During the final approach and landing phase of the X-38/Crew Return Vehicle, a steerable parafoil is used to maneuver and land at a targeted ground base under autonomous control. To simulate and verify performance of the onboard Parafoil Guidance, Navigation and Control system (PGNC), a commercial powered parafoil- vehicle, called the Buckeye consisting of a parafoil and vehicle two-body system like the X-38/CRV was modified to accommodate the avionics and scale-downed parafoil for aerodynamic similarity and a series of flight tests were conducted. Dynamic modelling and system identification results for the Buckeye are de- scribed in this dissertation. The vehicle dynamics are modelled as all 8 degrees-of- freedom system comprising 6 states for the parafoil and 2 states for the relative pitch and yaw motion of the vehicle with respect to the parafoil. Modal analysis for the linearized model from the nonlinear model shows the number and order of dynamic modes as well as the system is controllable and observable. For system identifica- tion, the overparameterized Observer/Kalman Filter Identification (OKID) method is applied to identify a linear model of the Buckeye two-body system from the flight data assuming that disturbances at a calm day are represented as periodic distur- bances. The identification results show that the overparameterized OKID works well for powered parafoil-vehicle two-body system identification under calm day condi- tions using flight data. For the data with possible discrete gusts the OKID shows limitation to identify a linearized model properly. Several sensor packages including airdata and Inertial Measurement Unit are designed and installed for the parameters for identification. The sensor packages successfully supply data of the parameters for identification and suggest a feasible, low cost method for the parafoil-vehicle two-body dynamic parameters.

Thesis (M.Sc.Eng.) / Guided airdrop systems lacking propulsion may be adversely affected by high winds. Strong winds encountered during Draper Laboratory flight testing prevented lightweight parafoil systems from landing accurately. This thesis introduces and compares multiple guidance strategies designed to address high wind scenarios in cases of differing wind knowledge fidelity. The algorithms presented significantly improve performance in high tailwind and shifting wind scenarios without compromising miss accuracy in standard wind conditions. This adds additional capability to parafoil guidance by substantially increasing the conditions under which accurate landings are possible.

Airdrop systems provide a unique capability of delivering large payloads to undeveloped and inaccessible locations. Traditionally, these systems have been unguided, requiring large landing zones and drops from low altitude. The invention of the steerable, gliding, ram-air parafoil enabled the possibility of precision aerial payload delivery. In practice, the gliding ability of the ram-air parafoil 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. Research and development work on guided airdrop systems has focused primarily on evolutionary improvements to the guidance algorithm, while the navigation and control algorithms have changed little since the initial autnomous systems were developed. Furthermore, the control mechanisms have not changed since the invention of the ram-air canopy in the 1960's. The primary contributions of this dissertation are: 1) the development of a reliable and robust method to identify a flight dynamic model for a parafoil and payload aircraft using minimal sensor data; 2) the first demonstration in flight test of the ability to achieve large changes in glide slope over ground using coupled incidence angle variation and trailing edge brake deflection; 3) the first development of a control law to implement glide slope control on an autonomous system; 4) the first flight tests of autonomous landing with a glide slope control mechanism demonstrating an improvement in landing accuracy by a factor of 2 or more in especially windy conditions, and 5) the first demonstrations in both simulation and flight test of the ability to perform in-flight system identification to adapt the internal control mappings to flight data and provide dramatic improvements in landing accuracy when there is a significant discrepancy between the assumed and actual flight characteristics.