Develop Risk Awareness Detection System for Autonomous Vehicles

It is still the biggest challenge that autonomous vehicles move safely in complex city environments. Especially in the blocked streets, where blind spots are in areas and unforeseen obstacles, existing systems may remain insufficient. The researchers developed a new approach to solve this problem.

The main findings of the work performed are:

• ** A new detection method called risk-aware swelling** was recommended; this system shows potential hazards in the map form
• A dense risk map is used instead of traditional “representation” (car, pedestrian, strip lines)
• There is a goal to characterize unspecified zones and complex interactions openly and evenly
• The system focuses on increasing safety especially in crowded and closed areas
• En-to-end provides better integrated in planning process for driving systems

The available methods are difficult to combine the tools and obstacles separately, which is difficult to combine this information in the planning phase. Risk-aware drives the decision-making process by showing all hazards on a single map. Interestingly, this method can also process unstructured obstacles and uncertain areas. This can provide more reliable autonomous driving in complex urban environments.

The researchers show that autonomous vehicles can be able to make smarter and secure decisions thanks to this system. In particular, in case of emergency or unforeseen scenarios, it may be reduced.