面向智能上下料系统的全局视觉AGV路径导航系统

Global Vision AGV Path Navigation System for Intelligent Loading and Unloading Systems

针对固定路径导航的AGV在智能上下料系统中的应用无法满足智能化工位布局不确定性的需求的问题,提出了基于全局视觉引导的AGV路径导航系统,该系统由路径生成与跟踪两部分功能组成。采用三次多项式插值曲线方程,结合全局视觉所提取的工位出入口、上下料区域和AGV位姿数据,将工位出入口和上下料区域串联得到适用于工位环境的数字化物流路径。融合传统PID控制器,将数字化物流路径和位姿偏差作为输入量,构建了基于全局视觉伺服的AGV运动控制模型。实验结果表明,无论是直线型还是曲线型工位,所提出的AGV路径导航系统都能准确生成并跟踪实际工位的物流路径,最大路径偏差为10.83 pixels,最大平均路径偏差为4.02 pixels,具有良好的适用性和稳定性。

The application of AGV with fixed path navigation in intelligent loading and unloading systems cannot meet the requirements of uncertainty in intelligent workstation layout.To address the above issues,an AGV path navigation system based on global vision guidance is proposed,which consists of two functions:path generation and tracking.Using cubic polynomial interpolation curve equation,combined with the workstation entrance and exit,loading and unloading areas,and AGV pose data extracted by global vision,the workstation entrance and exit and loading and unloading areas are connected in series to obtain a digital logistics path suitable for the workstation environment.Integrating traditional PID controller,the digital logistics path and pose deviation are taken as input variables to construct an AGV motion control model based on global vision servo.Experimental results show that the proposed AGV path navigation system can accurately generate and track the actual logistics path of workstations,whether they are linear or curved,with a maximum path deviation of 10.83 pixels and a maximum average path deviation of 4.02 pixels,which has good applicability and stability.