摘要
为研究目前编组站调车机车实时定位性差的问题,采用TI公司生产的CC2531芯片,构成基于Zigbee技术的编组站机车定位系统;传统基于Zigbee定位采用无线信号损耗模型,周围环境对该模型的参数设置影响较大,现实模型参数设置往往取经验值,因此定位精度、普适性都不理想;在研究无线信号损耗模型基础上,采用隐含层为30的BP神经网络进行拟合RSSI与距离的关系,再用最小二乘法进行由距离到位置坐标的计算;实际实验结果表明,该算法明显提高了调车机车的定位精度,同时有较高的普适性,可以提高机车司机在推峰过程中对机车速度控制,也有利摘钩人员对摘钩位置的判断,提高了编组站机车的工作效率。
Abstract ： In order to research the problem of real-- time location of the marshalling yard shunting locomotives, constitutes a marshalling yard shunting locomotives positioning system for uncoupling based on the zigbee protocol using the CC2531 chip produced by TI company. Most traditional positioning based on zigbee uses signal loss model to locate. The work surroundings has an important impact on the set of pa rameters in the model. If we take experience value in reality, both the positioning accuracy and the universal of this method are not in ideal. Based on research the traditional model, a BP neural network useing the hidden layer of 30 predict the fitting relationship between RSSI and distance, and then convert distance into position coordinates through the least squares method on the basement of signal loss model. Actual experimental results show that the algorithm can not only improve the positioning accuracy, degree of universality significantly, also help lo- comotive driver to control speed, judge the position of excising the hooks for staff, improve the marshalling yard locomotive efficiency.
出处
《计算机测量与控制》
北大核心
2013年第11期3035-3037,共3页
Computer Measurement &Control
关键词
机车定位
ZIGBEE
接受信号强度指示
BP神经网络
最小二乘法
locomotive positioning
Zigbee
received signal strength indication （RSSI）
BP neural network
least squares mett
