Abstract: Due to limitations of Monte Carlo localization algorithm, such as low sampling efficiency and big sampling number, LAASO algorithm is proposed based on adaptive sampling optimization. Anchor box and prediction area are used to further optimize sampling area. Sampling number is adaptive defined by sampling area. Curve fitting in SOMCL algorithm is take to optimize the weight of samples. Simulation test results show, in the condition that speed change ratio is 25 m/s and the maximum speed is less than 60 m/s, location accuracy of nodes is respectively increased by 40% and 36% than that of MCL and SOMCL, while sampling number is decreased by 20% and 31.5% compared with that of MCL and SOMCL. LAASO is more suitable for high operation environment.

Key words: wireless sensor networks, Monte Carlo, localization, samping optimization, adaptive

摘要： 针对蒙特卡罗定位算法采样效率低和采样次数多等缺陷，在SOMCL算法的基础上提出一种基于自适应采样优化的定位算法LAASO。该算法采用锚盒子与预测区域进一步优化采样区域，通过采样区域的大小自适应确定样本数目，利用SOMCL算法中的曲线拟合对样本权值进行优化。仿真测试表明，当速度变化率为25 m/s，且最大速度小于60 m/s时，相比MCL算法和SOMCL算法，LAASO算法定位精度分别提高了40%和36%，采样次数分别降低为20%和31.5%，且更适应于高速运行环境。

关键词: 无线传感器网络, 蒙特卡罗, 定位, 采样优化, 自适应