Experimental platform for autonomous collision avoidance strategies
Many car manufacturers are now working on fully automated vehicles, able to offer a more pleasant driving experience but also able to improve the traffic safety. Today, driver behaviour is a contributing factor in over 90 percent of all accidents. Hence a huge step to improving traffic safety is the implementation of proactive safety systems to avoid collisions between vehicles. The technology for implementing such a system is available today, but to be able to implement it, rigorous and expensive tests are needed. This project aims to propose a solution to this problem, based on a hardware in-the-loop platform.  
Aim of the project
The objective of this project is to develop a platform, containing a simulator and a test vehicle as hardware in the loop, that can be used to verify collision avoidance strategies.
A simulator named CAS (Collision Avoidance Simulator) has been developped and is used in  a experimental platform using a real test vehicle as hardware in the loop, for verifying the functionality of collision avoidance strategies. In this platform, the real and virtual vehicles are able to detect each other and hence react as if a real collision is about to occur. In this way, collision avoidance strategies can be tested both in a cheaper and more efficient way. The platform contains a simulator and a test vehicle as hardware in the loop. With this platform, we are able to make the test vehicle and the virtual vehicles detect each other and apply a decentralized strategy for collision avoidance by only controlling the longitudinal movement of the vehicles. We consider a three or four way intersection on flat ground and vehicles interacting with the platform are assumed to be fully automated with predefined paths. 
Xiao Diyin, Fredrik Granum, Johan Gustafsson, Adam Fjeldsted
Gabriel R. Campos and Robert Hult


   Download report    


   Demonstration video


[ Click Image To Enlarge ]