基于脉冲神经网络的数字车辆数据处理研究OA
Research on Digital Vehicle Data Processing Based on Spiking Neural Network
针对数字车辆综合信息管理平台中多源实时数据处理资源消耗大、响应延迟高等问题,提出了一种基于脉冲神经网络的处理方法.通过脉冲编码将来自车速、位置、温度等多种传感器的实时数据转化为脉冲信号,结合脉冲时序依赖可塑性(STDP)学习规则与多层网络架构,实现对时序数据的深度挖掘.实验结果表明,该模型处理实时数据流准确率达 92.5%,响应时间为 35 ms,能够有效满足铁路运输系统中大规模实时数据流的处理需求.
To address the high resource consumption and long response latency of multi-source real-time data processing in the digital vehicle integrated information management platform,a processing meth-od based on spiking neural networks is proposed.Real-time data from various sensors,including vehicle speed,position,and temperature,are first converted into spike trains through spike encoding.The Spike-Timing Dependent Plasticity(STDP)learning rule is then combined with a multi-layer network architec-ture to enable deep mining of temporal data.Experimental results demonstrate that the proposed model performs effectively in processing real-time data streams,achieving an accuracy of 92.5%and a response time of 35 ms,thereby effectively satisfying the demands of large-scale real-time data stream processing in railway transportation systems.
吴乐华;于东洋
中国铁路南昌局集团有限公司,江西 南昌 330009北京京天威科技发展有限公司,北京 100080
信息技术与安全科学
铁路数字化管理数字车辆脉冲神经网络实时数据流处理脉冲编码
railway digital managementdigital vehiclespiking neural networksreal-time data stream processingspike coding
《机械与电子》 2026 (8)
18-25,8
国铁集团科技创新项目(CRNCJ2024073)
评论