论文部分内容阅读
针对大规模无线传感器网络(WSN)定位算法普遍存在时间复杂度过高的问题,实现了WSN邻近节点间逐对“比较关系”矩阵到位置坐标的快速可视化映射.算法首先引进快速映射(FastMap)计算过程,把参考节点作为定位的轴点,选择距离最长的对角线作为轴线,避免了相对坐标到绝对坐标的转换过程;将FastMap运算的概略坐标作为MDS(multi-dimensional scaling)的输入,提高了定位精度.在MATLAB软件中设置600m×600m的定位区域,利用无线信号衰减模型产生虚拟测试点,分别针对包含3 600,1 600,900,576,400个节点的无线传感器网络进行仿真实验.结果表明:与随机型和经典MDS算法相比,所提出的算法在保持高的定位精度的前提下,大大降低了时间复杂度.算法被应用于智能超市导购系统,21辆购物车的平均定位误差为0.158 5m.
In order to solve the problem of large time complexity of large-scale WSN location algorithm, a fast visual mapping between node-by-pair “comparison relation ” matrices to location coordinates of neighboring nodes of WSN is implemented.The algorithm firstly introduces fast mapping FastMap), the reference node is taken as the pivot point of the positioning, the diagonal line with the longest distance is selected as the axis, avoiding the conversion process from relative coordinate to absolute coordinate; the general coordinate of FastMap operation is taken as MDS (multi-dimensional scaling) Of the input to improve the positioning accuracy.In the MATLAB software to set 600m × 600m positioning area, the use of wireless signal attenuation model to generate virtual test points, respectively, for 3 600,1 600,900,576,400 nodes of the wireless sensor network simulation experiment results show : Compared with the random and classical MDS algorithms, the proposed algorithm can reduce the time complexity greatly while maintaining the high positioning accuracy.The algorithm is applied to the intelligent supermarket shopping guide system, the average positioning error of 21 shopping carts is 0.158 5m.