数据驱动的舞蹈机器人设计  

Design of data-driven dance robots

作　　者：宋燕宇 陈定君 SONG Yanyu;CHEN Dingjun(Ankang University,Ankang Shaanxi 725000,China;Chongqing Vocational College of Public Transportation,Jiangjin District,Chongqing 402246,China)

机构地区：[1]安康学院,陕西安康725000 [2]重庆公共运输职业学院,重庆402246

出　　处：《自动化与仪器仪表》2023年第11期187-191,196,共6页Automation & Instrumentation

基　　金：陕西省艺术科学规划项目《古诗词在幼儿园舞蹈教学活动中的应用与研究》(2022HZ1684)。

摘　　要：针对市场上传统舞蹈机器人稳态误差较大、控制精度偏低的问题,提出以机器人手臂作为研究对象,利用数据驱动方法进行舞蹈机器人设计。在进行设计时,通过拉格朗日法进行运动学模型构建,该模型能够准确描述驱动力矩、重力、离心力、正定惯性矩阵四者的关系;构建带有积分功能的状态空间模型作为预测控制模型,该模型具有较高的控制精度,且积分功能的加入,能够有效降低驱动力和驱动力距计算时的稳态误差;利用改进的粒子群算法进行代价函数加权因子的寻优,能够快速且准确地确定代价函数加权因子最佳取值;最后通过仿真试验验证相关设计的性能。实验结果证明,设计的机器人手臂控制器具有较高数据计算能力和计算平稳性,能够实现对关节的紧密跟踪,对机器人运动突变的情况具有一定的应对能力,基本满足设计需求。In response to the problem of large steady-state error and low control accuracy of traditional dance robots in the market,a data-driven method is proposed to design dance robots using robot arms as the research object.In the design,the kinematics model is built by Lagrange method,which can accurately describe the relationship between driving torque,gravity,centrifugal force and positive inertia matrix;Construct a state space model with integration function as a predictive control model,which has high control accuracy.The addition of integration function can effectively reduce steady-state errors in driving force and driving force distance calculation;Using an improved particle swarm optimization algorithm to optimize the weighting factor of the cost function can quickly and accurately determine the optimal value of the weighting factor of the cost function;Finally,the performance of the relevant design was verified through simulation experiments.The experimental results prove that the designed robot arm controller has high data computing power and computational stability,can achieve close tracking of joints,and has certain adaptability to sudden changes in robot motion,basically meeting the design requirements.

关 键 词：动力学模型 预测控制模型 粒子群算法 

分 类 号：TP392[自动化与计算机技术—计算机应用技术]