Computer Engineering and Applications ›› 2009, Vol. 45 ›› Issue (24): 225-227.DOI: 10.3778/j.issn.1002-8331.2009.24.068

• 工程与应用 • Previous Articles     Next Articles

Interacting multiple model algorithm applied to robot target tracking

YUAN Zhu,YAN Bao-ding
  

  1. College of Electronic Information Engineering,Henan University of Science and Technology,Luoyang,Henan 471003,China
  • Received:2008-05-06 Revised:2008-07-28 Online:2009-08-21 Published:2009-08-21
  • Contact: YUAN Zhu

交互式多模型算法在机器人目标跟踪中的应用

袁 铸,阎保定   

  1. 河南科技大学 电子信息工程学院,河南 洛阳 471003
  • 通讯作者: 袁 铸

Abstract: The Interacting Multiple Model(IMM) algorithm is used to solve the maneuvering targets tracking problem for visual servoing robot.This paper uses Constant Velocity(CV) model and Constant Acceleration(CA) model to match the target’s two motion states,uses Markov chains to change the models,and gives the predicted value of the target’s motion state based on the former motion state and the current observation.The Monte Carlo simulation is carried out using Matlab software,which is based on IMM algorithm and Kalman Filter algorithm,the results show that the IMM algorithm can track the target more accurately than the Kalman Filter,no matter the target is at the uniform motion state or the maneuvering motion state.These prove that the IMM algorithm can improve the tracking precision of maneuvering target.

Key words: Interacting Multiple Model(IMM) algorithm, Kalman filter, maneuvering targets tracking, visual servoing robots

摘要: 将交互式多模型(IMM)算法应用于视觉伺服机器人对机动目标的跟踪。使用匀速运动(CV)和匀加速运动(CA)模型表示目标的两种运动状态,利用马尔可夫链进行模型切换,根据目标前一时刻的状态和当前的观测值,预测目标当前的状态。在Matlab上对IMM滤波算法和Kalman滤波算法进行了仿真实验研究,结果表明,不管目标处于何种运动状态,IMM算法估计量的误差均值都比Kalman滤波算法的误差均值小,尤以目标作机动运动时更为突出,证明了应用IMM算法可以提高跟踪机动目标的精度。

关键词: 交互式多模型算法, 卡尔曼滤波, 机动目标跟踪, 视觉伺服机器人

CLC Number: