仿人机器人头部设计与目标跟踪运动控制

发布时间：2018-04-03 05:21

本文选题：仿人机器人　切入点：头部　出处：《农业机械学报》2017年10期

【摘要】：针对仿人机器人视觉目标跟踪任务需求,设计了一种高精度、灵巧型仿人机器人头部系统,构建了视觉感知模块与3自由度机械臂运动机构。利用虚拟连杆方法对仿人机器人头部进行运动学建模,将目标跟踪问题转换为机械臂的运动学逆解问题,通过计算机构可操作度与条件数,显示机器人头部对目标跟踪任务具有良好的运动学灵活性。在此基础上,基于梯度投影法建立了目标跟踪运动学逆解与规避关节极限位置的同步优化算式,并针对视觉测量系统带宽较窄、容易造成较大跟踪误差的问题,提出了在两次视觉测量间隔内用轨迹预测的方法获得目标位置的估计值,以提高跟踪精度。仿真与实验结果表明,在测量盲区内采用轨迹预测方法,可将跟踪精度提高约80%,实现高精度目标跟踪。所提出的仿人机器人头部及控制方法,对动态目标具有良好的跟踪性能。
[Abstract]:A high precision and dexterous humanoid robot head system is designed to meet the task requirement of humanoid robot visual target tracking. A visual perception module and a 3-DOF manipulator motion mechanism are constructed.The kinematics modeling of humanoid robot head is carried out by using virtual connecting rod method. The target tracking problem is transformed into the inverse kinematic solution problem of the manipulator. The operational degree and the condition number of the mechanism are calculated.It shows that the robot head has good kinematics flexibility for target tracking task.On this basis, based on the gradient projection method, a synchronous optimization formula of inverse kinematics solution of target tracking and avoiding the limit position of joint is established, and aiming at the problem that the bandwidth of vision measurement system is narrow, it is easy to cause large tracking error.In order to improve the tracking accuracy, a trajectory prediction method is proposed to estimate the target position in the interval between two visual measurements.The simulation and experimental results show that the tracking accuracy can be improved about 80% by using the trajectory prediction method in the blind area of measurement, and high precision target tracking can be realized.The proposed head and control methods of humanoid robot have good tracking performance for dynamic targets.
【作者单位】：中国科学院沈阳自动化研究所机器人学国家重点实验室;中国科学院大学;
【基金】：国家自然科学基金项目(51605474)
【分类号】：TP242

【相似文献】