差速无人喷药车的自主导航及路径规划的仿真研究  

Simulation Research on Autonomous Navigation and Path Planning of Differential Unmanned Spray Vehicle

作　　者：张梓峭 冯卫君 马志凯[1] 范晓飞 Zhang Ziqiao;Feng Weijun;Ma Zhikai;Fan Xiaofei(College of Mechanical and Electrical Engineering,Hebei Agricultural University,Baoding 071000,China;Beijing Keyin Jingcheng Technology Co.,Ltd.,Beijing 100080,China)

机构地区：[1]河北农业大学机电工程学院,河北保定071000 [2]北京科银京成技术有限公司,北京100080

出　　处：《农机化研究》2025年第8期279-285,共7页Journal of Agricultural Mechanization Research

基　　金：河北省重点研发计划项目(2232720D);河北省省属高等学校基本科研业务费研究项目(KY2022018)。

摘　　要：随着农业装备现代化的发展,效率较低的人工喷药方式已不适用于大规模生产的需要。为此,针对机械喷药方式代替人工喷药方式的方案,提出了使用Pixhwark自驾仪作为无人喷药车的控制器,探讨了PX4开源自动驾驶软件对差速无人喷药车的控制,以及与ROS的Navigation导航功能包结合的适用性。实验在Ubuntu 20.04系统搭建仿真环境,建造了由PX4固件控制并搭载激光雷达的差速车模型;在进行自主任务巡航时,通过比较自身位置和目标位置来计算无人车的期望角度、加速度,引入L1算法比较自身位置和目标航线的横向偏差,以计算修正角度。路径规划实验中,在Gazebo仿真软件中搭建周围环境模型,通过激光雷达数据对周围环境进行SLAM建图。通过Navigation的路径规划交互接口move_base将规划的路径以线速度和角速度方式通过MAVROS通信机制发送到PX4端,实现了ROS的路径规划接口move_base与PX4端的通信,并进行了仿真系统验证。实验发现:PX4固件无人车的控制中,在摩擦力等因素的影响下,无人车不能达到期望的角度。通过PX4的路径规划实验和外部控制程序分析,发现PX4对于无人车的外部控制上存在转向角度开环处理的弊端,导致PX4的外部控制不能按照规划的期望角度控制无人车。With the development of modern agricultural equipment,manual spraying methods have proven to be inefficient and unsuitable for large-scale production needs.In order to replace manual spraying with mechanical spraying,proposed the use of the Pixhawk autopilot as the controller for an unmanned spraying vehicle.It explored the applicability of the PX4 open-source autopilot software for controlling a differential drive unmanned spraying vehicle and its integration with the navigation package in ROS.The simulation environment was set up on Ubuntu 20.04 system.A differential drive vehicle model controlled by the PX4 firmware and equipped with a laser range finder was constructed.During autonomous mission cruising,the desired angle and acceleration of the unmanned vehicle were calculated by comparing its current position with the target position.The L1 algorithm was introduced to calculate the corrective angle by comparing the lateral deviation between the current position and the target path.For the path planning experiment,the surrounding environment model was built in the Gazebo simulation software,and the laser range finder data was used for SLAM mapping of the environment.The planned path was sent to the PX4 side through the MAVROS communication mechanism in the form of linear velocity and angular velocity using the navigation’s path planning interaction interface move_base.This achieved communication between ROS’s path planning interface move_base and the PX4 side,and the simulation system was validated.During the experiment,it was found that in the control of the PX4 firmware unmanned vehicle,factors such as friction prevented the vehicle from reaching the desired angle.Through the PX4 path planning experiment and analysis of external control programs,it was discovered that the PX4 firmware had a drawback in its open-loop handling of steering angles,which resulted in the external control of the PX4 not being able to control the unmanned vehicle according to the planned desired angle.

关 键 词：差速无人喷药车 自主导航 路径规划 PX4 ROS SLAM建图 

分 类 号：S24[农业科学—农业电气化与自动化]