跟随型智能小车边缘计算的边云协同系统设计

Design of Edge Cloud Collaborative System Based on Edge Calculation of Follow Intelligent Car

本文研究自动跟随行人的智能小车,由于智能小车自身的计算资源十分有限,因此本文采用了边云协同方法来解决该问题。本文在树莓派上安装摄像头组成一台树莓派小车,利用云端计算机结合目标检测算法对小车采集到的视频流进行实时处理,将计算结果回传给智能小车以指导其运动。实验结果表明,该边云协同系统可以实现自动跟随行人,且实时性好,稳定性强。本文所提出的方法解决了智能设备自身运算能力不足的问题,可以为未来智能设备的研发提供参考。

In the paper,the intelligent car is studied which can automatically follow pedestrians.Because the computing resources of intelligent car are very limited,this paper uses the method of edge cloud cooperation to solve the problem.Specifically,a raspberry pi car is composed of a camera installed on the raspberry pi.The cloud computer combined with the target detection algorithm is used to process the video stream collected by the car in real time,and the calculation results are sent back to the intelligent car to guide its movement.The experiment results show that the system can automatically follow pedestrians with good real-time performance and strong stability.The method proposed in this paper solves the problem of insufficient computing power of intelligent devices,and can provide reference for the future research and development of intelligent devices.