采用概率密度分布和粒子滤波的室内跟踪算法

Indoor tracking algorithm via probability distribution and particle filter

提出了一种新的室内定位跟踪算法,采用了直方图法和核函数法估计参考点处的接收信号强度的概率分布,并将其作为该参考点的位置指纹,描述了该参考点处无线信道的特性;利用粒子滤波解决了非线性状态空间模型下的在线跟踪问题,仿真结果表明基于概率密度分布和粒子滤波的跟踪算法收敛速度快,且对环境变化不敏感,性能优于卡尔曼滤波算法。

This paper presents a new indoor location tracking algorithm,which uses signal strength probability distribution estimated by histogram method and kernel method to address the noisy wireless channel,and particle filter is used to deal with the nonlinear system model,which can approximate the optimal Bayesian estimate.Numerical simulation shows the new algorithm outperforms the tracking algorithm using Kalman filter.