In the prediction of numerical simulation for the crowd evacuation behavior under inundated underground, accurate evaluation of the hydrodynamic interaction forces between violent flows and evacuees will be demanded. However, this kind of simulation brings high computational load due to the computation of violent flows in the wide underground area. Increasing the feasibility of the inundated crowd evacuation simulation needs increasing computational efficiency. The fluid force function regarding the hydrodynamic force acting on the evacuee based on the numerical results using DEM-MPS coupling model is proposed in the present study and indicating this function will allow to save computational time sufficiently.

 When a fire breaks out in a road tunnel, it is important to secure a safe evacuation environment, and for that purpose, reliable early fire detection is required. In order to enable the detection of smoke in the early stage of the fire before ignition, smoke sensors, laser gas sensors, and photodiode sensors with different detection methods were selected, and the qualitative correlation between smoke concentration and each sensor output was confirmed in preliminary experiments. However, the detection characteristics of each sensor were investigated by a smoke cylinder and pan fire experiment in a full-scale tunnel. As a result, it was found that within the scope of this experiment, each sensor has the possibility of early detection of fire at the smoldering stage before ignition. In particular, the photodiode sensor suggests the possibility of early fire detection regardless of the smoke color, because the smoke enters the bird's-eye view even before the smoke reaches the sensor installation location.

 The Japanese road tunnel smoke ventilation method, which takes the longitudinal ventilation system as the mainstream, is a great difference with the European point extraction ventilation method. Even in the fire safety assessment, there is a very different concept between the two, that is, Japan focuses on maintaining a safe evacuation environment, while Europe concerns with eliminating fatal accidents. This paper taking these differences into account, assuming longitudinal ventilation (zero-flow-control) and point extraction ventilation system are applied to the same size tunnel individually, then assessed the number of smoke-exposed evacuees by the Japanese method, the other is to assess the injury degree of the evacuees based on the European’s, in order to compare the fire safety of the two smoke ventilation systems. As the results show, in the case of longitudinal ventilation (zero-flow-control) could maintain a safe evacuation environment probably with a small gradient, but it may lead to injury risk if the gradient increases and evacuation delayed. In the case of the point extraction system, except for the gradient of 0%, many evacuees could be surrounded by smoke, however, the evacuees injured incident almost not occurred.

 Humans navigate urban spaces by relying on two different types of spatial memory, namely route maps and survey maps. Conventional experiments that reveal human behavioral characteristics in urban spaces are not only costly in terms of the personnel required for measurement and analysis of large amounts of data but also lack reproducibility of the actual experimental environment. To solve these problems, we propose a system to construct an urban space in virtual reality (VR) using physical movements and exploration. However, it is not clear what type of cognitive maps are formed in the human brain when exploring urban spaces using survey maps. Therefore, the primary purpose of this study is to quantitatively evaluate the cognitive map formed in VR space using geometrical and statistical properties. Experiments were conducted in both spaces, and the geometrical properties of distortion of the cognitive map did not show any specific tendencies for deviation from the real survey map, whereas the statistical properties showed high correlations with the real map. These results indicate that there are no differences in the cognitive maps formed by experiments conducted in two different spaces.

 The surface to line of the expressway tunnel offered open is often dirty by a soot and oil, contamination water and eluviation thing. In addition, the measures that installed a raw water transmission gutter are performed that a leak of water often produces the tunnel constructed in Yaita method of construction and covers the occurrence point to change such as leaks of water. By these factors, as for the image shooting check of the run type, there is concern to influence accuracy of the grasp to change in the check by the proximity viewing from the start. This article carried out the washing of the surface to line for the purpose of further improvement of the grasp to change to occur on the surface to line and examined influence to give the grasp that washing changed and examined the effectiveness of the transparence type raw water transmission gutter for the purpose of the visualization of changing it of the point covered by an installed raw water transmission gutter.

 Sensor-based monitoring is expected to be a method for long-term evaluation of radioactive waste disposal facility. Especially, distributed optical fiber sensor (DFOS) is one of the best candidates, because it has various advantages, such as small and lightweight, no power for sensor, and multi-channel sensors which lead to less impact on radioactive barrier. The authors experimentally investigate to evaluate applicability of DFOS in the facility, while parameters to be monitored after closure of the facility have been selected. Specifically, long-term durability and pressure measurement feasibility is evaluated by laboratory tests. In mock-up test, temperature and strain distribution measurement are applied, and crack on the lower diffusion layer is successfully detected by local strain changes.