Self-Driving Car Application of a Stream-Oriented Accelerator Framework
暂无分享,去创建一个
In our self-driving car design for FPT'19 design competition, a stream-oriented acceleration framework is ported to Zynq SoC to enable easy co-design of software and hardware. The framework, OpenFC is primary intended to build a multi-FPGA acceleration cluster with HLS programmability easily but also powerful for embedded acceleration with multiple acceleration modules for image recognition. With the OpenFC framework, programmers can send and receive data streams between the microprocessor and accelerator modules called SPE (streaming processing element.) This paper describes briefly about our self-driving car design with the framework.
[1] Yasunori Osana,et al. Performance Evaluation of a CPU-FPGA Hybrid Cluster Platform Prototype , 2017, HEART.