Development of a test bed for real-time guidance and control of an autonomous vehicle

Abstract

An operational prototype of a ground vehicle that may be used as a test bed for future experiments in autonomous guidance, navigation, and control is developed. Key topics in the research include vehicle dynamics and modeling, inertial navigation, hardware/software integration, and system identification. The vehicle dynamics are derived from a simplified 3 degree-of-freedom (2 translational axes and 1 rotational axis) model, and inertial navigation equations are developed for the 2-dimensional, planar case. Descriptions of the hardware subsystems, which include the vehicle, onboard computer, sensors, actuators, and communications radio are presented. The execution of the real-time, onboard software used to perform control input/output as well as the ground station software application is outlined. Finally, the methodology and results of system identification experiments used to estimate the servo actuators and vehicle modeling parameters are given. Field testing demonstrated successful execution of the onboard and ground station software applications as well as the sensors and actuators subsystems. The vehicle was operated open-loop by uploading a sequence of steering and throttle commands from the ground station, and the sensor measurement data was recorded. The recorded data was used to estimate modeling parameters of the servo actuator mechanisms and the vehicle. The servo actuator models obtained from identification resulted in good matching between the simulated and physical responses with little deviation. The estimated vehicle parameters obtained from identification also resulted in good matching between the simulated and physical responses. Deviations between the simulated and physical responses are likely due to measurement errors present in the calculation of the vehicle states. The results of this research provide a proof-of-concept and demonstrate the viability of constructing an operational platform for vehicle autonomy.