Autonomous object detection and tracking robot using Kinect v2

Abstract

Object detection and tracking are one of the key features of a robust autonomous mobile robot, allowing it to navigate places and avoid obstacles. The Mobile robotics market and proliferation has been growing and the Covid-19 era has added another boost to this area where more and more interest is being drawn to the autonomous capabilities of these machines. In this paper we propose a hardware based model to detect and track objects based on color. We propose robust object detection and tracking with minimum environmental constraints to improve accuracy using our algorithm, and capable of behaving well in unknown environmental conditions. At the end of the analysis, the robot was able to detect the object and track it well. We also show frequency analysis, compression and error analysis of the underlying technique. Experimental outcomes verify improved accuracy of our algorithm.