Improved Mobile Robot Manoeuvring Using Bayes Filter Algorithm Within the Planned Path

Abstract

This research introduces a novel approach for object tracking, capitalizing on the Bayes filter algorithm within the constraints of a single-camera setup. Object tracking is a pivotal aspect of computer vision, significantly influencing system performance in diverse applications. The integration of the Bayes filter algorithm provides a probabilistic framework, effectively addressing challenges posed by occlusions, lighting variations, and unpredictable object movements in real-world scenarios. Our methodology not only streamlines the tracking setup by utilizing a single camera but also enhances practicality, making it particularly relevant for applications with resource constraints. The paper offers a comprehensive exploration of this approach, delving into the theoretical foundations and technical intricacies that underlie the fusion of advanced object tracking techniques with the Bayes filter algorithm. Through empirical evaluations across varied tracking scenarios, our approach demonstrates superior effectiveness compared to traditional methods, showcasing the algorithm's ingenuity in improving tracking accuracy and adaptability. It achieved a dynamic simulation efficiency of 97.025%, a sensitivity of 96.2616%, and an overall system quality (F-score) of 97.0493%. This research contributes valuable insights to the evolving landscape of object tracking methodologies, presenting a practical and efficient solution that combines the Bayes filter algorithm's power with the simplicity of a single camera setup. The findings presented herein offer a nuanced perspective for researchers and practitioners seeking to elevate the precision and real-time adaptability of object tracking systems in diverse applications.