Advanced Driver Assistance Systems (ADAS) provide a significant contribution to increasing automotive safety. ADAS systems provide the driver with increased situational awareness, helping to reduce collision and accidents. To provide the driver with increased situation awareness ADAS systems can be categorized as providing external or internal awareness. External ADAS systems monitor such aspects as blind spots and lane detection, while internal systems monitor the occupants and particularly the driver themselves such as Driver Drowsiness Detection.
Both internal and external ADAS systems rely heavily upon embedded vision systems, implementing these embedded vision systems depending upon the task at hand can be computationally intensive. This computational complexity can reduce the performance of the system introducing latency and reducing the validity of the information provided to the driver. The use of hardware programmable logic enables the implementation of a low latency high performance system. However, industry standard development techniques such as the use of OpenCV cannot be used due to high development cost and timescales.
This webinar will demonstrate how an ADAS driver drowsiness detection application can be implemented using a Zynq heterogeneous SoC which combines programmable logic with high performance ARM cores. This example will demonstrate how a System Optimizing Compiler can be used in conjunction with the Zynq to create the ADAS application using high level languages and industry standard frameworks. The use of the System Optimizing Compiler enables seamless acceleration of C functions into the programmable logic, enabling a significant performance increase.
- What is ADAS & Driver Drowsiness detection?
- Implementation challenges
- Architecture of the solution
- Zynq architecture & Benefit of TySOM for prototyping ADAS solution
- Xilinx SDSoC and TySOM – platform creation and acceleration
- Live demo – unaccelerated and accelerated solution – demonstrate the difference