Improvement of dispatchers' cognitive ergonomics in central rail suburban traffic control using decision-support system.

Haramina, Hrvoje ; Brabec, Dean ; Grgurevic, Ivan 等

1. INTRODUCTION

The central rail suburban traffic control is an extremely complex process. It requires great experience and full concentration on the work of the dispatcher who monitors the events in rail traffic and based on this makes timely decisions on further traffic continuation. Based on such decisions the train control system acts on trains and thus determines the planned method of their movement. Regarding the density of train traffic, and its influence on the dispatcher's workload, the rail track network has been divided into a number of control areas so that the work scope of a single dispatcher covers a single control area. The dispatcher's task is to make decisions in order to optimize the train traffic, with the aim of maximum realization of the planned schedule. Based on the dispatcher's decisions regarding the train traffic sequence, undesired events may be avoided, that may cause failure to realize the planned and published train schedule. Apart from some unforeseen factors that may influence the regular rail traffic flow, such as weather conditions, traffic accidents, and various malfunctions on the stationary and mobile rail plants. Most of these disturbances cause both directly and indirectly delays in the railway traffic. There are many different factors causing disturbances, most of them are of technical and organizational art. The delays caused by a certain disturbance can be divided into primary and secondary delays. The primary delays are a direct effect of the actual disturbance, while the secondary are the delays of the trains that interact with the primary delayed trains. The train dispatcher, who from a train traffic control centre plans, supervises, and controls the movements of the trains cause only a diminutive part of the total quantity of primary delays, but has a key role regarding the secondary delays. This phenomenon occurs at the moment when, due to the operation of a certain train, the parallel operation of other trains according to in-advance planned schedule is prevented, because they claim at the same time period, the same line section, which regarding safety reasons leaves the possibility of parallel operation of only one of the trains. This may lead to the delay of the mentioned trains, which in turn may cause continued propagation of this phenomenon to other trains in the network Mazzarello, M.; Ottaviani, E.(2005). By making changes in the original plan (timetable) the dispatcher tries to handle the effects of the disturbances, i.e. by changing the times and locations for the train meetings and the train overtakes he tries to minimize the secondary delays. This re-planning process is a demanding task due to the great complexity of the system. (Sandblad et al., 2003).

Modern solutions of the train traffic control system allow timely collection of a large number of data on the traffic condition that the dispatcher may use in a very useful way to make decisions regarding further method of traffic control. Today, dispatchers have essentially reached the limits of maintaining rail network productivity and quality on heavily used dense railroad networks essentially through human judgment and minimized interference. In the future, new more advanced traffic control systems will be necessary to develop, implement and communicate new real-time schedules quickly and accurately. Regarding the relatively large number of the provided process data, and sometimes very complicated traffic situations that need to be solved in such a way as to foresee with maximum precision the future development of the traffic events, there is increasing emphasis on the need to apply the expert systems in order to improve the dispatcher's cognitive ergonomics in the decision-making process during rail traffic control. (Luethi et al., 2006). In the planning process of this system it is very important to define a man-machine interface concerning e.g. workload, situation awareness and automated cognitive processes, limitations in human memory capacity, cognitive work environment problems, human error performance and dynamic decision processes. (Wikstrom et al., 2006).

2. MODEL OF DECISION SUPPORT SYSTEM IN THE CENTRAL RAIL SUBURBAN TRAFFIC CONTROL PROCESS

Railroad dispatchers play the key role in the rescheduling process. They are responsible for developing new routings for delayed trains and changing the train sequencing as well as communicating this information to the actors. However, dispatchers can only make optimal decisions quickly when they have a good overview of the actual state of the network, accurate information and a good forecasting tool. The dispatcher who controls traffic centrally often encounters challenges which affect the reliability of realizing the planned schedule. These challenges refer to the traffic situations in which trains operate in a way that had not been planned, which may lead to serious consequences regarding reduced efficiency of the rail traffic as well as the quality of providing services to rail users (Rebreyend, 2006). It is very important, namely, during the traffic control to make timely and valid decisions regarding the methods of the implementation because this may have significant influence on the solution of the mentioned situations. It should be noted that the quality and speed of making such decisions depends greatly on the workload and the ergonomic conditions of the dispatcher (Sandblad et al., 2003). This is precisely the reason why there is need to introduce a support system that would help the dispatcher in making such decisions. This paper defines the dispatcher support model that may help in solving one of the basic problems encountered by the dispatcher during the rail traffic control process. The support system facilitates the dispatcher's work regarding making the decision about the extent to which it is suitable to carry out the technological procedure of overtaking of certain trains on a single-track railway line. The dispatcher has to decide, namely, which of the two trains operating in the same direction along a single-track line (the preceding or the consecutive one) should be given priority in leaving the next traffic place of work at which the order of their operation may be changed.

