摘要
为了能够及时准确地定位病虫害发生的位置,最大程度减少农户经济损失,农业大棚巡逻机器人的定位研究非常必要。但机器人使用的传统里程计常因为机械结构、误差积累、功耗发热等问题导致定位失效。为此,在已知农业大棚地图的基础上,将BP神经网络应用于巡逻机器人定位,提出采集机器人实际的位置信息和雷达信息,利用仿真软件进行模型训练,选择能够满足实际需要的激活函数,从而确定BP神经网络。通过实验并对比其他方法,证明使用本方法不但能实现精准定位,而且也能够提高机器人续航能力,有效提升农业大棚的生产效率。
In order to timely and accurately locate the location of pests and diseases and minimize economic losses for farmers,it is necessary to conduct research on the positioning of agricultural greenhouse patrol robots.However,traditional odometers used by the robots often fail to locate due to issues such as mechanical structure,error accumulation,power consumption,and heat generation.Therefore,on the basis of known map of agricultural greenhouse,BP neural network is applied to patrol robot positioning.Actual position information and laser information of the robot are collected,model training is carried out on simulation software and activation functions that can meet practical needs are chosen,and the BP neural network that can be used for localization is determined.Results from experiments and comparison with other methods verifies that the proposed method can not only achieve precise positioning,but also improve the endurance of robots and effectively enhance the production efficiency of agricultural greenhouses.
作者
董立国
侯继红
黎浩阳
周宇航
邓子杰
Dong Liguo;Hou Jihong;Li Haoyang;Zhou Yuhang;Deng Zijie(Department of Intelligent Manufacturing,Guangzhou Vocational College of Technology&Business,Guangzhou 511442,China)
出处
《机电工程技术》
2024年第7期56-59,63,共5页
Mechanical & Electrical Engineering Technology
基金
2022年度广州科技贸易职业学院校级科研项目(2022YB08)
2021年度广州市科技计划项目(202102080606)
2023年度广州市高等教育教学质量与教学改革工程现代产业学院项目(2023XDCY008)
2022年度广州市高等教育教学质量与教学改革工程职业院校教师教学创新团队项目(2022JSJXCXTD028)。
关键词
BP神经网络
农业大棚
巡逻机器人
定位
里程计
BP neural network
agricultural greenhouses
patrol robot
location
odometer