园艺电动旋耕机器人设计与试验

Design and Experiment of Autonomous Electric Rototiller Robot for Horticulture

摘要:

针对设施园艺旋耕作业存在的机具缺失、传统拖拉机牵引污染大、功耗高、运动灵活性差等问题,本文设计了一种新能源园艺纯电自走式旋耕机器人。基于作业场景与农艺需求,确定了整机结构方案,开发了隔离型DC-DC电气系统架构、模块化旋耕机组及电液提升系统。通过虚拟样机技术构建了考虑底盘俯仰角与旋耕机组倾角的耕深MAP图,并采用BP神经网络PID控制器调控耕深电液提升系统。为提高控制性能,提出了一种改进果蝇算法,融合自适应步长、高斯随机游走及种群浓度判定值优化策略,用于BP神经网络PID初始权值整定。结果表明,改进果蝇算法较传统方法减少9次迭代,目标函数值降低6.14%;在阶跃信号下基于该算法整定的BP神经网络PID控制器调节时间仅需0.31 s,无稳态误差与超调,且对正弦信号与动态斜坡信号具有快速响应特性。田间验证显示,机器人旋耕速度均值为2.81 km/h,平均偏移量为0.2112 m,偏移率为0.524%,最小转弯半径为458 mm,满足灵活行走转向要求;耕深50、100、150 mm工况下,稳定性系数分别为89.51%、88.47%、92.85%,土壤破碎率达70.51%、75.07%、89.14%,均优于碎土率大于65%、耕深稳定性系数大于85%的作业标准。研究结果为设施旋耕机械设计与精准控制提供理论依据。

Abstract:

In response to the issues of equipment shortages,high pollution,excessive power consumption,and poor maneuverability in traditional tractor-tillage operations in facility horticulture,an energy-driven,purely electric self-propelled tillage robot was designed. Based on operational scenarios and agronomic requirements,the overall structural scheme was determined,and an isolated DCDC electrical system architecture,modular tillage units,and an electro-hydraulic lifting system were developed. Using a virtual prototype simulation approach,a tillage depth MAP chart considering the chassis pitch angle and tiller inclination was constructed,and a BP neural network PID controller was employed to regulate the electro-hydraulic lifting system for tillage depth. To enhance control performance,an improved fruit fly optimization algorithm incorporating adaptive step size,Gaussian random walk,and population concentration judgment value optimization strategies was proposed for the initial weight tuning of the BP neural network PID. The results indicated that the improved fruit fly algorithm reduced the number of iterations by 9 and decreased the objective function value by 6.14% compared with that of traditional methods. The BP neural network PID controller,tuned based on this algorithm,achieved a regulation time of only 0.31 s under step signal conditions,with no steady-state error or overshoot,and exhibited rapid response characteristics to sinusoidal and dynamic ramp signals. Field validation demonstrated that the robot achieved an average tillage speed of 2.81 km/h,with an average deviation of 0.2112 m,a deviation rate of 0.524%,and a minimum turning radius of 458 mm,meeting the requirements for flexible maneuvering. Under tillage depth conditions of 50 mm,100 mm,and 150 mm,the stability coefficients were 89.51%,88.47%,and 92.85%,respectively,and the soil fragmentation rates reached 70.51%,75.07%,and 89.14%,all exceeding the operational standards of a soil fragmentation rate greater than 65% and tillage depth stability above 85%. The research findings can provide a theoretical basis for the design and precise control of facility tillage machinery.