电-气混合式农业机器人抓手设计及自适应主动柔顺抓取控制研究

The developments of agricultural robots can largely reduce the farmers' labor intensity, improve working conditions, as well as increase work efficiency. Grasping, carrying and placing of fragile agricultural materials are the fundamental capabilities and key tasks for agricultural robots and robotic manipulators. Secure grasping not only requires contacting the agricultural materials but also preventing potential slip and damage while the materials are picked and placed. In the automatic grasping processes, material damage depends on the aggressiveness of the grippers and on the biomechanical properties of grasped material. On the basis of our considerable experience in physical property analyzing, biomechanical modeling and mechatronics in agriculture, the design of hybrid electric-pneumatic robotic self-adaptive gripper, sensor configuration, the mechanism of non-destructive robotic grasping of agricultural materials, as well as the flexible grasping strategies and control methods based on multimodal feedback sensory data will be investigated. Dynamic simulation analysis and grasping experiments would be conducted to verify the non-destructive grasping control. Special attention would be paid to the two aspects, both the design of self-adaptive flexible robotic gripper and flexible grasping control based on sensory data, to improving the grasping performance of agricultural robotic grippers. The results would improve the grasping performance of the agricultural robots, and provide potential information for other advanced robots in expanding applications in agriculture and food engineering.

农业机器人的发展对降低民工劳动强度、改善劳动环境、提高作业效率具有重要意义。抓取操作柔嫩易损的农业物料是农业机器人进行自动化作业过程中面临的一项共性关键技术。农业物料本身固有的生物-力学特性和机械手抓取的柔顺性是影响机器人安全无损抓取的两个重要方面。结合申请者在物料物性分析、机电控制等方面的知识储备，研究电-气混合式机械手自适应柔性设计及传感器优化布置，探索农业物料的无损抓取机理；针对农业物料几何形态和生物-力学特性的多样性、多变性特点，重点研究多模态视触融合反馈的电-气混合式机械手无损抓取策略优化与自适应主动柔顺控制系统构建方法，并通过仿真分析和番茄抓取开展实验验证。项目充分融合抓手自适应柔性设计和多模态融合信息反馈控制的优点提高机械手抓取的稳定性和柔顺性，研究结果对于改善农业物料操作机器人的主动柔顺抓取性能具有重要意义，同时可以为传统机器人在农业、食品加工等服务领域的应用拓展提供参考。

抓取操作异形易损果蔬等农业物料是农业机器人进行自动化作业的基本功能和智能化系统的集中体现。果蔬等农产品具有形态多样、柔嫩易损、粘弹流变等特性，农业机器人抓手的自主顺应特性、多元感知能力及柔顺控制性能直接关系到机器人自主作业的成败。本项目面向农产品、食品的流水线自动化分拣及鲜食果蔬的选择性采收，沿着柔嫩目标柔性抓取的研究思路，提出了一系列创新性研究方案，覆盖了抓手柔性设计、分布式触觉传感、视觉伺服控制、柔顺抓取控制等农业机器人抓取研究诸环节。研究内容与成果主要包括：（1）提出了农业机器人电-气混合抓手的结构设计、构型综合和系统总成新思路，创新研发了3D柔性打印末端、柔性仿生驱动系统，面向流水线分拣和果蔬采摘，开发了多款新型刚柔耦合机械手；（2）开展了视觉、触觉、弯曲等多种传感器的优化布置研究，实现了视觉-触觉传感器的网格化集成、标定、目标位姿估计、动态触觉感知、状态辨识和抓取规划等；（3）基于多模态信息反馈开展了农业机器人无损抓取策略优化，建立了农业机器人自适应主动柔顺控制系统。本项目从农业机器人抓手的结构、传感与控制三个方面，探索了农业机器人主动柔性抓取的关键技术与无损抓取策略，提升了抓手抓取性能，减少了抓取变形与损伤。同时，该研究还可以有效拓展传统机器人及抓手在精准农业、食品加工、餐饮服务等多样化柔性作业场景的应用。