Modleing And Control Of Dual-Arm Picking Robot For Autonomous And Continuous Operation

There are many domestic and international researches on fruit picking robots,but there seems to be a long way to carry out continuous operation in a real greenhouse environment.The difficulties that impede the large-scale commercial use of the picking robot are mainly embodied in two aspects: hardware and software.In the aspect of hardware,there is no hardware device for fast picking and a structured picking environment that matches it.In the aspect of software,there lacks thorough analysis and modeling of the picking operation environment,and an intelligent software system to adapt to the continuous operation scene.In response to the first problem,a dual-arm picking robot is built,which has the characteristics of dual-arm and multi end effectors,and can be well adapted to the semi structured picking environment of greenhouse.This paper mainly focuses on the second problem,and then develops an intelligent software system for the continuous operation of the dual-arm picking robot.The main researches are as follows1)For the software and hardware design of picking robot in the semi structured environment,the characteristics of semi structure of the greenhouse environment is firstly analyzed,and several difficulties in the continuous picking of tomatoes are summarized as: identification and positioning problems under variable light conditions,intelligent picking problem in multi cluster growth environment,and complex multitask programming problem.Secondly,the characteristics of the hardware design of the tomato dual-arm picking robot are summarized.At last,an environmental perception scheme for a dual-arm picking robot in a semi structure environment is proposed,and the intelligent computer software system is designed.2)For the difficulty of data characterization and modeling in the semi structure environment of greenhouse,the characteristics of this environment are investigated,and a spatio-temporal database system based on dynamic association is developed.The spatio-temporal modeling and dynamic correlation of tomato objects and robot objects during the continuous picking process are carried out.The spatio-temporal database is built on the server side using the SQL server.The function of the system debugging and storage is normal,and the communication of the software system and the host computer is normal.3)For the difficulty of developing the software system,a hierarchical modular software system framework is designed.Moreover,a scheduling method of functional modules is designed based on the idea of finite state machine for the complex multi task planning problem of tomato continuous picking process.The task scheduling design based on the finite state machine(FSM)reduces the difficulty of development work and improves the efficiency of development.4)Based on the previous design,programing thought and the ROS system,the sub node programming of visual module,motion module,eye hand coordination module and task planning module are implemented.The program is tested to be sufficient and successful on the industrial control machine,and the continuous picking of the robot is realized successfully.The experimental results shows that if total target fruit within visual field is not more than three,the average fruit picking time is 35 s.