硕士学位论文 基于气动肌肉驱动的仿蛙腿跳跃机构 控制系统研究 RESEARCH ON CONTROLSYSTEM OF FROG-INSPIRED HOPPING LEG POWERED BY PNEUMATIC MUSCLE 焦磊涛 哈尔滨工业大学 2013年7月
硕士学位论文 基于气动肌肉驱动的仿蛙腿跳跃机构 控制系统研究 RESEARCH ON CONTROL SYSTEM OF FROG-INSPIRED HOPPINGLEG POWERED BY PNEUMATIC MUSCLE 焦 磊 涛 哈尔滨工业大学 2013 年 7 月
国内图书分类号：TP242.2 学校代码：10213 国际图书分类号：681.5 密级：公开 工学硕士学位论文 基于气动肌肉驱动的仿蛙腿跳跃机构 控制系统研究 硕士研究生：焦磊涛 导 师：赵杰教授 申请学位：工学硕士 学 科：机械电子工程 所在单位：机电工程学院 答辩日期：2013年7月 授予学位单位：哈尔滨工业大学
国内图书分类号：TP242.2 学校代码：10213 国际图书分类号：681.5 密级：公开 工学硕士学位论文 基于气动肌肉驱动的仿蛙腿跳跃机构 控制系统研究 硕士研究生：焦 磊 涛 导 师：赵 杰 教授 申请学位：工学硕士 学 科：机械电子工程 所 在 单 位：机电工程学院 答 辩 日 期：2013 年 7 月 授予学位单位：哈尔滨工业大学
Classified Index:TP242 U.D.C:681.5 Dissertation for the Master Degree in Engineering RESEARCH ON CONTROL SYSTEM OF FROG-INSPIRED HOPPING LEG POWERED BY PNEUMATIC MUSCLE Candidate: Jiao Leitao Supervisor: Prof.Zhao Jie Academic Degree Applied for: Master of Engineering Speciality: Mechatronics Engineering Affiliation: School of Mechatronics Engineering Date of Defence: July,2013 Degree-Conferring-Institution: Harbin Institute of Technology
Classified Index: TP242 U.D.C: 681.5 Dissertation for the Master Degree in Engineering RESEARCH ON CONTROL SYSTEM OF FROG-INSPIRED HOPPINGLEG POWERED BY PNEUMATIC MUSCLE Candidate： Jiao Leitao Supervisor： Prof. Zhao Jie Academic Degree Applied for： Master of Engineering Speciality： Mechatronics Engineering Affiliation： School of Mechatronics Engineering Date of Defence： July, 2013 Degree-Conferring-Institution： Harbin Institute of Technology
哈尔滨工业大学工学硕士学位论文 摘要 仿生跳跃机器人凭借其优越的越障能力以及环境适应能力已成为目前国内 外研究的热点。本文对已有的仿青蛙跳跃机器人进行分析，提出有效可行的机 构改进和优化的方案，并利用ADAMS进行性能仿真验证。同时采用实验的方 法建立了跳跃机构驱动元件一气动人工肌肉的数学模型，并研究气动人工肌 肉的位置控制策略以实现机器人的位姿调整。在此基础上设计搭建了跳跃机构 的控制系统，进行跳跃机构跳跃性能实验，实现跳跃机构不同类型的跳跃。 首先，推导气动人工肌肉的理想数学模型，在此基础上考虑影响肌肉特性 的各个原因，推导出改进的肌肉驱动数学模型，并将理想模型、改进模型和实 验曲线进行对比。考虑到设计控制器对肌肉模型的要求，提出一种建立简单肌 肉经验模型的方法，利用此种方法通过实验建立机器人所用的两种肌肉的经验 模型，并在实验中验证此模型的正确性。 其次，考虑到跳跃机构在跳跃过程中跳跃角度对跳跃性能的影响，基于人 工肌肉的数学模型，利用模糊控制方法和双层PD控制方法设计了人工肌肉的 位置控制器以用于实现机器人的姿态调整，在MATLAB Simulink中进行系统 仿真，并在实验中验证所提控制方法的可行性。 此外，针对前一代仿生跳跃机器人在机构设计中存在的不足，对其进行机 构改进和优化，设计加入关节力闭环和角度数据采集单元，以实现跳跃机构在 跳跃过程中关节角度信息的实时检测和力、位置的闭环控制。并对改进后的跳 跃机构进行ADAMS仿真，验证了改进后跳跃机构的跳跃性能。 最后，搭建仿蛙腿跳跃机构的控制系统实验平台，在此基础上进行跳跃机 构位姿调整实验和跳跃实验，以验证本文所提出的气动肌肉位置控制策略和控 制系统设计的有效性。同时在机器人跳跃实验中通过多次测试机器人的跳远、 跳高和翻越障碍物的能力，总结出影响机器人跳跃性能的主要因素，为仿生跳 跃机构以后的进一步优化和改进奠定基础 关键词：仿蛙腿跳跃机构：人工肌肉：肌肉模型：位置控制：模糊控制
哈尔滨工业大学工学硕士学位论文 I 摘 要 仿生跳跃机器人凭借其优越的越障能力以及环境适应能力已成为目前国内 外研究的热点。本文对已有的仿青蛙跳跃机器人进行分析，提出有效可行的机 构改进和优化的方案，并利用 ADAMS 进行性能仿真验证。同时采用实验的方 法建立了跳跃机构驱动元件——气动人工肌肉的数学模型，并研究气动人工肌 肉的位置控制策略以实现机器人的位姿调整。在此基础上设计搭建了跳跃机构 的控制系统，进行跳跃机构跳跃性能实验，实现跳跃机构不同类型的跳跃。 首先，推导气动人工肌肉的理想数学模型，在此基础上考虑影响肌肉特性 的各个原因，推导出改进的肌肉驱动数学模型，并将理想模型、改进模型和实 验曲线进行对比。考虑到设计控制器对肌肉模型的要求，提出一种建立简单肌 肉经验模型的方法，利用此种方法通过实验建立机器人所用的两种肌肉的经验 模型，并在实验中验证此模型的正确性。 其次，考虑到跳跃机构在跳跃过程中跳跃角度对跳跃性能的影响，基于人 工肌肉的数学模型，利用模糊控制方法和双层 PID 控制方法设计了人工肌肉的 位置控制器以用于实现机器人的姿态调整，在 MATLAB Simulink 中进行系统 仿真，并在实验中验证所提控制方法的可行性。 此外，针对前一代仿生跳跃机器人在机构设计中存在的不足，对其进行机 构改进和优化，设计加入关节力闭环和角度数据采集单元，以实现跳跃机构在 跳跃过程中关节角度信息的实时检测和力、位置的闭环控制。并对改进后的跳 跃机构进行 ADAMS 仿真，验证了改进后跳跃机构的跳跃性能。 最后，搭建仿蛙腿跳跃机构的控制系统实验平台，在此基础上进行跳跃机 构位姿调整实验和跳跃实验，以验证本文所提出的气动肌肉位置控制策略和控 制系统设计的有效性。同时在机器人跳跃实验中通过多次测试机器人的跳远、 跳高和翻越障碍物的能力，总结出影响机器人跳跃性能的主要因素，为仿生跳 跃机构以后的进一步优化和改进奠定基础 关键词：仿蛙腿跳跃机构；人工肌肉；肌肉模型；位置控制；模糊控制
哈尔滨工业大学工学硕士学位论文 Abstract The research on bionic hopping robots is becoming a popular subject recently due to its characters of strong obstacle-overleaping ability and great field and environment adaptability.Analyzing the previous frog-inspired hopping robot,the optimization and improvement programs are proposed and verification for the new robot is made in ADAMS.The mathematical models of the pneumatic muscles were built by experimental method and the position control method for pose alignment of the robot was designed.Besides,a control system is proposed in this dissertation and to realize different jumping modes. The ideal mathematical model of the pneumatic artificial muscle was derived. Based on the ideal model and considering the factors which affect the muscle characteristics,a new improved mathematical model was deduced.Then the improved model was compared with the experimental curves and ideal mathematical model.Considering the demands of designing