机械工程学报
機械工程學報
궤계공정학보
CHINESE JOURNAL OF MECHANICAL ENGINEERING
2010年
1期
68-73
,共6页
球面3-RRR并联机构%动力学%拉格朗日方法%迭代学习控制
毬麵3-RRR併聯機構%動力學%拉格朗日方法%迭代學習控製
구면3-RRR병련궤구%동역학%랍격랑일방법%질대학습공제
3-RRR spherical parallel robot%Dynamics%Lagrange method%Iterative learning control
采用Lagrange方法建立球面3-RRR并联机构基于动平台姿态参数的动力学模型.考虑在计算过程中采用线密度、厚度忽略等近似计算及工作过程中机构的磨损等微小变化造成的参数不确定性,进一步建立带参数不确定性的动力学模型.针对其在振动测试中的重复性动作等特点及参数不确定性设计鲁棒-自适应迭代学习控制器,并对此控制器进行稳定性证明.该控制器利用自适应算法对未知定常数的强大学习能力来补偿系统动力学模型的参数不确定项;利用迭代学习算法对期望轨迹进行零误差重复跟踪.由于该控制器吸取了系统动力学模型的已知信息,避免了一般模型未知系统迭代控制时必须满足的Lipschitz约束条件.仿真结果表明,在此控制器的作用下球面3-RRR并联机构能够很好地重复跟踪期望轨迹.
採用Lagrange方法建立毬麵3-RRR併聯機構基于動平檯姿態參數的動力學模型.攷慮在計算過程中採用線密度、厚度忽略等近似計算及工作過程中機構的磨損等微小變化造成的參數不確定性,進一步建立帶參數不確定性的動力學模型.針對其在振動測試中的重複性動作等特點及參數不確定性設計魯棒-自適應迭代學習控製器,併對此控製器進行穩定性證明.該控製器利用自適應算法對未知定常數的彊大學習能力來補償繫統動力學模型的參數不確定項;利用迭代學習算法對期望軌跡進行零誤差重複跟蹤.由于該控製器吸取瞭繫統動力學模型的已知信息,避免瞭一般模型未知繫統迭代控製時必鬚滿足的Lipschitz約束條件.倣真結果錶明,在此控製器的作用下毬麵3-RRR併聯機構能夠很好地重複跟蹤期望軌跡.
채용Lagrange방법건립구면3-RRR병련궤구기우동평태자태삼수적동역학모형.고필재계산과정중채용선밀도、후도홀략등근사계산급공작과정중궤구적마손등미소변화조성적삼수불학정성,진일보건립대삼수불학정성적동역학모형.침대기재진동측시중적중복성동작등특점급삼수불학정성설계로봉-자괄응질대학습공제기,병대차공제기진행은정성증명.해공제기이용자괄응산법대미지정상수적강대학습능력래보상계통동역학모형적삼수불학정항;이용질대학습산법대기망궤적진행령오차중복근종.유우해공제기흡취료계통동역학모형적이지신식,피면료일반모형미지계통질대공제시필수만족적Lipschitz약속조건.방진결과표명,재차공제기적작용하구면3-RRR병련궤구능구흔호지중복근종기망궤적.
A dynamic model based on attitude parameters of moving platform for a 3-RRR spherical parallel robot is developed by using Lagrange method. Further, a dynamic model with parameter uncertainties is developed considering the parameters uncertainties caused by tiny changes such as approximately computing in linear density, thickness, and mechanism abrasion during working process. Aiming at its characteristic of acting repeatedly in vibration measuring and parameter uncertainties, a robust-adaptive iterative learning controller (ILC) is designed for this mechanism, and its stability is proved. The adaptive algorithm, which is powerful in leaning unknown constants, is used to compensate the uncertain parts of dynamics model, and the iterative learning algorithm is used to track the desired path without errors. Because of the utilization of certain information of dynamics model, the Lipschitz condition necessary for most unknown systems during iterative learning control is avoidable. The simulation results indicate that this robot can accurately track the desired path repeatedly.