The final decision regarding which train is to leave first the respective traffic place of work, i.e. whether there will be change in the sequence of the mentioned trains, is made by the dispatcher independently or based on the support system proposal. In case when the support system is not applied in decision making, such a decision is partly based on the in-advance defined and regulated rules, and partly on dispatcher's own intuition, whose efficiency is based on many years of experience. Based on such experience of the dispatcher, it is possible to create rules in the expert system knowledge base that may represent a significant support in decision making with the aim of reducing the workload and increasing the efficiency of the dispatcher's work.

The support system model means the collection of data about the condition in traffic that are relevant for making such a decision. The collection can be performed in several ways depending on the level of equipment regarding technical means available to the railways for this purpose (Schaefer & Pferdmenges, 1994). The modern approach in this respect is the application of the information equipment onboard train which collects all the relevant data about its condition that enter the base of the facts of the decision support system (e.g. European Train Control System--ETCS). Besides, this approach understands also the implementation of the wireless radio transmission of the data to the control centre at which the decisions are made regarding traffic control (e.g. Global System for Mobile Communications--Railway--GSM-R). The mentioned model is presented in the block scheme in the Figure.

[FIGURE 1 OMITTED]

Fuzzy logic is used for the deduction needs of the dispatcher support system in railway traffic control. The model understands the fuzzy deduction system which represents the core of the decision support system, which is designed in the module of Fuzzy Logic Toolbox of the MATLAB software package. The mentioned expert system model represents the possible tools for the solving of the described problem in rail traffic control. Through its implementation the dispatcher may obtain proposals regarding the solutions in the decision making about the future method of train traffic operation.

The implementation of fuzzy logic in the dispatcher decision support system in the process of central rail traffic control provides certain advantages in relation to the application of the classical expert systems. Such systems, namely, have fewer rules and are therefore easier to maintain and are closer to the human, approximate method of thinking, being thus easier to understand which results in easier and more reliable collection and usage of knowledge.

3. CONCLUSION

Centralized train and traffic control should be supported by efficient user interfaces that allow the train dispatcher to be continuously updated and able to evaluate future traffic conflicts so that these can be taken care of in time. By applying the described model of decision support system, the dispatcher can at the right moment, i.e. sufficiently fast, based on the facts collected from the real-time traffic process, which do not have to be precisely determined, make the most acceptable decisions in the rail traffic control process. In this way the dispatcher can focus more on the tasks in train traffic coordination between their own and adjacent control areas. This opens up the possibility of improving the efficiency of train traffic control in a wider area of the rail network as well as the operative rail traffic control as a whole.

4. REFERENCES

Mazzarello, M. & Ottaviani, E. (2005). A Traffic Management System for Real-Time Traffic Optimisation in Railways, Available from: http://www.sciencedirect.com Accessed: 2008-06-11

Rebreyend, P.; (2006). A Dispatching Tool for Railway Transportation, Available from: http://patat06.muni.cz/doc/patat06_088.pdf Accessed: 2009-04-18

Sandblad, B.; Andersson, A.W.; Bystrom, J. & Kauppi, A. (2003). New control strategies and user interfaces for train traffic control, Available from: http://www.it.uu.se/research/project/ftts/reports/comprail02b s.pdf Accessed: 2009-05-22

Schaefer H. & Pferdmenges, S. (1994). An expert system for real-time train dispatching Available from: http://library.witpress.com/pages/PaperInfo.asp?PaperID=1 1381 Accessed: 2008-11-14

Wikstrom, J.; Kauppi, A.; Andersson A.W. & Sandblad, B. (2006). Designing a graphical user interface for train traffic control, Available from: http://www.it.uu.se/research/publications/reports/2006 025/2006-025-nc.pdf Accessed: 2009-03-12

Luethi, M.; Weidmann, U.; Laube, F. & Medeossi, G. (2006). Rescheduling and Train Control: A New Framework for Railroad Traffic Control in Heavily Used Networks, Available from:http://www.ivt.ethz.ch/ oev/publications/papers/2007A_ML.pdf Accessed: 2009-03-12