Sensor Data Fusion and Submerged Test Results of a Pectoral Fin Propelled Uuv

An inertial navigation system (INS) complete with extended Kalman filters (EKFs) to mitigate sensor measurement errors is described in this paper for a six-degree-of-freedom (6-DOF) pectoral fin propelled unmanned underwater vehicle (UUV). Simulation results are discussed and the vehicle model is validated with initial experimental data. After modeling the sensor performance characteristics, EKFs are implemented to fuse imperfect gyroscope and accelerometer data from the INS and eliminate the integration errors caused by measurement biases of these sensors. Vehicle responses to standard maneuvers are simulated including straight line motion, constant radius turn, and climbing. In addition to simulating the vehicle with sensor models in the loop, an updated set of control laws is shown to be robust with respect to variations in actuator performance. The control architecture designed and tested in simulation, which utilizes combinations of fin kinematics sets as well as fin rotation biases to generate desired control force vectors, is implemented in the vehicle onboard microcontroller. This paper addresses initial results of our submerged test vehicle utilizing two updated control techniques. Evaluation of results reveals agreement between experimental and simulated results in vehicle speed and maneuverability which serves to validate our model.