Development of an FPGA based Object Tracking Robot: Tracking Algorithm for Object Tracking Robot

Chang, Andy Shin Weng (2019) Development of an FPGA based Object Tracking Robot: Tracking Algorithm for Object Tracking Robot. Final Year Project (Bachelor), Tunku Abdul Rahman University College.

Text Andy Chang Shin Weng_Full Text.pdf Restricted to Registered users only Download (2MB)

Abstract

This research focuses on implementing an object tracking system and collision avoidance system on a robot to track an object and at the same time avoid obstacles. It begins with receiving image data in the form of matrix and keep track of the image by initiating movements on the robot. At the same time, it will avoid obstacles by initiating movements on the robot as well. In the implementation, Pixel Tracking algorithm is introduced for object tracking. This algorithm tracks an object by comparing the number of pixels and area of frames to decide whether to move forwards and backwards or left and right respectively. As for the collision avoidance system, an input distance is received from the frontend to initiate appropriate movements to avoid obstacles. Before the hardware implementations, a HDL based test bench is created on each sub-block as a test driven development method for verification purposes in Modelsim of Mentor Graphics environment. The main focus of this research is low computation time. Therefore, the proposed architecture uses Altera Cyclone IV EP4CE115F29C7N FPGA device to run the systems. Outputs of the systems will be connected to the inputs of motor drivers to control the movements of the motor. The current design methodology of the system is less complex and uses logic elements of less than 3%. In addition, the power dissipation of this system is very low that is 326:52mW. The accuracy of this object tracking system and collision avoidance system is 35.714%.