Statistical Safeguards: Redefining Col lision Avoidance with Probability Theory: Employing Statistical Decision-Making to Enhance Safety in Mixed Traffic Environments

Abstract

This thesis introduces a probabilistic collision avoidance system that employs statistical decision-making in order to enhance the safety of mixed traffic environments. Central to this approach is the representation of vehicle positions as normal probability distributions, which are convolved with real-time sensor data to assess risks more accurately and reduce the unnecessary emergency stops. The research develops and implements a dynamic collision probability threshold, that is derived from safety integrity levels (SIL), which is imperative for complying with the rigorous safety standards and regulations. Simulations and analytical methods were used to validate the effectiveness of the proposed algorithm and demonstrating its potential in decision-making in emergency situations. Thus a a scalable solution for collision avoidance is presented in the form of an algorithm that can be integrated into existing safety systems, in order to enhance the operational efficiency for mixed traffic environments.