Autonomous systems on the test bench for maximum safety

Driver assistance systems and autonomous vehicles are increasingly being used in road traffic. In the meantime, man and machine share the roads. For the safety of all road users, road signals or signs must be equally easy for both to recognise and interpret. The research company RISC Software and its partners ASFINAG and Johannes Kepler University Linz are creating a secure basis for this.

Vehicles should stop at “STOP” even if individual LED elements of a traffic light have failed or are defective. Humans are usually able to recognize this circumstance and react correctly. Autonomous systems may interpret the defective display incorrectly. This could result in accidents.

The research company RISC Software puts autonomous systems to the test and investigates to what extent disturbances in traffic signals lead to misinterpretations or misbehaviour. The findings from the analysis are to be used to make systems based on artificial intelligence even safer and more robust. The systems will be evaluated using real traffic signals and road signs and disturbances will also be simulated.

The aim of the research work is to ensure maximum security of autonomous systems and to increase confidence in artificial intelligence. RISC Software will evaluate the technical methods. In cooperation with the Johannes Kepler University, it will ensure that research and development is based on ethical principles and that the system ultimately implements ethical criteria. The partner ASFINAG brings in concrete requirements from Austrian road maintenance companies. The data obtained will be made available to Austrian companies in the fields of mobility, road infrastructure and autonomous driving. This is intended to create a strong basis for further developments – such as traffic signs that are equally readable for man and machine.

You can also find further information in our news section: https://risc-software.at/projektstart-safesign