基于雅可比矩阵的仿生视觉运动模型研究

Research on Motion Model of Humanoid Vision System Based on Jacobi Matrix

针对仿生视觉系统在目标跟踪过程中眼、颈角度分解对精度的影响问题,提出了以速度作为参变量的雅克比矩阵角度分解最优化方法。首先,分析人的眼、颈运动规律,建立眼、颈2级4自由度系统模型;其次,构建雅克比矩阵,得到关于转动速度的眼、颈协调运动的角度分解数学模型;最后,通过仿真和物理试验分析各自由度转动角速度在最优化条件下对角度分解的影响,得到了基于所述系统的目标跟踪角度分解最优化方案。试验结果表明,本文所述仿生视觉运动控制平台的目标定位精度在70%以上。

An optimized angle decomposition method based on Jacobi matrix with velocity as the parameter is put forward to decrease the angle decomposition effect on system accuracy.Firstly,by analyzing the motion rule of human eye and neck,the two-stage four-degree′coordinate is established.Secondly,based on Jacobi matrix,the mathematical model of tracking angle decomposition of eye and neck is built.Lastly,by simulation and experiments,the influence caused by different rotation angular velocity on angle decomposition is studied under the condition of optimization,resulting in the optimized angle decomposition method.The experimental results show that the target tracking accuracy of system mentioned in this paper is above 70%.