Abstract: A new algorithm for satellite positioning under the circumstance of unfavorable geometry is proposed.GDOP，which is of great importance to the accuracy of positioning，will deteriorate when the relative geometry between satellites is bad，thus the variation of positioning will inflate.To counter this problem，we proposed a new biased estimator which is called rigid regression estimator.This algorithm improves the position accuracy by inducing biases in the estimator.It can provide position much accurately than the LMS estimation under bad geometry，so it has applicable value in military operation.

Key words: satellite positioning, rigid regression estimation, biased estimation, geometry dilution of positioning

摘要： 提出一种在可见卫星相对几何构型不好时提高定位精度的新算法。卫星间相对几何位置对定位精度的影响很大，在可见卫星几何布局不好的情况下，定位精度会大大降低。针对此问题，提出一种新的卫星定位算法。该算法通过引入有偏估计，以最小化参数估计的方差为目标，用轻微牺牲无偏性的代价大大提高了定位精度，解决了可见卫星布局不好情况下不能精确定位的问题，在军事上有很高的应用价值。

关键词: 卫星定位, 岭回归估计, 有偏估计, 精度衰减因子