Identification of safety risks in mixed traffic concepts in industrial sites

Automated driving technologies are developing rapidly for off-road vehicles and industrial machinery. However, highly automated heavy industrial vehicles and work machines still need to be separated from manual machines and human workforce. Different operating environments and work processes require different solutions to ensure efficient and safe operation and interaction with autonomous or semi-autonomous machines. The traditional
system safety analysis methods like Preliminary Hazard Analysis (PHA) and Operating Hazard Analysis (OHA) have been successfully applied in different domains and different automated machinery applications. These methods continue to have an important meaning when they are effectively used in the system safety engineering. However, these methods do not cover or consider autonomy aspects especially safety critical interactions and their possible deviations between system controllers (human and technical). The new challenge is that machine’s
autonomous behaviour cannot be fully predetermined because the key element in machine autonomy is adaptability to dynamically changing environment based on the perception of the available information. There is need for new methods to be able to identify and analyse autonomy related uncertainties, hazards, and hazardous situations. The System-Theoretic Process Analysis method (STPA) brings in new views for the analysis of autonomy aspects by
supporting the identification and analysis of unsafe control actions in different hierarchy levels of the system control. STPA method includes the modelling of the hierarchical control structure of the machinery system and complements the perspectives of traditional system safety methods. STPA method provides a formal presentation to connect losses, system level hazards, possible unsafe control actions and loss scenarios so that this information
can be used for defining system safety requirements. This study is part of the EU and Business Finland funded research project ‘Artificial Intelligence using Quantum measured Information for real-time distributed systems at the edge’ (A-IQ Ready). The safety research in the project aims for improving safety and productivity of automated vehicle operations outdoors in co-existence with human workers in so called ‘mixed traffic’ operations. The main target of the safety research is to develop data fusion concepts and risk conscious situational awareness information for autonomous driving and load handling.