This paper gives a number of recent examples of natural disasters that have occurred in Japan, and the major problems they caused. This paper also describes progress made in response to these problems. Three types of method need to be developed: first, a method to detect weaknesses, second, a method for evaluating weather conditions, and finally, a method to evaluate stability in case of meteorological changes. As weather conditions become more severe, the speed of work for developing disaster prevention technology must increase. As a result, it is necessary for RTRI to cooperate as many other research organizations as possible.

Following recent developments in information and communication technology (ICT), various research and development projects are ongoing around the world to propose new train control technology for safer and more stable transportation services. In particular, 5^th generation mobile communication network systems (5G) which will be partially available in 2019 are expected to be used for critical roles in various safety related systems in the future, such as for automatic driving systems for road vehicles and train control systems. This paper outlines recent trends in basic studies for utilizing various communication network techniques including content centric network techniques and 5G systems to sophisticate signalling systems and train control systems. This paper also introduces a future plan for the challenges of the developing autonomous train operation and control system to realize safer, more stable and low-cost train operation services.

Railways are composed of overhead contact lines, vehicles, tracks and structures, and display a complex dynamic behavior as a result of interactions between these components. The Railway Dynamics Division of RTRI covers all areas related to the dynamic aspects of these railway components, and carries out R&D aimed at improving the safety and maintainability of the railways. The main mission of this division, is to develop railway simulators using advanced numerical simulation techniques and high performance computing. This paper introduces the latest R&D on numerical simulation techniques in railway dynamics.

To evaluate the influence of tsunami inundations on railway lines and implement measures for evacuation, it is necessary to predict the depth of tsunami inundations at sites of railway facilities. The author therefore developed a real-time prediction method of tsunami inundations for railway lines, immediately after large earthquakes occur at sea. This method first predicts tsunami water levels in the coastal areas, using tsunami water levels at sea and the tsunami propagation functions. Secondly, the method predicts the depth of tsunami inundations for the target railway line, using a database of tsunami inundations for scenario earthquakes, created in advance.

Since gusts such as tornados, gust fronts, and downbursts are localized and short time phenomena, they cannot always be detected with existing anemometers along railways used for train operation. Therefore, to ensure keep the safety of train operations against these gusts, an algorithm was developed for extracting the strong wind areas on the ground. This algorithm is based consists of two parts. The first is to detect gust indices in upper air layers from echo and Doppler wind velocity observed with existing weather radar, and the second is to estimate the surface wind velocity of ground surface from upper air gust indices using calculation results of previous gust events using a meteorological model.

The natural frequency of a bridge pier is used as an index to evaluate the soundness of pier foundations. However, using conventional approaches, it is sometimes difficult to obtain the natural frequency derived from microtremor, because the peak of the Fourier spectrum of the bridge piers derived from the microtremor does not appear clearly in many cases compared with that derived from the impact vibration test. In this paper, we propose a method that can automatically identify the natural frequencies using only derived from microtremor measurement results at both sides on the top of piers, without renewing the latest result of impact vibration test.

In Japan, there are around 33,000 level crossings and over 90% are equipped with automatic barriers. Despite of all the efforts made by government and railway operators, every year more than 200 people are injured due to accidents inside level crossings. In order to detect pedestrians trapped inside level crossings, we propose a new Obstacle Detector (OD) using infrared camera and image processing algorithm. Evaluation result of the proposed method shows no false negative output and by applying filtering methods, probability of false operation of the OD can be reduced more than 10 times, leading to stable operation.

An on-board train positioning and train integrity system combining inertial sensors and tachometer generators has been developed to detect train position and presence in a section using a radio train control system. In this system, curves and slopes which are registered as distinctive points on the track in an on-board database are used to detect train position and calculate train length. Additionally, train splitting can also be detected by the difference in acceleration between the head vehicle and the tail vehicle. Running tests using fail-safe processors equipped with these functions have demonstrated the feasibility of this system.

In Japan, automatic train operation has been introduced on a restricted number of lines, such as those with no level crossings. More recently there have been plans to expand the introduction of automatic train operation to conventional lines which have level crossings, which increases the importance of having methods for automatically detecting humans and obstacles on railway lines. This paper introduces a test system using 90 GHz band millimeterwave radar and reports the results of tests conducted to confirm its performance specifically for detecting the position of humans on railway tracks. The paper also presents future plans for building a comprehensive railway track monitoring system using 90 GHz band millimeterwave radar.

This study proposes a hybrid simulation (HS) method for pantograph/catenary systems based on a dynamically substructured system (DSS) framework. In this method, the contact force between an actual pantograph and a hydraulic actuator is utilized to calculate the motion of the catenary in real-time, whilst the actuator is driven according to the calculated motion of the catenary. The advantage of the proposed method, when compared with commonlyused methods such as the inverse transfer function approach, is that DSS is better able to avoid instability that can be caused, for example, by pure delay characteristics in the actuator dynamics. The proposed method is also able to accurately represent dynamic interaction between the pantograph and the catenary. In this paper, the DSS methodology is introduced and then the proposed method is validated via simulation and experiment.

Large vibrations due to resonance is a major issue for high-speed railway bridges. In order to efficiently and frequently inspect bridge resonance, this study proposes a detection index, amplification factor, based on the vertical acceleration of the head and tail vehicles of a passing train. Field tests with actual trains and bridges revealed that the amplification factor tends to increase with bridge resonance and the impact factor. In addition, train-bridge dynamic interaction simulation clarified that the correlation of the amplification factor with the impact factor was over 0.9. It also revealed that the amplification factor of a train with short vehicles could be used to detect the potential resonance of bridges.

This paper focuses on the size of the machined contact surface on the wheel flange to clarify its influence on wheel climb derailment after wheel turning. To investigate transient characteristics between machined contact surfaces and smooth contact surfaces, tangential force experiments were carried out using a pair of small cylindrical specimens. Experimental results show that the coefficient of friction on the contact surface increases due to repeated rolling and sliding frictional force. The tangential force coefficient on the machined contact surface is small in comparison the same force measured on the smooth contact surface in the range of the slip ratio of less than about 1.0% due to the difference in the contact patch and surface properties, while the tangential force coefficient with both contact surfaces is almost the same in the range of a slip ratio of more than about 1.0%. This means that the influence of the machined contact surface on the steady and transient characteristics of the tangential force of the wheel/rail is small if only ordinary wheel turning is conducted.