Intelligent and situation-aware pervasive system to support debris-flow disaster prediction and alerting in Taiwan

https://doi.org/10.1016/j.jnca.2006.06.008Get rights and content

Abstract

Effective information transmission through robust communications is critical to prevent and alert for natural disasters. However, disasters always destroy the wired communication environment. Moreover, effective information needs to reveal the real situations of the disaster, e.g., the accurate position and the real-time image/video of accident events. An accurate disaster prediction model is useful to reduce casualties and prevent disasters from occurring. An effective disaster prediction is based on the accurate disaster decision model, which can be achieved through the situation-aware information communications between the disaster area and the rescue-control center. This study proposes and designs an Intelligent and Situation-Aware Pervasive System (ISPS), which successfully alert people the occurrence of debris-flow disasters. ISPS is a three-tier architecture consisting of mobile appliances, intelligent situation-aware agents (ISA) and a decision support server based on the wireless/mobile Internet communications. Furthermore, the Location-aware Routing Prediction Method (LRPM) was developed to decrease the transmission traffic and latency of pictures pushing the maps of the disaster to mobile clients. Based on the database of the pre-analyzed 181 potential debris flows in Taiwan, accurate debris flow prediction models were built to prevent debris flow using case-based reasoning (CBR) in the decision support server.

Introduction

With the rapid progresses on consumer electronics and wireless/mobile communication technologies, handheld devices providing Internet and ubiquitous connections, such as personal digital assistants (PDAs) and Tablet PCs combining smart cellular phones have become popular for applications such as on pervasive entertainment, healthcare and disaster rescue (Joseph, 2004; Mohan et al., 1999; Varshney, 2003; Wei et al., 2003). The mobile communication and computing capabilities of the pervasive disaster system provide real-time and effective information about accidents and are very useful for disaster rescue (Kung and Ku, 2003; Satyanarayanan, 2001; Yu, 2003). The disaster rescue operations have three fundamental considerations (Lorincz et al., 2004):

  • (i)

    Real-time and effective information transmitted via robust communications is very important and critical for the disaster prevention and alert. However, the wired communication environment is always destroyed when a disaster occurs, particularly a debris-flow disaster.

  • (ii)

    Situation-aware information about the disaster area for the rescue-control center is very useful for making rescue decisions. The situation-aware information includes the location-based and some context-aware information, such as the mobile client's location and the real-time pictures/videos of accident events (Abowd et al., 2002; Banerjee et al., 2002; Lum and Lau, 2002; Patterson et al., 2003). However, owing to the narrow bandwidth in mobile networks and light computing capabilities in handheld devices, pervasive multimedia services are difficult to deliver (Hare, 2002).

  • (iii)

    An effective disaster prediction model is helpful for reducing the level of casualties and even preventing a disaster. The accurate occurrence probabilities of debris-flow hazards are based on effective prediction models, which can be derived from the well-analyzed knowledge of debris flows and intelligent reasoning engines (Tsai et al., 2002; Kawamura et al., 2003).

This study proposed and developed an Intelligent and Situation-Aware Pervasive System (ISPS) to provide real-time and effective information and to prevent and alter debris flows in Taiwan. The proposed ISPS is a three-tier architecture composed of the mobile clients, the intelligent situation-aware agents (ISA), and the Decision Support Server (DSS) based on the pervasive computing environment. Mobile clients use handheld devices, such as a PDA combined with a cellular phone, to transmit and receive multimedia debris-flow information including texts, pictures and audios based on real-time and interaction communications via GSM/GPRS networks. Mobile clients obtain customized information from ISA, which are located at the fronts of mobile networks. Eight intelligent agents are designed to achieve situation awareness. These agents are the user-interface, monitoring, rain, news, alert, pre-load, collaboration and location-aware agents (LAA). (Due to the limited network bandwidth and computing capability of handheld devices, the Location-aware Routing Prediction Method (LRPM) was proposed to push the required and minimum relief maps from the nearest ISA to the mobile clients in accordance with the direction of the mobile client. The Decision Support System provides the reasoning engine to forecast the probability of debris flow based on the proposed prediction models. To increase the probability of debris flow, DSS applies global position system (GPS)/GIS and Remote Sensing (RS) techniques (Liu and Yang, 2004) to identify the occurrence factors of debris flows. This study proposes three prediction models based on the case-based reasoning (CBR) method (Juell and Paulson, 2003) to predict debris flow effectively.

