农业机器人技术为我国重大科技战略重点需求。目前农业果实采摘因复杂环境及劳动力结构性短缺，导致自动化智能化程度仍然很低。本项目以自主研发的果实采摘机器人为研究对象，建立水果采摘阶段的果园管理系统，开展果实成熟度检测与定位、路径规划与避障导航、分数阶滑模控制方法研究。重点研究:.(1)自然光照及遮挡环境下基于深度学习的果实检测与定位方法研究，提高系统的实时采摘和识别准确率；.(2)基于激光雷达和RGBD相机的果实采摘机器人路径规划与避障导航，促进采摘机器人路径规划与避障导航能力；.(3)采摘机器人的分数阶滑模轨迹跟踪控制算法，改善系统的稳定性和控制精度，有效地削弱系统抖振，增强采摘机器人的抗扰能力。.本项目探索复杂动态环境下果实采摘机器人关键技术，实现果实采摘作业的机械化、自动化和智能化，保证果实的适时高效采摘、降低收获作业成本,解决当前国家面临的劳动力短缺问题具有重大现实意义。

Agricultural robotics technology is the key demand of our country's major science and technology strategy. At present, the automation and intelligence of agriculture fruit picking is still low, due to the influence of the complicated environment and the structural shortage of labor force. .In this project, we will establish the orchard management system of fruit picking stage and research the apple maturity degree detection and location, path planning and obstacle avoidance navigation, fractional sliding mode control through self-developed fruit picking robots. The main contents include:.(1)Study the visual detection and location for apple maturity under the natural light and block environment based on the deep learning method, improve recognition accuracy and real-time performance of the apple picking robot system;.(2)Research the path planning and obstacle avoidance navigation algorithm of fruit picking robots based on lidar and RGBD camera to promote the capability of navigation and obstacle avoidance for the fruit picking robot;.(3)Proposed fractional sliding mode tracking control algorithms to improve robustness and control precision for picking robot system, weaken the system chattering and enhance immunity of the picking robot.. The project explores the key technology of the fruit picking robot under he complex and dynamic environment in order to achieve mechanization、automation and intelligent of the fruit picking robot, ensure the timely and efficient for fruit harvesting, reduce the costs of harvesting operations. It is of great practical significance to solve the current labor shortage problem in our country.