AI-Based Approaches for Autonomous Vehicle Emergency Handling and Response

Abstract

For all safety-critical systems, before deployment safety must be controlled in terms of both functional safety and cyber security hazards which generally comprises of numerous potential uncanny sources of unsafe, unwanted, and even ghost detections to name a few supplied unexpected inputs, counterfeited and collective byzantine assessments, overtly dangerous counter distinction, among others. For the potential real-time counter identification and management puzzlement – free gray box perception of an entire system is in inexistent. The key unaddressed impediment for trust implicit important fault-free particularly, mostly – autonomous regulators, is the evaluation of complex autonomy algorithms of this type which is beside the others, everything from the especially admissible list of potentials. Potential unusual features of offered examinations should include characteristics like the removal of some initialization phase observations and a conveyed collection of some environmental uncertainties influenced factors.