SAFE4CAR
System to guarantee minimum risk conditions in the event of failures in autonomous vehicles:
Perception for situational awareness and environment simulation
PID2022-140554OB-C32
Cooperative, connected and autonomous driving is currently an important technological advance that will influence and change our mobility and quality of life in the future. Although it has had an important development in recent years, there are still limitations to reach the
highest levels of automation.
Specifically, this project focuses on the conditions that can cause an autonomous mode disconnection and how the vehicle should react.
According to the levels of automation, in systems below Level 3, the response (failure backup) in the event of conditions that were not contemplated or when failures occur are performed by human drivers. However, in the case of levels 4 and 5, the fallback functions can no
longer be addressed by human drivers, but by the autonomous system itself.
In this sense, the main objective of the coordinated project is to identify and analyze failure situations, erroneous or potentially erroneous behavior, or conditions in which the autonomous vehicle cannot drive, and to design safe fallback strategies to guarantee actions with
minimal risk in those circumstances.



The subproject ” System to guarantee minimum risk conditions in the event of failures in autonomous vehicles: Perception for situational awareness and environment simulation” focuses on the incorporation in the SAFE4CAR system of perception around the vehicle 360º to
detect and identify possible escape routes in a fallback situation. Sensor failures in sensory fusion due to degradation or partial occlusion of the elements involved in the creation of the knowledge of the environment will be analyzed and simulated. In addition, an advanced simulator will be used to allow the integration of all these sensors and the
analysis of the optimal distribution to determine the perception elements along the autonomous driving and the areas of posible circulation in the face of these failures.
In this way, in cooperation with the other subprojects, the aim is to overcome the limitation of manufacturers, who only contemplate braking as a fallback response, but do not extend the possibilities to lane changes or degraded modes. They have also not been able to take
advantage of all the possibilities that a complete situational awareness, derived from the information of the perception and communication systems, can provide.
Additionally, the actions of the vehicle and the fallback measures are analyzed from the point of view of the human factor, to verify their naturalistic behavior, and the degree of acceptance and perceived risk. This assessment is carried out on automated vehicles with a configurable control unit in which the SAFE4CAR system is incorporated in real driving scenarios that exemplify the situations that entail possible problems for the perception and decision-making systems.