针对在无人运输领域中,AGV小车在工位对接时停车位置不够精确的问题,本文利用激光雷达和深度相机传感器并通过相应的坐标变换对障碍物的距离进行测量,将传感器得到的数据通过选择恰当的传递函数和学习算法利用RBF神经网络进行训练。实验结果表明,用该方法得到的结果测距误差低于0.1%,速度上相比于传统的BP算法提高了24%,可用来辅助AGV小车停车时进行重定位,同时通过融合得到的数据在以后建图方面也具有一定的优势。
Abstract
In the field of unmanned transportation, the parking position of AGV trolley is not accurate enough when docking at the workstation.This paper uses lidar and depth camera sensors, and through the corresponding coordinate transformation, measures the distance of obstacles and obtains the sensor.Through the selection of the appropriate transfer function and learning algorithm, the RBF neural network is used for training, and the simulation verification results show that the range error of the result obtained by this method is less than 0.1%, and the speed is compared with the traditional BP algorithm.It is increased by 24%, which can be used to assist the relocation of the AGV trolley when it is parked.At the same time, the data obtained through fusion also has certain advantages in future mapping.
关键词
无人运输 /
RBF神经网络 /
激光雷达 /
视觉相机 /
数据融合
Key words
unmanned transportation /
RBF neural network /
lidar /
vision camera /
data fusion
{{custom_sec.title}}
{{custom_sec.title}}
{{custom_sec.content}}
参考文献
[1] 战艺.移动机器人SLAM 与路径规划方法研究[D].桂林:广西大学, 2016.
[2] LEE D H, CHANG J, YHEO Y S.Stereo matching using cost volume fusion for high dynamic range scenes[J].Electronics Letters, 2017, 53(23):15221524.
[3] ZHAO X, KAN J.A trunk ranging system based on binocular stereo vision[C]//The 9th International Conference on Digital Image Processing (ICDIP 2017) International Society for Optics and Photonics, 2017:10431048.
[4] 刘坤, 翟敬梅.室内阴影环境下障碍物检测与测量[J].计算机工程与设计, 2017(10):27992803.
[5] 卢艳军, 王诗宇, 张太宁, 等.针对目标跟踪过程的双目视觉测距方法研究[J].电光与控制, 2020, 27(1) :6063.
[6] 李大华, 沈洪宇, 于晓, 等.一种改进Census变换的双目匹配测距方法[J].激光与光电子学进展, 2019, 56(11):189194.
[7] SHANG E, AN X, WU T, et al.Li DAR based negativeobstacledetectionfor field autonomous land vehicles[J].Journal of Field Robotics, 2016, 33(5):591617.
[8] 刘家银, 唐振民, 王安东, 等基于多激光雷达与组合特征的非结构化环境负障碍物检测[J].机器人, 2017, 39(5):638651.
[9] 孙贝贝.基于异质传感器信息融合的移动机器人避障研究[D].太原:中北大学, 2018.
[10] 李蔚东, 谢再晋, 李敏.基于多传感器融合的扫地机器人算法研究[J].电子设计工程, 2021, 29(2):118122, 128.
[11] FETIC A, JURIC D, OSMANKOVIC D.The procedure of a camera calibration using camera calibration toolbox for Matlab[C]//MIPRO Proceedings of the 35th International Convention IEEE, 2012:17521757.
[12] 周鹏程.基于机器视觉的自主式救援机器人的研究[D].南京:东南大学, 2016.
[13] 王中立, 牛颖.基于多传感器信息融合的机器人障碍物检测[J].中国测试, 2017, 43(8): 8085.
[14] 黄琴.基于贝叶斯估计与分布的关联分析方法研究[D].广州:华南理工大学, 2016.
[15] 张春英, 冯晓泽, 刘凤春, 等.基于DS证据理论的三支决策推理模型[J].模糊系统与数学, 2020, 34(6).
[16] Liu Q, Wang X, Rao N S V.Artificial neural netwoksfor estimation and fusion in longhaul sensor networks[C]//International Conference on Information Fusion, 2015:460467.
基金
*国家自然科学基金项目(61701244);江苏省产业前瞻与关键核心技术重点项目(BE20200062)。
PDF(1172 KB)