controller for pneumatic muscle,in this dissertation,a modeling method which builds simple mathematical model of pneumatic muscle was proposed,and the mathematical models for the pneumatic muscles used in the robot were built with the proposed modeling method.Furthermore the correctness of this model was verified by the experiment. Considering the influence of the take-off angle on the jumping performance and based on the mathematical model of the pneumatic muscle,a control strategy for pneumatic muscle by the fuzzy control and double-layer PID control was proposed in this dissertation to adjust the take-off angle of the robot.The system simulation was made in the MATLAB Simulink and the accuracy of the control strategy was tested in the experiment. With improving and optimizing the mechanism design deficiencies,a new robot is designed.The force sensors and angle sensors is equipped on the new robot, thus the robot has the ability of force/angle closed-loop control and angle acquisition of joints.Besides,ADAMS simulation was performed to test the jumping ability of the new robot. Experimental platform of the bionic hopping robot control system is bulit,and the robot pose alignment and jumping experiments are conducted based on the experimental platform to validate the proposed muscle position control strategy and the effectiveness of control system.In jumping experiments,the robot's abilities to broad-jump,high jump and climbing over obstacle were tested repeatedly to find out the factors which affect the jumping performance.This experiment laid a
哈尔滨工业大学工学硕士学位论文 II Abstract The research on bionic hopping robots is becoming a popular subject recently due to its characters of strong obstacle-overleaping ability and great field and environment adaptability. Analyzing the previous frog-inspired hopping robot, the optimization and improvement programs are proposed and verification for the new robot is made in ADAMS. The mathematical models of the pneumatic muscles were built by experimental method and the position control method for pose alignment of the robot was designed. Besides, a control system is proposed in this dissertation and to realize different jumping modes. The ideal mathematical model of the pneumatic artificial muscle was derived. Based on the ideal model and considering the factors which affect the muscle characteristics, a new improved mathematical model was deduced. Then the improved model was compared with the experimental curves and ideal mathematical model. Considering the demands of designing controller for pneumatic muscle, in this dissertation, a modeling method which builds simple mathematical model of pneumatic muscle was proposed, and the mathematical models for the pneumatic muscles used in the robot were built with the proposed modeling method. Furthermore the correctness of this model was verified by the experiment. Considering the influence of the take-off angle on the jumping performance and based on the mathematical model of the pneumatic muscle, a control strategy for pneumatic muscle by the fuzzy control and double-layer PID control was proposed in this dissertation to adjust the take-off angle of the robot. The system simulation was made in the MATLAB Simulink and the accuracy of the control strategy was tested in the experiment. With improving and optimizing the mechanism design deficiencies, a new robot is designed. The force sensors and angle sensors is equipped on the new robot, thus the robot has the ability of force/angle closed-loop control and angle acquisition of joints. Besides, ADAMS simulation was performed to test the jumping ability of the new robot. Experimental platform of the bionic hopping robot control system is bulit, and the robot pose alignment and jumping experiments are conducted based on the experimental platform to validate the proposed muscle position control strategy and the effectiveness of control system. In jumping experiments, the robot's abilities to broad-jump, high jump and climbing over obstacle were tested repeatedly to find out the factors which affect the jumping performance. This experiment laid a