An improved braitenberg policy for object tracking

Abstract

In vision-based unmanned vehicle tracking of moving vehicles, it is difficult to achieve accurate tracking of motorized vehicles with conventional fixed-parameter PID controllers due to their varying speeds and changes over time. To generate control signals for the systems and move the controlled object, the PID controller performs calculations based on percentage, integral, and derivative data. This is the discrete PID controller method. In this paper an improved tracking method, the Braitenberg policy, is proposed and a series of experiments are conducted in Simulink. The experimental results verify that the method can achieve stable tracking in simple environment for both general and extreme cases. The method can be used for future experiments for both general and extreme cases. The method can be used for future experiments in more scenarios. Braitenberg policy provides stronger and more stable tracking performance in extreme cases compared to conventional PID control. The algorithm provides more opportunities for tracking research and a small idea for the integration of multiple algorithms.