摘要
随着人工智能的快速发展,机器人领域的一些先进技术的发展和进步已经对许多工业生产和社会发展做出了巨大的贡献。移动机器人的自主性是一个关键问题,一个完全自主的移动机器人必须具备对环境信息的认知能力以及遇到障碍物时的避障能力,因此,多目标识别就变得非常关键。论文借鉴先进的深度学习研究成果,优化并应用于ROS平台的移动智能体。以YOLO网络结构为基础,结合移动智能体的处理平台以及移动过程中实时性的要求,对网络模型进行改进优化。改进后的网络在确保精确度的前提下显著提高处理帧率,满足移动智能体的实时性要求。
With the rapid development of artificial intelligence,the development and progress of some advanced technologies in the field of robotics have made tremendous contributions to many industrial production and social development.The autonomy of mobile robot is a key problem.A fully autonomous mobile robot must have the ability to recognize the environmental information and avoid obstacles when encountering obstacles.Therefore,Multi-object detection becomes very important.This paper uses advanced deep learning research results to optimize and apply mobile agents in ROS platform.Based on YOLO network structure,combined with the processing platform of mobile agent and the real-time requirement of mobile process,the network model is improved and optimized.The improved network can significantly improve the frame rate and meet the real-time requirement of mobile agent under the premise of ensuring accuracy.
作者
陈浩
刘镇
CHEN Hao;LIU Zhen(School of Computer,Jiangsu University of Science and Technology,Zhenjiang 212000)
出处
《计算机与数字工程》
2020年第5期1108-1113,共6页
Computer & Digital Engineering
