Abstract: In order to resolve the nonlinear optimization problem in the traditional TDOA-based location, this paper firstly turns it to a problem of searching peak value by using the max likelihood estimation and then presents a new improved immune algorithm based on the clone selection theory. The new algorithm adopts a float point encoding mode to improve efficiency of the algorithm and uses the Gauss mutation which is controlled by the affinity to improve the ability of searching part of area. Numerical simulations show that this new algorithm has higher accuracy and better performance in global searching.

Key words: Time Difference of Arrival（TDOA）, Gauss mutation, clone selection, Maximum Likelihood Estimation（MLE）

摘要： 为解决传统TDOA 定位估计所带来的非线性优化问题，首先通过最大似然估计将其转换为峰值搜索优化问题，再提出一种基于克隆选择思想的改进型免疫算法对其进行求解。该算法采用浮点数编码方式，提高了运算效率；引入高斯变异算子和变异控制变量，加强了局部搜索能力。仿真实验表明，在保证一定抗体数目的前提下，该算法适应性强，性能稳定，能快速逼近全局最优的解，且算法定位精度更高。

关键词: 到达时间差定位法, 高斯变异, 克隆选择, 最大似然估计