Drone as an autonomous aerial sensor system for motion planning

Abstract

A system capable of both autonomous navigation and remote fire detection was developed from first principles. Using a complementary filter and attitude fundamentals, an Euler coordinate system representation of the aircraft’s orientation was measured using a widely available low-cost inertial measurement unit incorporating MEMS accelerometers, gyroscopes and magnetometers. Using line-of-sight guidance principles, navigation trajectories could be calculated in real-time providing autonomous navigation between user designated waypoints. A stabilisation control system using proportional-integral-derivative (PID) controllers was developed in order to achieve stabilised flight on the calculated navigation trajectories. Fire detection was facilitated with the use of three low-cost air composition sensors fed into an artificial neural network. A motion planner was developed to assist in flight planning by using topology information for the flight area to construct an optimal flight path based on distance and climb penalties.