【作者】 何海波；

【导师】 杨元喜；

【作者基本信息】 中国人民解放军信息工程大学 ， 大地测量学与测量工程， 2002， 博士

【摘要】 精密GPS动态测量能实时或事后地给用户提供高精度的位置、速度、加速度、姿态和时间信息，它在陆地、海面或航空中的导航和测量应用越来越广泛。本论文针对单基线动态测量模型，就短基线模糊度快速解算、中长距离模糊度解算、GPS动态测量(包括定位、测速和姿态测量)以及周跳检测与修复等质量控制问题展开研究。论文主要贡献有以下几点： 1．针对短基线模糊度的快速解算，提出了一种实时估计观测值方差-协方差矩阵的方法，改进了观测值的统计模型。算例显示这种方法能提高模糊度瞬时解算的成功率。 2．针对中长基线中电离层延迟的影响，将电离层加权模型引入了动态测量。在顾及电离层加权模型的情况下，利用参数消去法，推导了加权最小二乘法估计模糊度实数解的公式。 3．针对中长基线中的模糊度在线解算(OTF)，研究了电离层延迟、对流层延迟的影响及其相应的处理方法。研究结果表明，对流层模型改正可以大大提高模糊度解算的成功率和可靠性。将电离层延迟作为零均值的随机游走过程(电离层加权模型)，将对流层延迟作为静态参数，采用非递推形式的加权最小二乘法来估计，可以提高模糊度解算的成功率和可靠性。但是，如果将电离层延迟或对流层延迟作为随机过程，用递推形式的加权最小二乘法或Kalman滤波来处理，则会影响模糊度的解算。 4．针对模糊度在线解算问题，对加权最小二乘法和经典Kalman滤波这两种算法进行了综合分析和比较。结果显示，不论在模糊度解算的成功率和可靠性上，还是在计算速度上，加权最小二乘法都优于Kalman滤波。 5．提出了一种周跳检测与修复的新方法，即最小二乘搜索法。综合应用电离层残差法、伪距载波组合观测值法、最小二乘搜索法等方法，可修复失锁时间长达数分钟的周跳。 6．针对中长基线的动态定位，研究了电离层延迟、对流层延迟和多路径效应等误差对位置结果的影响以及相应的误差处理措施。结果表明，对流层模型改正可以大大改善定位结果的精度，不过仍存在未模型化的对流层延迟误差。将电离层延迟作为随机过程来处理，可以提高定位精度；而将对流层延迟作为随机过程来处理，则会影响定位精度。 7．针对高精度动态定位结果的精度，对最小二乘法和经典Kalman滤波这两种算法进行了综合分析和比较。算例显示，在高精度GPS动态测量中，最小二乘法可以提供厘米级精度的位置结果，而Kalman滤波算法不但不能提高定位结果的精度，反而会给定位结果引入米级的偏差。 8．针对观测值粗差和载体机动误差，提出了抗差自适应Kalman滤波的新算法，并对该算法作了进一步的改进。 9．详细分析了GPS测速中各种误差的影响，并对位置差分法、原始多普勒频移观测值法、载波相位中心差分法等三种测速方法进行了综合分析及比较。 10．讨论了GPS姿态测量中模糊度解算、姿态角解算等问题，并基于非专业姿态测量接收机，给出了GPS姿态测量的初步结果。更多还原

【Abstract】 Precise kinematic GPS surveying can provide users with the precise position,velocity,attitude and time information of the moving platform in real-time or post-processing modes. It is now increasingly used for many surveying and navigation applications on land,at sea or in the air. Precise kinematic GPS surveying in single-baseline mode is investigated with the emphasis on the (near) instantaneous ambiguity resolution for short-baselines,on the ambiguitiy resolution for long-baselines and on quality control. The main achievements of this paper are as follows:1. A real-time estimation of a prior variance-covariance of GPS observations is developed for the (near) instantaneous ambiguity resolution for short-baselines,which improves the stochastic model of the observations,and then the success rate and the reliability of ambiguity resolution.2. The ionosphere-weighted model is introduced into precise kinematic GPS positioning to deal with the ionosphere over long-baselines. Further,based on the method of pre-elimination of parameters,the formulae of the partly-weigthed least squares estimation with constraints on parts of the parameters are derived to include a prior information.3. The effects of the atmosphere errors on ambiguity resolution on-the-fly are investigated for long-baselines. The results show that the ionosphere-weighted model or the tropospheric estimation,integrated with the partly-weigthed least squares,can improve,the success rate and the reliability of ambiguity resolution;However,if the ionospheric delay or the tropospheric delay,which is modeled on random walk process or first-order Gauss-Markov process,is estimated with the Kalman filter,it will reduce the success rate and the reliability of ambiguity resolution. The results also show that Saastamoinen/Niell model can remove the most of the tropospheric delay,and then significantly improve the success rate and the reliability of ambiguity resolution.4. A comparision between the partly-weigthed least squares estimation and the Kalman filter ismade on ambiguity resolution on-the-fly,which shows that the partly-weigthed least squares estimation is much better than the Kalman filter for the success rate,the reliability and the computational efficiency of ambiguity resolution.5. A cycle-slip detection and correction technique is developed to detect and correct cycle slips in dual-frequency carrier phase,named as least-squares-search method for cycle slips,which can correct almost all cycle slips when data gap is shorter than a few minutes,with the help of the geometry-free phase combination and the pseudorange minus phase linear combination.6.The effects of the atmosphere errors on the kinematic positioning solutions are assessed for long-baselines with the fixed ambiguities. The results show that Saastamoinen/Niell model can remove the most of the tropospheric delay and then significantly improve the kinematic positioning solutions;the ionosphere which is modeled on random walk process can also improve the kinematic positioning solutions;however,the troposphere which is modeled on random walk process will bias the kinematic positioning solutions.7. A comparision between the partly-weigthed least-squares estimation and the Kalman filter is made on precise kinematic GPS positioning with the fixed ambiguities,which shows that thepositioning results from the partly-weigthed least squares,which accuracy is at the level of 10 cm,are much more accurate than the results from the Kalman filter,which accuracy is at the level of a few meters.8. A general estimator for adaptively robust filter is developed,which includes the estimators of classical Kalman filter,adaptive Kalman filter,robust filter,sequential least squares (LS) adjustment and robust sequential adjustment. The procedure can not only resist the influence of outlying kinematic model errors,but also control the effects of measurement outliers.9. The precise velocity can be determined respectively with precise positioning results,receiver raw Doppler measure更多还原