基于改进目标检测算法的AGV避障方法研究

目的 针对食品包装的抓取和装配，利用人工成本高且容易装错等问题，基于方位特征(POC)方程的并联机构结构综合方法，提出一种以3PUPaR并联机构为主体的抓取机器人机构。方法 计算3PUPaR并联机构的自由度、方位特征集以及耦合度等拓扑特性，验证机构具有空间纯移动的运动特性，建立机构的运动学逆解方程，基于此，采用数值法得到机构的工作空间与运动灵巧度。并研究参数尺寸大小对工作空间与全局灵巧度的影响，以全局灵巧度和工作空间为优化目标，选择具有Pareto解的NSGA-II算法完成多目标参数优化。结果 结果表明，机构具有较大且连续的工作空间。结论 该新型三自由度并联机器人能够满足食品包装的抓取和装配时运动范围。

AGV Obstacle Avoidance Method Based on Improved Target Detection Algorithm

The work aims to propose a grasping robot mechanism with 3PUPaR parallel mechanism as the main structure based on the synthesis method of parallel mechanism structure by the position and orientation characteristics equation (POC) equation aiming at the problems of high labor cost and wrong loading in the process of grasping and assembling food packaging. The degree of freedom, the position and orientation characteristics set, the coupling degree of 3PUPaR parallel mechanism and other topological properties were calculated to verify whether the mechanism had the motion characteristics of space pure movement and establish the inverse kinematics equation of the mechanism. Based on this, the numerical method was used to obtain the mechanism workspace and dexterity and study the influence of different parameters on workspace volume and dexterity. Taking the global dexterity and workspace as the optimization objective, NSGA-II algorithm with Pareto solution was selected to complete the multi-objective optimization. The results showed that the mechanism had a large and continuous workspace.The New 3-DOF parallel robot can meet the motion range of food packaging during grasping and assembly.