Real-time users cannot carry out real-time precise point positioning (RT-PPP) because they cannot receive real-time service (RTS) products from the international GPS service (IGS) in the case of interrupted communication. We address this issue by introducing a stable particle swarm optimized wavelet neural network (PSOWNN) to predict the short-term satellite clock bias in real time accurately. The predicted sequences of the new model are compared with those of the conventional linear polynomial, quadratic polynomial, gray system (GM (1,1)), and Kalman filter models. The results show that the accuracy of the proposed model is better than that of these four models. The average prediction accuracy of the 30-min and 60-min forecasting has improved by approximately (79.3, 82.4, 79.1, 97.4) % and (97.4, 82.9, 87.7, 98.9) % and is better than 0.3 ns during 30-min and 1-h forecasting. The RTS products can thus be replaced with the short-term clock products predicted by the PSOWNN model to meet the precision requirements of RT-PPP.

中文翻译：

---

具有稳定性能的卫星时钟偏差高精度短期预测方法

实时用户无法执行实时精确点定位（RT-PPP），因为在通信中断的情况下，他们无法从国际GPS服务（IGS）接收实时服务（RTS）产品。我们通过引入稳定的粒子群优化小波神经网络（PSOWNN）来解决这一问题，以实时准确地预测短期卫星时钟偏差。将新模型的预测序列与常规线性多项式，二次多项式，灰色系统（GM（1,1））和Kalman滤波器模型的预测序列进行比较。结果表明，所提模型的准确性优于这四个模型。30分钟和60分钟的预测的平均预测准确性提高了大约（79.3、82.4、79.1、97.4）％和（97.4、82.9、87.7、98.9）％，并且优于0。在30分钟和1小时的预测期间为3 ns。因此，可以用PSOWNN模型预测的短期时钟产品来代替RTS产品，以满足RT-PPP的精度要求。