基于关节变量反馈的六足机器人闭环控制系统的研究

【作者】 段玉波；

【导师】 吴光敏；

【作者基本信息】 昆明理工大学 ， 物理电子学， 2012， 硕士

【摘要】 六足机器人因其各向性较好,对行走路面的要求很低,能够适应各种不规则路面,并且在运动中具有良好的稳定性,在机器人研究领域具有十分重要的意义。本论文的研究对象就是一种结构典型的六足机器人。本文基于深圳市德普施科技有限公司的重载六足机器人平台Nomad HDATS,对其运动结构和三角步态进行了分析,通过“D-H”方法建立起机器人各关节和杆件的运动学模型,写出它们之间的变换矩阵,推导出该六足机器人的正向运动学方程和逆向运动学方程。采用拉格朗日算法(Lagrange)建立了机器人的动力学方程,解决了动力学逆向问题,计算出机器人各关节执行器所需要的驱动力矩。机器人的控制一般是闭环控制系统,它能让机器人在运动过程中根据实时信息进行自主决策,实现自主控制。机器人闭环控制系统的条件是机器人能够获取外部环境和机器人自身姿态的信息,外部环境信息一般通过传感器来获得,如视觉、触觉、光电位置检测等传感器；机器人姿态信息则主要通过角度传感器来确定,如水平陀螺仪、指南针传感器等。本文的研究要点就是控制系统如何确定机器人姿态的问题。本论文对这一问题的研究并未使用角度传感器,而是考虑通过机器人各关节变量的值来确定机器人的姿态。经深入研究,发现机器人所使用的普通模拟舵机也能够有效反映出关节变量的实际值。基于以上的思考与分析,本文针对所研究的具有相同的六条三自由度腿的机器人,设计出一套基于关节变量信息反馈的六足机器人闭环控制系统。本文所设计的机器人闭环控制系统的基本原理是基于关节变量反馈系统,将采集到的反映各关节舵机输出轴实时位置的模拟信号经过A/D芯片转换成数字信号并传送给单片机,再由单片机根据推导出的转换算法编写的程序将其换算成对应的舵角位置信息或机器人距离地面的高度信息。最后,将此设计方案嵌入到机器人原有的开环控制系统中,控制器能够通过此套反馈系统扫描全部关节舵机的实时位置信息,并以此作为参考依据综合评价出机器人此时的姿态和可能碰到的状况,从而决定其下一步的运动。总结之,本文对所研究的六足机器人进行了相关理论的研究,提出了基于关节变量信息反馈的机器人闭环控制系统的设计方案,并对其进行了分组测试,测试结果达到预期,证实了系统的可行性与有效性。系统的设计思路比较新颖,成功解决了普通模拟舵机无法通过实时位置信息获取机器人姿态以实现闭环控制系统设计的问题,为今后机器人控制系统设计领域提供了一定的参考思路和借鉴意义。更多还原

【Abstract】 Hexapod robot plays very important significance in robot research field for its performance in each direction is so good that there is low demand of pavement and it can adapt to all kinds of irregular road surface. More over, its stabilityis good in moving. The research object of this paper is a kind of hexapod robot with typical structure.Based on the hexapod robot Nomad HDATS(Heavy Duty All Terrain System) of Shenzhen Depush technology Co., LTD, this paper analyzes the sports structure and tripod gait. Through the "D-H" method to establish the kinematics model of the robot joints and stems, writing the transformation matrixes between them, deducing the positive kinematics equation and the inverse kinematics equations of the hexapod robot. Using Lagrange algorithm to establish the robot dynamics equation, the paper solves the problem of inverse dynamics, calculating the driving moment needed by each joint actuators of the hexapod robot.The control of the robot is usually closed loop control system, its principle is that the robot makes independent decision to realize the independent control according to the real-time information in moving process. Robot closed loop control system requires the robot can obtain external real-time environment and its own posture, the external environment information can be obtained through the sensor in general, such as visual, touch, photoelectric position detection sensors, etc; The robot posture can be determined by angle sensor, such as horizontal gyro, compass sensor, etc. The paper studies that control system how to determine the robot posture. This thesis doesn’t use angle sensor on the problem, but considers that we would be able to make sure the posture of the robot through robot joint variables. After further study, we find that common simulating steering gear used by our robot also can effectively reflect the actual value of the joint variables. Based on the above thinking and analysis, this paper designs a set of robot closed-loop control system based on information feedback of joint variables to the hexapod robot we studied which has six equal limbs with3DOF.Based on the joint variable feedback system, the basic principle of the closed-loop control system designed is to convert analog signal reflecting the real-time position of the joint steering gear output shaft into digital signal by A/D chip and send it to the single chip microcomputer. Then according to the programming algorithm, the single chip microcomputer converts the digital signal into servo motor’s position information or heigth that the robot is from the ground. Finally, adding the design scheme into original control system, the controller will be able to scan real-time position information of all joint servo motors using the designed feedback system and reference to the real-time position information to judge posture of the robot and the posible situation at that time, and then decide the next movement.In general, the paper makes some research and analysis to the theory related to the hexapod robot, puts forward design scheme of the robot closed-loop control system based on the joint variable information feedback, and make group testing, the result of which achieves the expected confirming the feasibility and effectiveness of the system. The designing idea of the system is new, succeeding in solving the problem that common simulation steering gear can not be used for designing the closed-loop control system to get the robot’s posture, providing some reference ideas for robot control system design in future.更多还原