|导师||喻俊志 ; 吴正兴|
|关键词||仿生机器海豚 机动控制 深度控制 路径跟踪 协同追踪 水质监测|
一、结合复杂动态水域实时水质监测的需求，以虎鲸为仿生对象研制了一种面向水质监测的机动型机器海豚系统。整体采用流线型设计以降低流体阻力；静动密封相结合的密封方式增加其耐压性；模块化的结构设计方便安装与更换；基于中枢模式发生器的底层运动控制器提供稳定的多模态运动。这些均有助于其完成高效率、大范围、无污染的自主水质监测任务。2016 年 9 月，在青海省玉树市禅古水库成功开展了野外水质监测应用实验。基于机器海豚的动态水质监测实验的成功不仅为三江源水源保护提供重要的数据支撑，而且为动态自主水质监测提供了新的技术手段。
二、针对机动型机器海豚的定深控制问题，提出了一种基于视线导航法和滑模观测器的滑模模糊控制算法，实现了机器海豚最小平均深度误差为 0.18 cm 的定深运动。采用视线导航法将定深控制问题转化为俯仰角跟踪控制问题，利用滑模观测器实时估算机器海豚游动速度；基于速度跟踪误差与俯仰角跟踪误差分别构建了滑模控制器获取定深运动过程中的驱动力和俯仰力矩，并进行了李雅普诺夫稳定性判据；利用模糊推理将驱动力及其变化率和俯仰力矩及其变化率分别映射到机器海豚腰、尾关节的拍动频率和鳍状肢的偏转角上；仿真和对比实验验证了所提控制方法的有效性。
四、针对双机器海豚协同目标追踪问题，提出了一种基于改进的快速遍历随机树（Rapidly-exploring Random Tree，RRT）路径规划算法和基于行为的协同追踪策略，实现了双机器海豚对静态目标与动态目标的协同追踪。为满足协同路径快速规划的需求，给出了一种基于经典 RRT*-Connect 算法的优化算法。仿真结果表明所提算法具有更快的收敛速度、更短的全局路径和更少的路径点。为了减少算法执行时间、提高追踪效率，采用基于行为的协同追踪策略和集中式结构框架将目标追踪问题分解为多个简单的子行为，通过简单子行为的有机组合来完成复杂的追踪任务。最终，野外实验验证了所提追踪策略的有效性。
Nature is the source of inspiration and the cradle of innovation. After the long-term natural selection, dolphins are fairly well endowed with many novel mechanisms and amazing locomotion abilities, which is well-suited to aquatic environments. The development of bio-inspired robotic dolphins based on the bionics aims to lay theoretical and technical foundations for the design and control of high-performance underwater vehicles. This thesis is devoted to several critical problems of bio-inspired robotic dolphin, involving the modular design of the overall robotic system, the mobile application of water quality monitoring, depth control, maneuvering control, path-following control, and cooperative target tracking for dual robots. The technical contributions made are summarized as follows.
Firstly, a novel mechanical design modeled after a killer whale and used for mobile water quality monitoring is proposed in complex and dynamic aquatic environments. For the sake of hydrodynamic performance, the conceived robot adopts a well-streamlined body shape to offer an expected lift-to-drag ratio. Glue sealants and dynamic seal unit are utilized to enhance waterproofing performance. In particular, a modular approach to robot design facilitates installation and replacement purposes. Furthermore, a Central Pattern Generator (CPG)-based controller is then built to govern the multimodal locomotion of the robotic dolphin. All these measures together assist the robotic dolphin in enhancing the effect of water quality monitoring. To examine the operational performance of the developed robotic dolphin, some field experiments had been conducted at Changu Reservoir of Yushu, Qinghai Province, in September 2016. The success in tests not only provides precious data for ecological and environmental protection in Three River Sources areas but also offers technical support to the real-time dynamic water quality monitoring tasks.
Secondly, a sliding-mode fuzzy controller is developed on the basis of a line-of-sight (LOS) guidance and a sliding-mode observer (SMO) for the maneuverable robotic dolphin. The controller successfully regulates the vertical displacement of the robot within the minimum average depth error as small as 0.18 cm. The LOS guidance method and SMO are employed to obtain the desired pitch angle and speed estimator. A sliding-mode controller is then developed to overcome systematic uncertainties and environmental disturbances, and the Lyapunov stability theory is utilized to analyze the stability and convergence properties of the closed-loop system. Furthermore, depth control in the actual robot is realized by utilizing a fuzzy logic controller, which is represented as the mapping of the input propulsion forces/moments and output control parameters, like flapping frequency and deflection angle. Numerical and experimental results demonstrate that the proposed control strategy successfully steers the robot towards and along the desired depth.
Thirdly, according to the path following problem of the maneuverable robotic dolphin, a sliding-mode fuzzy control method is presented to follow some paths and smooth transition in the horizontal plane. In order to generate agile and precise turning maneuvers, modified mechanical flippers are designed for rich turning patterns. After analyzing the pros and cons of each pattern, it lays the groundwork for path following application. As for the path-following control, a global camera is employed to obtain the real-time position of the robot. Then, the LOS guidance law is presented to acquire the desired course angle. After that, a sliding mode controller is constructed to obtain the yaw moment, and the asymptotic stability of the control scheme is investigated via the Lyapunov stability theory. Furthermore, a fuzzy logic controller is built to establish the mapping relationship between the outputs from SMC and turning patterns and flapping frequency for the robotic dolphin. Both simulations and experiments are conducted to validate the effectiveness and robustness of the proposed method.
Finally, a modified rapidly-exploring random tree (RRT)-based path planner and behavior-based cooperative tracking strategy are presented for a dual robotic dolphin system to fulfil the cooperative target tracking task. Specifically, with full consideration of the demand of fast path planning, an optimization algorithm based on the RRT*-Connect is employed to generate paths for the dual robotic dolphins with fewer waypoints, faster convergence rate, and better stability. Furthermore, a behavior-based approach in conjunction with centralized architecture is implemented to achieve high-level decision-making. Finally, both simulations and field experiments are carried out to verify the effectiveness of the proposed control scheme. The obtained results also shed light on the control and implementation of cooperative multi-robot target tracking in complex aquatic environments.
|刘金存. 机动型仿生机器海豚运动控制与应用研究[D]. 北京. 中国科学院大学,2018.|