基于深度学习的烟厂制丝车间智能巡检机器人自主导航系统设计

Design of Autonomous Navigation System for an Intelligent Inspection Robot in the Tobacco Factory's Silk Production Workshop Based on Deep Learning

烟厂制丝车间通常存在各种设备、管道等复杂的环境条件,智能巡检机器人难以准确辨别和识别这些复杂的场景,导致避障精度下降;为了提高智能巡检机器人的避障精度,提出了基于深度学习的烟厂制丝车间智能巡检机器人自主导航系统设计;通过电源模块的设计为系统提供电能,结合射频模块和基带信号处理模块的设计,完成巡检机器人导航系统的硬件设计;在系统的软件设计中,根据导航路径标记图像的角点在视觉差异、距离和颜色等维度的特性,识别智能巡检机器人导航路径标记,通过计算发生碰撞的预警距离,得到碰撞概率的估计值,利用深度学习的卷积神经网络模型,规划巡检机器人避撞路径,实现智能巡检机器人的自主导航;测试结果表明,文中系统可以使巡检机器人成功绕过障碍物,将避障精度提高到95.5%,为提高车间巡检效率和安全性提供了一种新的解决方案。

There are usually complex environmental conditions such as various equipment and pipelines in a tobacco factory's tobacco production workshop,it is difficult for an intelligent inspection robot to accurately identify and recognize these complex scenes,resulting in a decrease in obstacle avoidance accuracy.In order to improve the obstacle avoidance accuracy of the intelligent inspection robot,a deep learning based autonomous navigation system design for the intelligent inspection robot in the tobacco factory's tobacco production workshop is proposed.The power module provides a power to the system and the RF module is combined with the baseband signal processing module to complete the hardware design of the inspection robot navigation system.In the software design of the system,according to the characteristics of the corners of the navigation path marking image in the dimensions of visual difference,distance and color,the navigation path marking of the intelligent patrol robot is recognized.By calculating the warning distance of collision,the estimated value of the collision probability is obtained.The convolutional neural network model based on deep learning is used to plan the collision avoidance path of the patrol robot,and achieve the autonomous navigation of the intelligent patrol robot.Experimental results show that this system can enable inspection robots to successfully bypass obstacles,improving the obstacle avoidance accuracy to 95.5%.It provides a new solution for improving the efficiency and safety of workshop inspection.