Mechanism and neural network based on PID control of quadcopter

This paper describes mechanism the quadcopter with on-board sensors. This quadcopter consists of four rotors, four straight legs, and a disk-shaped body. The body is mainly composed of a lightweight, very rigid carbon fiber reinforced polymer (CFRP) and resin composite. A 9-axis inertial measurement unit (IMU) that contains accelerometer and gyroscope and magnetometer is installed in the body to measure rotation angles and angular velocities. In addition, a robust control method based on a neural network-based PID control method capable of dealing with payload and wind disturbances is proposed. In the control method, the gains of the PID controller are adjusted in real-time. Through several hovering experiments of the quad copter, the effectiveness of the mechanism and the control method is verified.