As Autonomous Driving (AD) technologies roll out from development to deployment, vendors are demanding production-ready embedded solutions. Arm IP becomes a natural choice, combining high performance general-purpose compute with custom workload-specific acceleration in a reduced power envelope. However, porting existing AD software stacks to another platform is not without pitfalls. In this presentation, we discuss the challenges faced – with emphasis on hardware acceleration of neural networks – porting the open-source Autoware AD stack to an Arm-based system.
LIYOU ZHOUView website
Senior Software Engineer - Automotive (Arm Ltd)
Liyou is a software engineer working on Accelerating open-source self-driving software stack with Arm hardware and software IP.
Liyou has previously worked in IoT developing end to end secure firmware update solution.
Liyou is also active in the developer community, presented at mbed Connect and organised hackathons at the University of Cambridge.