An Integrative Approach for Enhancing Pedestrian Prediction in Self Driving Cars

Safe walkways and efficient driving are the goals of Pedestrian Intention Detection (PID), which is key for Advanced Driver Assistance Systems (ADAS) and self-driving cars. In this study, a comprehensive dive into PID methodologies is done, focusing on the latest computer vision techniques and machine learning algorithms. The study proves that the integration of additional features like pose estimation and drivable area detection enhances the decision-making capabilities of PID systems. By deeply testing and evaluating the proposed PID system on benchmark datasets, like JAAD, it was shown that it works well with an accuracy of 79.53 %. Furthermore, some possible future directions for this PID system will be mentioned to conclude the paper. The outcomes of this research are anticipated to pave the way for the integration of state-of-the-art PID systems into ADAS platforms and, therefore, provide safer and more efficient transportation systems for future generations.