摘要
为提升苹果采摘机器人在复杂农业环境下的采摘成功率,结合实际工作环境需求,设计了一种苹果采摘机器人系统。首先,采用激光扫描和机器视觉得到苹果果实和障碍物的位置信息,实现目标物苹果和障碍物的快速识别和精准定位;然后,通过区间二型模糊神经网络的模糊推理自适应能力和泛函逼近特性,并结合Q学习提出了多目标自主导航方法,完成了机器人系统的导航避障设计;最后,分别在仿真平台和实际应用中验证了算法的可行性。研究结果表明,所提方法对各种场景都有很好的泛化性能,在初始位置变化的情况下仍能保持良好的导航能力,在相同距离和相同路况下收敛时间总是最短;苹果采摘机器人能很好地完成避障导航工作,40次训练周期内,在226步后到达目标地点,用时为26.25 s。在实际应用中,完成一次采摘所需的平均时间为8.23 s,识别精度可达93.65%,实现了苹果采摘过程的高精度导航和智能控制,可为提升采摘效率、减轻人力负担和促进智慧农业发展提供有力支持。
In order to improve the apple picking success rate of the apple picking robot in the complex agricultural environment,the system composition of the apple picking robot was proposed based on the actual working environment requirements.Laser scanning and machine vision were used to obtain the location information of apple fruits and obstacles,so as to realize the rapid identification and accurate positioning of target apples and obstacles.Then,based on the adaptive ability of fuzzy reasoning and functional approximation characteristics of interval type-2 fuzzy neural network,and combined with Q learning,a multi-objective autonomous navigation method was proposed to complete the navigation obstacle avoidance design of the robot system.Finally,the feasibility of the algorithm was verified in the simulation platform and practical application respectively.The results showed that the proposed method had good generalization performance for all kinds of scenarios,can maintain good navigation performance under the condition of initial position change,and the convergence time was always the shortest under the same distance and the same route.The apple picking robot can complete the obstacle avoidance navigation work very well.In 40 training cycles,the apple picking robot reached the target location after 226 steps,and the time spent was 26.25 s.In practical application,the average time required to complete a picking was 8.23 s,and the recognition accuracy can reach 93.65%.The research successfully realized the high-precision navigation and intelligent control of the apple picking process,which provided strong support for improving the picking efficiency,reducing the labor burden and promoting the development of smart agriculture.
作者
张志伟
温耀龙
王晓晓
Zhang Zhiwei;Wen Yaolong;Wang Xiaoxiao(School ofDesign and Engineering,Chongqing College of Humanities,Science&Technology,Chongqing 401524,China;School of Fine Arts and Design,Chongqing University of Arts and Sciences,Chongqing 402160,China;School of Architecture and Civil Engineering,Chongqing Metropolitan College of Science and Technology,Chongqing 402167,China)
出处
《农机化研究》
2026年第7期143-149,177,共8页
Journal of Agricultural Mechanization Research
基金
重庆市教委科学研究规划项目(23SKGH305)。
