Towards an Advanced Self-Monitoring Tracking Module: Leveraging Statistical Hypothesis Tests and Subjective Logic Reasoning

In automated driving systems, monitoring and self-assessment of tracking algorithms is essential. This is especially necessary to meet today's safety and robustness challenges in an automated system. We propose a hybrid approach to develop a self-monitoring module for tracking algorithms. It makes use of well-known statistical hypothesis testing techniques. The results of which are fed into a subjective logic-based reasoning framework to produce robust and reliable self-assessment scores. Hence, we investigate the potential of combining these two approaches for monitoring and self- assessment systems and show the significance of this approach in experimental results.