论文部分内容阅读
根据城轨列车制动特性,对列车运动方程做2项简化。一是忽略空气阻力和坡度的影响;二是分别假设制动力传递有延时及减速度不变或减速度是初始速度的线性函数。由此推导出2个简化的列车停车误差估计模型和模型参数之间的线性关系函数式,并给出模型参数在线学习算法,以克服停车过程中的各种非线性因素的影响,提高停车精度。根据统计学原理,采用5个评价指标对模型的性能进行评价,采用停车误差估计判断停车精度是否满足停车可靠性的要求。利用实测停车数据对模型和在线学习算法进行验证和比较。结果表明:提出的简化模型和在线学习算法,能有效降低停车误差,并纠正误差分布的有偏性;停车误差在大于99.5%的情况下满足30cm停车精度的可靠性要求;模型1的效果比模型2略好。
According to the braking characteristics of the urban rail train, the train’s motion equation is simplified by 2 steps. One is to ignore the influence of air resistance and slope; the other is to assume that the transmission of braking force has a delay and deceleration, respectively, or deceleration is a linear function of the initial velocity. This paper deduces two simplified train parking error estimation model and the linear relationship between the model parameters and gives the model parameters on-line learning algorithm to overcome the various non-linear factors in the parking process to improve the parking accuracy . According to the principle of statistics, five evaluation indexes are used to evaluate the performance of the model, and the parking error estimation is used to judge whether the parking accuracy meets the requirements of parking reliability. Use the measured parking data to verify and compare the model and online learning algorithm. The results show that the proposed simplified model and on-line learning algorithm can effectively reduce the parking error and correct the bias of the error distribution. The parking error of more than 99.5% meets the reliability requirement of 30cm parking accuracy. The effect ratio of model 1 Model 2 slightly better.