Hovercrafts present unique control challenges; they are non-linear under-actuated systems. As a result they are difficult to operate manually. The desire was to implement a linearized control algorithm using inertial data and a GPS on a microprocessor. A real-time program loop autonomously controlled the hovercraft and allowed it to navigate between preset waypoints.
A hovercraft platform was constructed from a modified kit. A microprocessor with integrated IMU was used in conjunction with serial radio modules for teleoperation. A theoretical model was developed examining forces and moments acting on the platform. This model was then linearized and used as a basis to develop a control algorithm. The control algorithm allows input of waypoints and using data from the IMU autonomously navigates from point to point. A first pass at optimization of PID parameters was also performed.
A battery of tests were performed to examine the accuracy of the platform in following the waypoints, control stability, and robust error handling in response to unknown external perturbations.
The hovercraft uses an Inertial Measurement Unit including GPS, 3-axis accelerometer, 3-axis gyro, and 3-axis tilt-compensated magnetometer. An ATMEGA2560 microprocessor running custom firmware handles control, capable of real-time duplex telemetry/data with a ground-station via a 950 MHz serial radio connection . Waypoints are remotely set defining the desired path, the processor/IMU then employ a closed feedback linearized control algorithm that dynamically operates both throttle and rudder.
Project Report: Control Theory Implementation on RC Hovercraft
PowerPoint Presentation: Control Theory Implementation on RC Hovercraft