Implementation of system identification and flight control system for UAV

Abstract

This paper presents a methodology to implement a flight control system based on PID control design for PX4 autopilot system. The objective of the method is to find out the optimal controller gains on the same control structure of PX4 flight stack software without iterative controller tuning. This is achieved through a two-step procedure which consist of aircraft system identification and PID optimized control design. The first step to implement an autopilot system on an Unmanned Aerial Vehicle (UAV) relates to characterizing the UAV's dynamics using a mathematical model. To accomplish the accuracy of the particular UAV control, the process of system identification, which is the estimation of the parameters of the equation of motion, is essential. The measurement of inputs and outputs during manual flight is utilized to determine the unknown parameters of SISO mathematical model using a software of comprehensive identification from frequency responses. Subsequently, the model is utilized for an optimization-based tuning of these PID controller gains offline in order to minimize the requirement of numerous in-flight tuning. The controller is implemented on PX4 autopilot system. The results are demonstrated that the tracking control system has excellent dynamic performance in respect of simple design, high precision, and easy implement.