The development of a vision-orientated unmanned aerial vehicle platform based on the networked control system

Abstract

Unmanned Aerial Vehicles (UAVs) are rapidly becoming popular for many missions that are dangerous to pilots or require longer flight durations than are humanly possible. In this thesis, a vision-based multi-rotor UAV control platform based on the networked control system was developed. An on-board IP camera was used as the main sensor for motion estimation purpose. Position and velocity estimations in three directions of the earth frame were achieved from the camera feedback combined with the pose information from the gyroscope and acceleration sensor on the aircraft. Sensor feedback signals including camera, gyroscope, acceleration sensor and ultrasonic sensor, were sent through the wireless network to the ground station which is a powerful computing center. Computations of the optimal controller and video processing algorithm were completed in the ground station. Since the UAV control system was based on networked control, the delay compensation algorithm was used in the control system. Therefore, the processor’s computing capability is crucial to a UAV control system’s performances. In comparison to traditional solutions with a powerful micro-processor integrated on the aircraft body, the networked control system is able to connect the aircraft with an external processor to enhance the computing efficiency and reduce the whole weight of the aircraft body. The networked control system is a low-cost solution to some industrial UAV applications. The hardware and software architectures were detailed in this thesis, including implementations of the hardware platform, video processing algorithms and control strategies.

Date of Award	2013
Original language	English
Awarding Institution	The Hong Kong University of Science and Technology