Reconfigurable HW/SW Architecture of a Real-Time Driver Assistance System

Driver assistance systems significantly increase the driving comfort and can prevent accidents. On the other side, they require high performance and computations need to adopt to the environment to reach their goals. Reconfigurable architectures offer the requested flexibility and performance, if care is taken to partition the tasks in hardware and software parts. Elsewise, real-time requirements can't be fulfilled. In this paper, the implementation and a hardware/software partitioning of a driver assistance system for a reconfigurable architecture are presented. The assistance system detects taillights of ahead moving vehicles in dark environments or in tunnels and visualizes the results. A detailed description of the implementation on a reconfigurable platform is given. Furthermore, the experimental results demonstrate the effectiveness of the approach.