1 Investigation into the Positioning Accuracy Required for Traffic Management Systems on Different Types of Railway Services H. A. Hamid* † , G. L. Nicholson*, C. Roberts* *BCRRE, University of Birmingham, UK, † Hawija Technical Institute, Northern Technical University, Iraq Keywords: Traffic Management System, Train Positioning System, Position Accuracy. Abstract Nowadays, the digitisation and automation of railway systems are being carried out all over the world, in order to increase the systems’ accuracy, efficiency and reduce maintenance costs. As part of this trend, intelligent traffic management systems (TMS) are under investigation as a way to increase punctuality and automatically return trains back to the original traffic plan when an operational disturbance happens. A TMS needs a variety of input data to consider the current traffic conditions, predict the future state and decide on a new traffic plan when necessary. Numerous studies have proposed TMS designs; all the proposed systems need to read the train positions in real-time for monitoring and analysis purposes. However, the accuracy of the train positions that can be reported in real-time varies; it depends mainly on the control system design and type of positioning sensor used. Train position uncertainty can significantly influence the performance of a proposed TMS, although the impact has rarely been assessed. In this study, TMS positional accuracy requirements for different railway services are investigated. The influence of train positioning uncertainty is studied with respect to TMS of urban, inter-city and high-speed and mixed-traffic services. This is achieved by simulating the characteristics of these railway services in terms of different trains, tracks, a local TMS and train positioning systems with their associated uncertainty. The experiment is carried out first using exact position data; then it is repeated using position data containing stochastic inaccuracies. The TMS outputs are compared with respect to the train order of the traffic plan and the trains’ total delay. The results show that a small positioning deviation can influence the TMS performance of an urban service, while the TMS of high- speed service is affected less by positioning deviation than the TMS of other services. 1 Introduction High demand is currently placed on the railway for both freight and passenger services. Due to the high cost of building new railway infrastructure, railway engineers tend to upgrade the current infrastructure, where possible and required, by using intelligent and automated systems in order to increase railway safety, capacity and punctuality. A typical dispatching system nowadays depends on the experience of a human dispatcher, who applies simple rules to resolve operational disturbances to traffic. Many research studies have proposed the application of an intelligent traffic management system (TMS) in order to increase the effectiveness of the railway dispatching and the junction rescheduling process. Once an operational disturbance is detected, a TMS endeavours to return trains back to the original traffic plan by using a temporary traffic plan, which is based on retiming, rescheduling and/or rerouting the railway trains around and within the conflict area. A TMS needs a variety of input data to monitor the current traffic, predict the future traffic and solve potential traffic conflicts. The train position information in real-time is a vital input to TMS in all proposed systems. Train positioning methods are various and mainly depend on the control system design and type of positioning sensor used. The conventional train positioning method relies on the block occupation technology of a railway signalling system. The track occupation information is supplied to the traffic control centre (TCC) by trackside equipment, usually either track circuits or axle counters [1]. The magnitude of inaccuracies in train position reporting data when using block signalling technology rely on the block section length [2]. The train positioning system in the European Train Control System (ETCS) is based on odometry and balise technologies. In addition to the TMS, train position information is used by other railway systems and applications, such as automatic train protection (ATP), automatic train operation (ATO), driver advisory systems (DAS), passenger information systems and fleet management systems. The required train position accuracy is different from one railway application to another. Many studies propose new positioning technologies in addition to the block occupation-based technology, in order to increase train positioning accuracy. Examples are the Global Navigation Satellite System (GNSS), Inertial Navigation Systems (INS) and Inertial Measurement Units (IMU). There is a rich literature and overview of models and methods used for dealing with precise train positioning [3]. However, there are limited references addressing the required train positioning accuracy for some railway applications. There are studies investigating the impact of some input parameters on the proposed TMS systems, such as the impact of using a flexible timetable on the TMS results in [4] and the impact of local versus global optimisation strategies of conflict resolution in [5]. However, few studies have been conducted on evaluating the impact of the inaccuracies in