Research Status and Development Direction of Design and Control Technology of Fruit and Vegetable Picking Robot System
WU Jianqiao,FAN Shengzhe,GONG Liang,YUAN Jin,ZHOU Qiang,LIU Chengliang
DOI: https://doi.org/10.12133/j.smartag.2020.2.4.202011-sa004
2020-01-01
Abstract:Vegetable and fruit harvesting is the most difficult production process to achieve mechanized operations. High-efficiency and low-loss picking is also a worldwide problem in the field of agricultural robot research and development, resulting in few production and application equipment currently on the market. In response to the demand for picking vegetables and fruits, to improve the time-consuming, labor-intensive, low-efficiency, and low-automation problems of manual picking, scholars have designed a series of automated picking equipment in the recent 30 years, which has promoted the development of agricultural robot technology. In the research and development of fresh vegetable and fruit picking equipment, firstly, the harvesting object and harvesting scene should be determined according to the growth position, shape and weight of the crop, the complexity of the scene, the degree of automation required, through complexity estimation, mechanical characteristics analysis, pose modeling and other methods clarify the design requirements of agricultural robots. Secondly, as the core executor of the entire picking action, the design of the end effector of the picking robot is particularly important. In this article, the structure of the end effector was classified, the design process and method of the end effectors were summarized, the common end effector driving methods and cutting methods were expounded, and the fruit collection mechanism was summarized. Furthermore, the overall control scheme of the picking robot, recognition and positioning method, adaptive control scheme of obstacle avoidance method, quality classification method, human-computer interaction and multi-machine cooperation scheme were summarized. Finally, in order to evaluate the performance of the picking robot overall, the indicators of average picking efficiency, long-term picking efficiency, harvest quality, picking maturity rate and missed picking rate were proposed. The overall development trend was pointed that picking robots would develop toward generalization of picking target scenes, diversified structures, full automation, intelligence, and clustering were put forward in the end.