Development of a Miniature Self-Driving Car Using SoC and NPU

For the ICCE2025 design contest, we are developing a miniature self-driving car system using Raspberry Pi5 and Hailo-8L Neural Processing Unit (NPU). Previously, we had implemented the object detection part on AMD/Xilinx FPGA using DPU softcore IP. Now the hardcore NPU chip dedicated to neural network processing is available, we are implementing the object detection part using the NPU and building our system. We compare the performance of the NPU implementation with that of the previous FPGA implementation.