The rest of this study is organized as follows. Section 2 describes related works about debris-flow disaster information systems. Section 3 describes the infrastructure and system parts of the proposed ISPS system. Section 4 describes the LRPM control procedure. Section 5 describes the control mechanism design of the reasoning engine and the CBR mechanism of the DSS. Section 6 describes the implementation and the numeric evaluations in ISPS. Finally, Section 7 concludes this study.

Section snippets

Related works

In recent years, many studies have focused on developing feasible mechanisms and systems to predict the debris flow occurrence and to provide effective alerting mechanisms to reduce the damage caused by debris-flow disasters.

To monitor and alert the occurrence of debris flows, Chang and Lee (2002) proposed machine vision theory, which was based on the recognizing and analyzing captured images. Although the proposed method had the advantages of locale hazard avoidance and reusability, it needed

ISPS architecture and functionality

Fig. 1 shows the network architecture and related system components of an ISPS to provide real-time and emergency information to prevent and alter debris flows in Taiwan. The proposed ISPS is a three-tier architecture composed of Mobile Appliances (MA), intelligent situation-aware agents (ISA2), and a Decision Support Server (DSS) based on wireless/mobile and Internet communications. The following sub-sections describe the functionality of the three components.

Location-aware routing prediction method

In mobile communication environments, users continuously move and change their location continuously. However, the downloading relief latency maps of mobile appliance is very slow via the GPRS network. Moreover, the display delay of mobile appliances is difficult to tolerate due to their light computing capability. This study proposes the LRPM to push the required and minimum relief maps to the mobile appliance by predicting the direction of the mobile appliance, and to transmit and predict

The CBR scheme

CBR is widely used in many fields, e.g., speech recognition, medical diagnosis and legal arguments (Huang, 2000; Lorenzo and Perez, 1996; Pal and Pabitra, 2004). CBR scheme uses explicit and experienced cases to address and reasons new solutions to non-deterministic problems based on heuristic rules and theoretical models (Luger and Stubblefield, 1999). CBR enables the core of reasoning engine to learn from the experienced and historical data. After reaching a search-based solution for a

System implementation and evaluation

This section describes the implementation and performance evaluation of the ISPS system.

Conclusions

The availability of wideband mobile networks has led to a requirement for multimedia presentations embedded in the powerful handheld appliances providing QOS controls and location-aware services. This study proposed and designed an ISPS, which is a three-tier architecture and composed of the mobile appliances, the ISA, and the decision support server. ISPS provides the probability of debris flow occurrence in Taiwan, and real-time multimedia communications are based on mobile environments.

Acknowledgment

The authors would like to thank the National Science Council of the Republic of China, Taiwan for financially supporting this research under Contract No. NSC-93-2625-Z-020-004.

References (28)

  • G.D. Abowd et al.

    Context-aware computing

    Pervas Comput

    (2002)
  • S. Banerjee et al.

    Rover: scalable location-aware computing

    IEEE Comput

    (2002)
  • S.Y. Chang et al.

    Machine vision applied to detection of debris flow

    in: Proceedings of the second researches on mountain disasters and environmental protection across Taiwan Strait

    (2002)
  • S.C. Chen et al.

    Development of evacuation route and shelter system for debris flow disaster

    in: Proceedings of the second researches on mountain disasters and environmental protection across Taiwan Strait

    (2002)
  • G.W. Cheng

    Quasi-periodicity of debris flow bursting

    in: Proceedings of the third researches on mountain disasters and environmental protection across Taiwan Strait

    (2003)
  • H.T. Chou et al.

    Infrasound of slope hazards and infrasonic monitor system

    in: Proceedings of the fourth researches on mountain disasters and environmental protection across Taiwan Strait

    (2004)
  • C.B. Hare

    Redefining user input on handheld devices, 3G mobile communication technologies

    in: Third international conference

    (2002)
  • Z.W. Huang

    Estimate cost of software development using integrated model: cluster, rule-based reasoning, and case-based reasoning

    in: Proceedings of fifth conference on artificial intelligence and applications

    (2000)
  • C.D. Jan et al.

