基于LS-PSO的北斗伪距单点定位算法

Compass Pseudorange Single-point Positioning Algorithm Based on LS-PSO

北斗伪距单点定位具有易于实现、不存在整周模糊度、速度快等特点,具有很大的研究和应用价值;传统最小二乘法由于引入了线性误差、对初始值依赖性强而导致定位精度低;为了提高北斗伪距单点定位的精度,通过分析最小二乘法和粒子群算法的优缺点,提出了一种LS-PSO组合算法;首先利用最小二乘法定位计算接收机的大约位置,作为粒子群算法解的基准值并建立解的搜索空间,然后利用粒子群算法得到全局最优值,解算出精度更高的结果;经过实验验证,LS-PSO组合算法可以稳定的解算出m级精度的定位结果,并且三维方向偏差都在大约5m以内;最后通过与遗传算法的收敛情况和最小二乘法的定位精度进行对比,证明LS-PSO组合算法可以快速的收敛到最优解并且有效的提高了北斗伪距单点定位精度。

Compass pseudorange single-point has characteristics of easy to implement, without ambiguity problem, postioning fast. It has great value for research and application. Traditional least squares algorithm has linear error and strong dependence on initial values so that its accuracy of positioning is very low. In order to improve the accuracy of Compass pseudorange single-point positioning, proposed a combination algorithm of LS-PSO by analyzing the advantages and disadvantages of least squares algorithm （LS） and particle swarm optimization （PSO）. Firstly it obtains the approximate location of receiver for the base value of PSO by LS and establishes the search space of the solution. Then it calculates the global optimum by PSO so that the result has higher accuracy. Through the verification of experiments, LS- PSO can calculate m-level precision positioning results stably and the three-dimensional errors are all within 5 meters. Finally by compared with the positioning accuracy of LS and the convergence of genetic algorithm, it proves that LS-PSO can converge to the optimal solution quickly and improve the accuracy of Compass pseudorange single-point positioning effectively.