中文摘要：

传统的基于最小二乘法和奇偶矢量的接收机自主完好性检测（Receiver Autonomous Integrity Monito-ring,RAIM）算法均是利用冗余观测量进行一致性检验实现的,其检测统计量均由伪距残差构造得到。本文采用接收机钟差作为RAIM算法的判据,基本思想是根据钟差历史解算值建立接收机钟差二次多项式模型,从而得到当前时刻的钟差预测值。同时解算当前时刻的钟差值,将这两者之差与钟差门限值进行比较来判断系统当前时刻是否存在故障。本文首先分析了伪距偏移量对接收机钟差的影响,然后给出确定钟差门限值的方法。结合用户解算位置和接收机钟差利用子集比较法实现了故障识别的功能。因本文给出的故障检测算法不是基于冗余观测量的,故该方法在四颗可见星下就可进行故障检测,与传统的RAIM算法相比降低了RAIM算法对可见星个数的要求,该点在实验部分进行了仿真验证。另外还对可用星多于四颗的基于接收机钟差的RAIM算法进行了故障检测和故障识别部分的仿真,验证了本文方法的可行性。

英文摘要：

The conventional RAIM algorithms based on least square method and parity vectors both utilize pseudo-range residuals to develop test statistics.To use these methods,redundant pseudo-range measurements are necessary.Different from the previous algorithms,the clock bias is selected as a criterion of the RAIM algorithm which does not need redundant pseudo-range measurements.An analysis of the impact of pseudo-range bias vector on receiver clock bias is provided on the basis of the proposed method.And then a method for calculating the threshold value of clock bias is given.According to the solved values of clock bias over a past period of time,a quadratic function model is established to get the value of predicted clock bias in current time.On the other hand,the estimated value of clock bias in current time is obtained from the measurements in current time.After the difference between the predicted value and the estimated value of clock bias in current time is got,a comparison is made between the difference and the threshold value of clock bias to accomplish the function of fault detection.The values of user position together with that of receiver clock bias are used to carry out fault recognition by means of comparing among subsets of measurements.When there are only four satellites available,the above method can still be used for fault detection.And fault recognition can be fulfilled when only five satellites available,thus effectively freeing an additional measurement and increaseing availability.Simulations have been carried to testify the feasibility of the proposed algorithm and analysis of its performance are also provided.