基于测距的蒙特卡罗盒传感网定位算法研究

针对以蒙特卡罗算法为基础的无线传感器网定位算法普遍存在定位精度和采样效率低的问题，提出了一种基于测距的蒙特卡罗盒（R-MCB）定位算法。通过测距信息构造修正的包含有约束条件的方形边界框，使用从强约束条件中除去弱约束条件的启发法来提高采样效率，然后进行样本过滤和加权处理，并通过校准减少距离误差实现精确的定位。该R-MCB定位算法允许节点是静止或移动的，并且能够与可进行测距的节点和没有测距能力的节点协同工作。通过在传感器硬件上进行真实模拟定位算法证明，在多数情况下该R-MCB算法的定位误差，均要比WMCL算法（加权蒙特卡罗定位算法）的定位误差低10%左右。

Range-based Monte Carlo boxed localization algorithm for sensor network

A Range-based Monte Carlo Boxed （R-MCB） localization algorithm is proposed to solve the common problems in Monte Carlo algorithm-based localization algorithm for wireless sensor network, namely low localization accuracy and sampling efficiency. This paper constructs a corrected square bounding box containing constraint conditions by using ranging information, improves the sampling efficiency by heuristic method, namely to eliminate weak constraint conditions from strong constraint conditions, conducts sample filtering and weighing, and then reduces distance error to realize accurate localization through calibration. The R-MCB localization algorithm allows nodes to be static or mobile and it can work with nodes that can perform ranging as well as nodes that lack ranging capabilities. It is proved by really simulating the localization algorithm on sensor hardware that R-MCB is better than the range-free algorithm called Weighted Monte Carlo Localization （WMCL） in terms of localization error under many circumstances.