基于无线传感器网络的月球车定位方法及其滤波算法分析

【作者】 朱铭；

【导师】 刘济林； 陈国斌；

【作者基本信息】 浙江大学 ， 信息与通信工程， 2007， 硕士

【摘要】 月球的未知环境要求月球车能够稳定且精确地定位，这是研制月球车过程中需要解决的关键问题。本文提出了一种由中心计算机、月球车和无线传感器网络构成的分布式星球车定位系统框架。这种系统能够提供稳定的全天候定位服务，而且可以减小移动设备在复杂的计算时开销。同时，还详尽介绍了无线传感器网络测距的定位的原理，并提出了网络中的无线传感器节点和月球车定位算法。为了实现实时而准确地运用已知无线传感器网络定位，我们结合粒子滤波和卡尔曼滤波的各自特点，提出了粒子滤波-卡尔曼滤波月球车导航的定位算法。该滤波算法使用粒子滤波获取车体初始位置估计，之后使用扩展卡尔曼滤波继续跟踪星球车。当星球车遇到短暂无信号等情况时，该算法会根据系统状态在两种滤波算法中切换。仿真结果表明我们提出的定位算法具有很高的精度和稳定性以及较低的计算复杂度。更多还原

【Abstract】 The unknown environment of the moon in outer space requires that moon rover localize itself accurately, and the localization method should be robust. This is the key problem in developing the moon rover.A distributed moon rover localization system framework is introduced in this paper. The framework is composed of central computer, moon rover, and wireless sensor network. This system can not only provide robust localization service regardless of the weather, but it can also minimize the computing load when rover moves. We also analyze the ranging principles used in wireless sensor network. The localization system will localize sensors and moon rover in the network according to these principles.On basis of this framework, we propose PAKF (Particle - Extended Kalman Filter), a novel localization algorithm. PAKF has two modules: it first uses particle filter to acquire estimates of rover’s position; then it randomly samples estimates as the initial estimate of extended Kalman filtering. When sensor silence and other emergency events happen, it will switch between the two filtering modules. The experiment results illustrate that PAKF is a robust algorithm with high accuracy and computing efficiency.更多还原