【作者】 侯绪研；

【导师】 邓宗全；

【作者基本信息】 哈尔滨工业大学 ， 机械设计及理论， 2009， 博士

【摘要】 随着航天技术的发展和国民经济实力的增强,探月计划已被纳入日程。月球表面地形复杂、未知,加之月球车所携带的能源有限,因此,降低月球车行驶的能耗性,充分利用有限的能源,是保证探月任务顺利完成的关键。为消除六轮摇臂式月球车行进过程中的寄生功率损失,从而降低能耗,提出了运动协调控制模式。运动协调规划模型是从寄生功率产生原因入手,以月球车姿态角为状态变量,以电机转速为规划变量,电机转矩等为中间变量,以六轮瞬时功率之和为目标函数,以满足速度投影定理的运动学模型、准静态平衡力学模型、驱动电机额定功率、地面驱动力系数、电机最大驱动力等为约束条件的有约束最优化模型。在对规划模型进行整理后,基于Kuhn-Tucker条件求解规划模型,并将最优解作为依据控制月球车行进。仿真表明,运动协调控制模式可以有效减少寄生功率损失,降低能耗。根据摇臂式悬架的被动适应地形特性,提出了根据地形信息计算月球车姿态角的模型,结合运动协调控制模式进一步计算运动协调控制参数,最终计算通过路径所需能量。在此基础上,在分析了现有路径规划系统的片面性和局限性后,提出了以通过路径所需能量、所需时间、路径起伏状况、路径长度等为属性的路径多属性决策系统。该系统采用TOPSIS法,从若干条可通行路径中决策出最优路径。仿真表明,路径多属性决策系统所优选出的最优路径,兼顾低能耗性、快速性、平稳性等要求。作为路径多属性决策系统的拓展,提出了月球车群的路径多目标优选动态规划系统。该系统针对由两轮并列式探路月球车和六轮摇臂式主月球车组成的月球车群,提出了“探路月球车往复式前进,主月球车路径多目标优选动态规划”的月球车群路径决策方式。仿真表明,虽然路径多目标优选动态规划系统消耗更多的时间,但是降低了能耗。为避免由于模型不准确所产生的误差,降低控制系统对视觉系统的依赖,提出了结合运动协调控制模式的马尔可夫预测控制系统。该系统针对月球表面地形分布所具有的马尔可夫性和月球车摇臂式悬架的被动适应地形特性,通过采用马尔可夫预测方法预测下一状态的地形信息,预测下一状态的月球车姿态角,从而预测下一状态的运动协调控制参数。仿真验证了该系统的正确性。由于该系统是针对研究对象特性所选用的方法,从而避免了常规预测方法的通用模型,减少了计算量。为确定运动协调控制模式在实际应用中的控制效果,充分验证运动协调控制策略的有效性,进行了原理样机实验。实验以现有的六轮摇臂式月球车原理样机为本体,构建了由上位机、数模转换模块、驱动模块、传感器模块、电源模块等组成的硬件系统,编写了由控制策略算法程序、传感器数据读取程序、数模转换模块控制程序等组成的软件系统,最终构成原理样机实验平台。将相同驱动参数控制下和协调控制策略控制下月球车的能耗作为实验结果。通过十条随机路径上的实验结果对比表明,运动协调控制模式可有效降低能耗,最高为5.4%,最低为0.6%。通过分析可能的误差原因,提出了实验的改进措施。针对理论研究和实验过程中发现的问题,提出了若干条可以改进和深入研究的方向,其中包括运动协调控制策略在其他摇臂式月球车中的应用、硬质地面假设条件到松软土壤实际状况的转换、配备随动轮的协调控制装置等。这些问题不仅为运动协调控制模式投入实践奠定了基础,也为今后的深入研究指明了方向。更多还原

【Abstract】 With the development of space technology and the strength of the national economy,Lunar exploration plan has been included in the agenda. Lunar terrain is complex and unkown, and energy on luanr rover is limited, so it’s key to Lunar exploration that making full use of limited energy. And decreasing energy consumption has become more important.To eliminate parasitic power losses of six-wheeled rocker lunar rover and lower energy consumption, a coordinated motion control mode was brought out, which solved from a coordinated motion programming model. The coordinated motion programming model, aiming at causes of parasitic power losses, was an constrained optimization model, whose state variables were attitude of lunar rover, planning variables were speed and torque of motors, objective function was energy consumption of six driving wheels, and constrains were kinematics model following theorem of projection of velocity, quasi-static balance force model, rated power of driving motor, driving-force coefficient of ground and max driving force of motor. The optimal solutions of the coordinated motion programming model, solved based on the Kuhn-Tucker conditions, was applied as the coordinated motion control strategy to the control on lunar rover.According to the passively deformation adaptability of the rocker suspension, a calculation on the attitudes of lunar rover based on terrain information was carried out. The driving parameters were calculated associated with the coordinated motion control mode, and energy consumption for any path was got. On this basis, aiming at limitation and one-sidedness of existing path programming system, a multi-attribute decision-making system was brought, whose attributes were energy consumption, time consumption, roughness of path and length of path. TOPSIS law was brought in this system and optimal path was picked out from several tranversibal paths. Simulation shows that the optimal path from the multi-attribute decision-making system considers to low-power, fast and smooth.As the extension of the multi-attribute decision-making system, Dynamic Programming and Multi-objective Optimization on path for lunar rover group, consist of two parallel wheeled rover and six-wheeled rocker rover, was brought out. After slave rovers moving forwards and backwards, path of master rover was obtained from Dynamic Programming and Multi-objective Optimization on path of slave rovers. Simulation shows that although time consumption increased, energy consumption reduced. To avoid error from inaccuracy of model and reduce reliance on the visual system of the control system, a Markov predictive control system combine with the coordinated motion control mode was researched, based on the Markov of terrain on lunar surface and the passively deformation adaptability of the rocker suspension. The terrain information of next state was predicted, so the attitude of rover and the coordinated driving control parameters were predicted. Feasibility of this system was verified by MATLAB simulation. For the pertinence to characters of research object, general model of conventional method was avoid and computational complexity was reduced.To determine the effectiveness of coordinated motion control mode, an experiment on prototype was carried. The experiment platform was foundated on the six-wheeled rocker prototype. Hardware system of platform was host computer, D/A converter, driving module, sensing module and power module. And software system of platform was control strategy program, senor reading program and D/A converter control program. The energy consumption under coordinated motion control strategy was measured as the experiment results, compared with the energy consumption under same control parameters for six wheels. The results on ten random paths show that energy consumption was reduced by the coordinated motion control strategy, up to 5.4 percent and down to 0.6 percent. Some improvements were brought by analyzing the possible causes of error.For some phenomenon in theoretical research and experiment, several improvement and future research were brought, such as application on other rocker rovers of coordinated motion control strategy, conversion from rigid ground to soft ground, and coordinated motion control equipment with follower wheel. These problems not only lay a foundation for application of coordinated motion control strategy, but also point out the directions for the future research.更多还原