    Debris-flow rainfall-based warning model on Taiwan Application to Typhoon Mindulle

    in: Proceedings of the fourth researches on mountain disasters and environmental protection across Taiwan Strait

    (2004)
  • A.D. Joseph

    Gaming, fine art, and familiar strangers

    IEEE Pervas Comput

    (2004)
  • P. Juell et al.

    Case-based systems

    IEEE Intell Systems

    (2003)
  • M. Kawamura et al.

    Analysis of slope failures due to the 2000 Tokai heavy rainfall using high resolution satellite images

    in: Proceedings of IEEE international conference on geoscience and remote sensing symposium

    (2003)
  • H.Y. Kung et al.

    A real-time mobile multimedia communication system for the prevention and alert of debris-flow disaster, 2003

    IEEE Veh Technol Conf

    (2003)
  • H.Y. Kung et al.

    The design of numerical prediction models for the debris-flow disaster in Taiwan

    J Chiao Da Manage

    (2005)
  • Cited by (23)

    • A novel case adaptation method based on differential evolution algorithm for disaster emergency

      2020, Applied Soft Computing Journal
      Citation Excerpt :

      GA is used to make feature selection [21]. An accurate debris flow prediction model is built to prevent debris flow using CBR in the decision support server, and the emulation results have shown the high prediction probability of the proposed CBR model [22]. A CBR methodology is adopted to develop an intelligent fault detection system for the production of injection-molded drippers [23].

    • Earthquake disaster waste management reviews: Prediction, treatment, recycling, and prevention

      2019, International Journal of Disaster Risk Reduction
      Citation Excerpt :

      The material intensity coefficient varies according to the number and type of dwelling structure (detached dwelling, complex dwelling, etc.), type of building material (wooden, concrete, etc.), the household characteristics (number in household, property owner's age, years of occupation, income, etc.), location of dwelling (urban or rural), and type of disaster waste (wood chips, concrete, stone, rubble, scrap metal, etc.) Identification is essential because treatment and recycling methods vary for each category. Multiple studies attempt to estimate disaster waste generation instantly and accurately using remote sensing technology [31–36]. Remote sensing technology is able to observe the earth's surface in high resolution using satellites and can create a spatial record of the degree of destruction to buildings in the affected area before and after an earthquake.

    • Emergency decision making for natural disasters: An overview

      2018, International Journal of Disaster Risk Reduction
      Citation Excerpt :

      Based on the database of the 181 potential debris flows in Taiwan, accurate debris flow prediction models were built to prevent debris flow using CBR in the decision support server by Kung et al. [82]. Tong et al. [83] built a new framework and program of grassland fires disaster emergency spatial decision support system based on CBR, which can be used in the prevention and rescue work of the grassland fire, moreover, the system can learn new knowledge by itself. Another decision support system is combining machine learning and natural language processing.

    • Operational early warning systems for water-related hazards in Europe

      2012, Environmental Science and Policy
      Citation Excerpt :

      Recent years have seen a considerable growth of operational systems at regional/country level, due to improvements and proliferation of measuring and monitoring networks, computing facilities, telecommunications, as well as increased awareness of natural hazards and future scenarios possibly affected by climatic changes. Some noteworthy examples are the Japanese early-warning system for debris flows and slope failures (Osanai et al., 2010), the Hong Kong Government's Landslide Warning System (see Chen and Lee, 2004), operational since 1984, and more applications such as in USA (Baum and Godt, 2009 and references therein), Brazil (Ortigao and Justi, 2004), New Zealand (Keys and Green, 2008), Taiwan (Kung et al., 2008), Canada (Jakob et al., 2006), among others. Alarm systems for landslides and debris flows, based on NWP input, are not widely operated in Europe.

    • A collaborative GIS framework to support equipment distribution for civil engineering disaster response operations

      2011, Automation in Construction
      Citation Excerpt :

      The objective and constraints of those models include minimum time cost of logistics distribution with multiple resource types and quantities, spatial constraints for facility and target locations, and vehicle capacity for split deliveries [11,60,61]. Provided with these models, decision makers and first responders should be able to better utilize available information and assets for effective disaster response operations [30,35,58]. However, in disaster response operations, these models are hardly deployed.

    View all citing articles on Scopus
    View full text