智能驾驶中基于FPGA的高效目标检测方案

打开文本图片集

中图分类号：TN952 文献标识码：A 文章编号：2096-4706（2026）06-0007-05

Efficient FPGA-based Object Detection Solution for Autonomous Driving

LIU Yuzhou （Portland Institute，Nanjing UniversityofPostsand Telecommunications，Nanjing21oo23，China）

Abstract： With the rapid development of intelligent driving and unmanned systems，environmental perception imposes higher requirementsonthe acuracyandreal-timeperformanceofradarsensors.Thispaper presentsan effcient millimeterwave radar target detection system based on an FPGA-based hardware-software co-design approach.The system uses the PZ7020as thecoreprocesing unit，combinedwithLD2450andLD2451milimeter-waveradars，andachieves high-speedmultitarget informationtransmissionviaUART，andperforsdataacquisition，buffering，fltering，andparsingontheFGA.Onthe software side，adata frame parsing program basedoninterupt and buffer mechanisms is designed toensurethecompleteness andreal-timeperformanceofdatareception.Experimentalresultsshowthatthesystemcanstablyoutputarget distance，velocity and angle information， with an accuracy exceeding 86% and error ranges within acceptable limits for engineering applications， demonstrating the feasibility and practicality of the proposed solution.

Keywords： FPGA acceleration; millimeter-wave radar; Object Detection; embedded data parsing

0 引言

随着无人机及汽车智能辅助驾驶需求的提高，环境感知技术成为智能系统的关键。（剩余6894字）