach. The algorithm requires no a priori information about the wind profile. A six degree-of-freedom model is used to evaluate the performance of the integrated GNC algorithm. Monte-Carlo simulations were conducted for a number of wind profiles which were specifically chosen to test various design parameters. The combined results show a circular error precision (CEP) landing accuracy of 115 ft. It is shown that the primary factors limiting landing accuracy are the maximum actuator deflection and the low system bandwidth.