An Active Safety System for Wheeled Mobile Driving Simulators

Wheeled mobile driving simulators (WMDS) represent an innovative concept in driving simulator technology that differs from conventional systems in its independence from a fixed infrastructure. Instead, WMDS move like mobile robots on wheels within a given open space, requiring novel approaches to establish safety. Previous safety concepts ensured continuous controllability of the WMDS, but required the intervention of operators to protect against collisions. The following work addresses the research question of whether the unbound movement of a WMDS can be actively safeguarded with a likewise mobile safety architecture. For this purpose, requirements are defined and based on this, the feasibility is practically examined. It is deduced that an active safety architecture for WMDS requires two further safety functions: A workspace compliance function that forces the WMDS to maintain its prescribed workspace during a driving simulation, and a collision protection function to actively protect the test person and people in the WMDS environment from collisions. Minimum functional requirements are derived in terms of required measurement quantities and decision logics. This results in a concept that monitors the presence and distance of objects within a speed-adaptive protection zone around the WMDS as well as the compliance with local, position-dependent speed limits, demanding reliable information of the WMDS position and speed in the entire workspace. So far, no sensory systems for those measurement quantities have been realized for an application for WMDS. Therefore, this work investigates hardware and software components that can reliably perform the intended functions within the operational design domain (ODD) of a WMDS. An approach based on lidar sensors is chosen to implement all required measurement variables with the addition of artificial workspace landmarks. The hardware and software requirements are concretized, selected sensors are implemented on a physical prototype and software algorithms are presented. Finally, the resulting safety system is evaluated in a representative environment. The design goals and the evaluation address the safety of the intended function under all conceivable operational conditions, fault detection capability and robustness against undesired interventions during WMDS operation. If the functions prove themselves under the most difficult conceivable operational conditions, they are considered suitable as a safety relevant function. The results of the work show that landmark-based position and velocity detection can fulfil the requirements of a safety-related function for workspace compliance. For object detection, the fulfilment of the target function can be shown, but only under ODD limitations of the WMDS. The general applicability of lidar sensors for the active safety system is thus not considered to be falsified with the results, but limited by further requirements, e.g. on the ground conditions of the workspace. The findings of this work provide requirements, test cases and promising approaches for an active safety system for WMDS that can be followed up and optimised in future work