This year’s Student Innovation Challenge addresses the topic of analyzing sensor data in the Sensor Web and generating higher level information products. My research topic draws inspiration from smart city applications. To be more specific, IoT sensor web technologies and 3D city models play significant roles in building the smart city. I argued that most of the existing solutions to integrating 3D city models and sensor observations are usually customized and lack interoperability. Therefore, in order to improve the interoperability of smart city models, I conclude that the integration of the IoT sensor service and the 3D city model have to be based on open standards.

Smart city application – Disaster analysis and management

The rapid economic development and urbanization have led to the intensive use of land in urban areas. In order to accommodate more people and space, the interiors of buildings become more complex. The high capacity and centralized nature affect the design of escape and rescue routes, which could be more complicated or effective. In this situation, the issue of fire disaster prevention and management, including alarm notification and evacuation route simulation, is critical.

Fire disaster management using sensor web technologies and 3D city model

As evacuation maps are commonly displayed in a two-dimensional format, people may find it difficult to plan an ideal evacuation route during an emergency. In recent years, applications of the sensor web have been used to improve the capability of environment observation. The users can monitor the environment immediately and obtain the dynamical phenomenon via the sensor web. For a building model, GeoJSON permits the display of a site in three dimensions. This helps enable occupants to visualize and immediately understand the status of a fire or other emergency. JSON uses nodes and edges to form the network for interior space navigation in three dimensions. Effective fire disaster prevention methods integrate sensing technology, analysis, judgment, decision, and action functions. Therefore, the key aspects of this project includes (1) the fire escape perspective, which provides people with critical and accurate information and identifies the most appropriate routes for evacuation; and (2) the Integration System, which is to integrate and share the information by using a sensor web integration platform.

Project idea

This project applies GeoJSON and an indoor route network with JSON format in a 3D model for the disaster prevention that integrates information on fire prevention facilities, wireless sensors, evacuation route analysis and disaster prevention functions. The creation of a framework for an intelligent fire and disaster prevention system by integration GeoJSON and the route network works as follows. When the wireless sensors detect a high temperature or certain smoke levels the system sets off an alarm. It then monitors the fire outbreak locations and the evacuation routes, which are calculated by the operations in the integration 3D model. The system records the data about this accident in a sensor web database for future review and analysis. The proposed integration system is designed to enhance the timeliness and safety of evacuation actions.

Implementation prototype

The test was conducted in R3 building of the National Central University of Taiwan. Based on the open standard concept, I used GeoJSON for my building model and JSON for my indoor route network data formats. On the other hand, I used the OGC SOS (Sensor Observation Service) for my sensor service. My implementation was based on the assumption that each sensor represents each node of the indoor route network. Therefore, the route node and the sensor observation via the corresponding feature of interest in the sensor service could be integrated.

For the implementation, I simulated a fire in the building. There are some sensors that keep detecting the temperature and the smoke obscuration percentage in the building. Each sensor records the observation of the corresponding space. When a fire starts, the system uses the sensor data, the location of the evacuees and the geospatial 3D city model to analyze and calculate the most suitable escape route for the users. Most importantly, the data is updated dynamically. The evacuees can follow the real-time escape route for firefighters to get their location for rescuing. The system keeps updating the dynamic information and analyzing the sensor data. When a fire event happens, the users can use the browser on the mobile device to follow the route calculated from the analysis module to escape.

The system is written in open source Cesium.js, which displays the three-dimensional building model of GeoJSON via a web browser. In the following figure, the room 203 of the R3 Building was selected as the evacuee’s location. After the system analysis, the right staircase is marked as a dangerous area (the white dot), and the best escape route is via the left staircase and exiting the central exit. The path visualizes the results of the analysis based on the indoor network information (purple route in the figure).

If the evacuee is on the second floor and there is no proper escape route to the exit on the first floor, the system will automatically use the descending ladder on the second floor as the exit of the escape route and analyze the best route to the descending ladder. However, if the route from the evacuee’s location to each exit and the descending step are considered a hazardous area, the system will analyze the location of the evacuee and wait for the rescue in the safe area provided by the system.

Next Steps

The next step in this project could be the visualization of the result of integrating the 3D city model and sensor observations. For example, dynamically colouring the room based on the value of observation. On the other hand, considering more phenomenon, like wind speed to analyze the plume mode, in the warning module can make the system more complete.

Last but not least, I would like to thank 52° North for providing me with this opportunity of participating in this challenge and presenting the results at the Geospatial Sensor Webs Conference.

The code of this project can be found on GitHub and the slide of the presentation held at the Geospatial Sensor Webs Conference 2018 can be found here.