【作者】 陈诚；

【导师】 李伟； 金波；

【作者基本信息】 浙江大学 ， 机械电子工程， 2012， 博士

【摘要】 六足步行机器人是一种模仿多足动物运动方式的腿式运动结构。它作为一种移动机器人,相比与轮式和履带式机器人,具有落足点离散,主动隔振以及对地表状况适应性强等众多特点,使得它具备广泛应用于各种复杂地表环境的潜能,并因此成为研究热点。但在实际应用中它存在结构复杂、速度慢、控制难度大以及能耗较大等缺点,因而受到很大限制。本论文为改善六足机器人地表环境适应能力,通过设计具有半球形足端的腿部机构,建立了相应的运动学和步态修正算法。针对多足机器人能量利用率低的缺点,通过分析六足步行机器人的运动学、静力学和动力学问题,并建立系统的能量消耗模型,从多个方面对六足机器人的系统能耗进行了优化分析,并通过仿真和实验对优化结果进行验证。第1章以大量的国内外文献调研为基础,介绍了本课题的研究背景和研究意义。全面论述了多足步行机器人及其相关技术的发展历程,重点介绍了多足机器人腿部机构特征、步态生成和控制方法、足端脚力分配算法和能耗优化策略。最后,提出了研究的主要内容。第2章通过建立机器人的三维机构模型和运动学模型,对机器人步行运动下各驱动关节的角度/角速度矢量与各腿足端位置矢量之间的关系展开了正/逆运动学分析。在此基础上,进一步研究了半球形足端对静态稳定步行时关节轨迹生成和躯体运动轨迹所造成的影响,提出了基于半球形足端模型的六足机器人运动学修正算法,并基于三角步态和改进型波浪步态对修正算法进行了仿真和对比分析。第3章在建立电机能耗模型的基础上,推导出了整个机器人系统的能耗方程以及优化目标函数,指出了系统能耗和各关节负载转矩之间存在的联系。在运动学分析的基础上,对六足步行机器人进行准静力学分析,提出了六足机器人足端脚力分配问题。在分析了机器人的平衡条件,并结合摩擦锥等相关约束的基础上,建立了一系列机器人足端脚力的约束方程。利用动力学分析得到了机器人足端脚力与关节转矩之间的关系,将原有的脚力分配问题变换为关节力矩分配问题,获得了各关节力矩的线性等式约束和不等式约束条件,从而对以系统能耗最低为目标的优化目标函数利用序列二次规划法进行求解。与传统的脚力分配方法相比较证明,新的力矩分配方法能够有效节省系统能耗。第4章进一步研究了机器人在步行状态下具有不同步态参数时的系统能耗。利用多项式,对关节空间下的单腿各关节运动轨迹进行构造,运用遗传算法对机器人摆腿相足端轨迹进行了优化设计。以系统能耗比阻作为评估指标,通过仿真获得了机器人在不同总质量下的躯体结构的最优质量分配和躯体高度以及步态的最优步长等参数。最后以力矩分配算法为基础,采用蒙特卡洛法分析了机器人各腿部足端的工作空间及运动裕量,通过分析单腿能耗与该腿足端运动裕量之间的关系,寻求最优的机器人静立姿态。第5章设计搭建了具备仿哺乳类腿机构的六足步行机器人样机平台及其人机交互系统,在此基础上展开了诸如球型足端下的行走实验,静立姿态下的能耗实验以及不同步态参数下的步行能耗实验等一系列实验,对前述理论和仿真分析的结论加以验证。更多还原

【Abstract】 Multi-legged walking machine, compared with wheeled or tracked locomotion, is widely recognized as a much more effective and efficient transportation vehicle, especially on complex and unstructured terrains. Hexapod robots, as a kind of legged walking machines, generally have much more superior performance than those with fewer legs in terms of less complexity of the control method, more walking statically stability and faster walking speed and therefore has already become the research hotspot. However, the legged system today has its own disadvatages, such as low payload to machine load ratio and high energy consumption. Since power is a very limited resource in autonoumous robots, it becomes urgent and imperative for researchers to explore the power-consumption optimizaition techniques. Therefore, this dissertation focues on the minimization of energy expenditure in hexapod robots’standing and walking status.This dissertation is organized as follows:The first chapter of this dissertation performs the background and significance of the project. Based on lots of domestic and foreign literature and reference, this chapter generally introduces the development of multi-legged walking robots and their gait generation techniques. The advantages and disadvantages of variable foot-force distribution methods as well as power consumption optimization methods are especially discussed in this chapter. At the end, the main research purpose and contents are illustrated.In the second chapter, the kinematic model of the robot is built, while the forward/inverse kinematic analysis is used to describe the relationship between the joints’angles/angular velocity and the foot-point positions in the body reference frame. Additional, this chapter analyse the hip misplaced problem caused by round rigid foot and provides a hip-control algorithm for restoring leg coordination. The algorithm is implemented in simulation model with a large-radius ball foot in order to evaluate how the algorithm would perform if applied to a real robot.The third chapter derived the robot’s dynamic model with the foot-force distribution problem formulated. By using the joint torques as the primary variables, the distribtuion of required force and moments to the supporting legs of the hexapod robot is tackled as a torque-distribution problem. The objective function of this problem, which is constructed as the sum of the mechanical energetic cost and heat loss power, is formulated as to minimize the quadratic objective funtion with respect to linear equality and inequality constraints. In contrast to the force distribution method whose objective function corresponds to the sum of the squares of the tip-point force components, the torque distribution scheme could save the system energy cost obviously with appropriate walking velocities and duty factors of robot.In the fourth chapter, the optimal target of the robot standing posture about the dissipation power was proposed through utilizing the energy consumption mathematical model. Combining with the calculation of the foot-points workspace and the kinematic constraints, the Monte-Carlo Method was proposed to analyze the system energy consumption of the robot with different footholds. Furthermore, the protraction movement of a three-joint robot leg is optimized for minimum energy consumption. Foot-point trajectory is performed for various initial-final tip point positions of protraction by genetic algorithm. This chapter also provides analysis on parameters by simulation such as optimal stride length, duty factor, walking velocities, body height and lateral offset with various body mass and payload.In the fifth chapter, the mammal-like hexapod walking robot platform with its Human-Machine Interface are established. The hip-control algorithm lead by round rigid feet is implemented in the platform, while the power consumption experiments on standing posture and walking patterns with varies gait/structure parameters are also carried out. The optimal results of the former chapters derived from simulation are justified by those experiment.更多还原