摘要
针对卫星定位容易丢失信号的问题,提出北斗卫星导航技术与GSM-R技术相结合的定位方法。根据哈尔滨西至长春铁路线的实测数据,利用BP神经网络构造BDS/GSM-R定位信息的融合模型,采用遗传算法优化BP神经的权值和阈值,对比优化前和优化后的BDS/GSM-R组合定位方法。实验结果表明:优化后的BP神经网络训练结果与列车实际运行轨迹偏差更小,东向、北向、方位角定位误差的波动范围明显减小,可以保持列车定位的连续性和精确性。
Aiming at the problem that satellite positioning signal is easily lost, the positioning method that com-bine BeiDou satellite navigation technology with GSM - R technology was put forward. According to the measureddata from the Harbin West railway station to Changchun, onethe BDS/GSM -R positioning information fusionmodel was firstly built using BP neural network structure. The BP neural weights and thresholds were optimizedusing genetic algorithm. Finally, the results before and after optimization of combined BDS/GSM - R positioningmethod was compared. The results show that the difference between the training results of BP neural networkafter optimization and the practical operation tracks of train is reduced. The range of error located in the east andnorthoriention and azimuth error are obviously decreased. This method can keep the continuity and precision oftrain positioning.
出处
《铁道科学与工程学报》
CAS
CSCD
北大核心
2016年第3期552-556,共5页
Journal of Railway Science and Engineering
基金
国家自然科学基金资助项目(6147080)