Automatic Train Control Algorithms with Regulation Restrictions Adaptive to System State Changes
Abstract
Centralized automatic train control algorithm is proposed which allows effective adjustment of each train’s departure and running times or intervals when there are unscheduled delays. The novelty lies in that it employs three sets of methods, called regulative characteristics, which contain relationships between regulation restrictions (minimal station dwell times and train running times). These relationships are obtained a priori from the simulation of train circulation on the detailed railway or metro line model, saving later onboard real-time processing requirements, which is especially useful for those train control systems which are being only partially modernized. In the traditional algorithms these regulation restrictions are frequently considered fixed, while the changes of system’s state do affect them, so the usage of the values recurrently updated by the regulative characteristics allows to have more flexible running and station times for every train on the line. The efficiency of the new algorithm is additionally increased by the usage of forecasting of station dwell times and train running times based on the history of these values for each train or station. Proposed algorithm decreases the quantity of undesired stops between stations, allowing to save energetic resources spent on repeated train acceleration and to raise, in general, the quality of the transportation process algorithm has been tested on Mexico Metro Line simulator confirming efficiency increase of 3% for the worst case scenario and up to 10% for the best cases (busiest state of the line plus additional temporal speed limits).
Keywords
Download Options
Introduction
Centralized Automated Train Control (ATC) together with Automatic Train Operation (ATO) systems allow modern railways to increment traffic density in order to satisfy the raising passengers affluence while maintaining the desired quality of the transportation process in urban and suburban rapid railways. While the problem of rising passenger affluence in all kind of urban and suburban transport is being addressed from different points of view [1-7, 12-22] over the years, many efforts were spent on the subject of automation of the train control [1, 8-17]. In the core of any ATC-ATO system lies a family of algorithms which operates mainly with the values of train running and station well times. Normally [1, 13-25], there is at two level hierarchy present: the higher level calculates the timing difference between the desired and executed scenario (which is normally based on a schedule, an interval or both). The timing difference is caused by per turbations. If the per turbation is considered compensable, the higher level calculates the operating values (running and station times) for each train on the line. If the perturbation is non compensable, new scenario (schedule or interval or both) is generated. The lower level then executes the control commands generated by the higher level. The way in which the higher level calculates the operating times depends on a strategy concerning preferential usage of available timing resources (station time resource preference, running time resource preference or hybrid) and on algorithm type: there are algorithms based on schedule which operate with astronomical times, and algorithms based on intervals that use relative times, while hybrid algorithms use both [1, 10]. AllATC-ATO algorithms have basic regulation restrictions, such as minimal station well time (or minimal station time), minimal interval departure interval and minimal running time. Their absolute minimal values are dictated by line and rolling stock physical parameters and traditionally considered as fixed. However, these values can be considered as restrictions which depend on current system state, meaning that the changes of departure intervals and running times of all the trains in the line caused by perturbations affect them (moving them towards more restrictive values). In this article we present the algorithm which considers the modifications in the regulation restrictions caused by system state changes based on method, also presented in the article, of regulative characteristics of the station-to-station blocks, where the regulative characteristics are families of curves that relate train control parameters and are obtained a priority on the railway line simulator.
Conclusion
Results and methods conceived are mostly applicable for metro lines and can be kept for suburban railway lines with the stations located close to each other, so there is small deviation in running times and there are small differences in service patterns so that the dwell time can be forecast.
The core block of the new algorithm is based on the regulative characteristics, which are calculated a priori on the line simulator, while the forecasting only improves its efficiency in the conditions of highly dense traffic (that is when sequential perturbations are most likely correlated between each other). It might be noted that the usage of other forecasting techniques [2, 22, 28] could be adopted and studied in order to work within regulative characteristics methodology. In any case, following the presented methodology, the great majority of calculations are performed only once every time there is a constructional update on some station-to-station block, for which the regulative characteristics must be also updated, because the simulation parameters would change. This leaves few calculation cycles to be performed ―on the go‖ and makes the new algorithm particularly suitable for real-time regulation especially on outdated control systems with no great calculationpower onboard, which is the popular case of partial modernization. On the other hand, energy saving and general transportation process quality improvement makes presented algorithms recommendable for practical usage.