Development in building fire detection and evacuation system-a comprehensive review

ReceivedFeb27, 2020 Revised Jun6, 2020 Accepted Jun 18, 2020 Fire is both beneficial to man and his environment as well as destructive and deadly among all the natural disasters. A fire Accident occurs very rarely, but once it crops up its consequences will be devastating. The early detection of fire will help to avoid further consequences and saves the life of people. During the fire accidents, it is also important to guide people within the building to exit safely. Because of this, the paper gives a review of literature related to recent advancements in building fire detection and emergency evacuation system. It is intended to provide details about fire simulation tools with features, suitable hardware, communication methods, and effective user interface.


INTRODUCTION
With the rapid growth in population, buildings are also growing vertically due to the shortage of occupancies. However, with these buildings growing vertically, there arise problems regarding the safe evacuation of people during an emergency like a fire. Similarly, with the growth of insulating materials which catch fire easily and their excessive use in building, a threat to the life of building and human is also increasing. Many people are killed every year, and more are hospitalized due to the fire in India. An average of 520 injuries and 40 deaths occurs per year according to the data provided by NFPA.
A fire Accident occurs very rarely, but once it occurs it's consequences will be devastating. As a result, there is substantial attention given by researchers worldwide for the development of intelligent building systems. Many of these casualties can be avoided if we detect a fire early and guide people to a safe location. To build a suitable fire detection and safe evacuation system it is necessary to focus on parameters such as appropriate sensors, software and hardware tools, and combination techniques and at the end effective user interface.
This article gives systematic reviews of intelligent fire detection and evacuation system as a combination of fire detection (sensors, technology), evacuation assistance (Static/Dynamic Display) and crowd monitoring and prediction (Video, non-video, GPS, Bluetooth). Besides, the review is evaluated based on (i) fire and building simulation software, (ii) suitable hardware (with their merits and demerits), (iii) fire and smoke detecting sensors, (iv) communication methods and (v) effective user interface for a safe exit. In this, efforts have been taken to identify research gaps, and suggestions are made to fulfill the need. Further, a novel architecture is recommended for future fire detection and safe evacuation system.

DEVELOPMENT IN FIRE DETECTION AND EMERGENCY EVACUATION 3.1. Fire sensing, sensors and devices
It involves a Multi-sensor module for sensing parameters like heat, smoke, flame, gas etc.as depicted in Figure 1, with signal conditioning, amplifier circuit, processing unit, alarm system, and safe evacuation path display [33].

Hardware platforms
After detecting a fire event it needs to be processed for further course of actions, for which effective hardware platform plays a key role. Figure 2 shows the block diagram of the fire alarm system. Many hardware devices are used by researchers till now, the details are given below:

Raspberry pi
It is cheaper and minicomputer that can easily plug into a monitor. Connecting the internet to Raspberry pi is very easy as it is having an Ethernet port and USB Wi-Fi dongles. Multiprocessing is possible using this board. Programming languages like Python, C, C++, and Ruby can be used.It is a low-cost computer based on Linux OS, featured with an Ethernet controller and two USB ports. It is having ARM-architecture, with 512 MB data memory, onboard graphics processing unit (Videocore IV). It supports multiple programming languages and software installation. We can install Apache HTTP Webserver to serve the webpage.

Arduino board
Arduino boards are equipped with many digital inputs and output pins that can be interfaced with different expansion boards or shields and other required circuits. These boards featured with serial communications interfaces such as USB, which are also used for loading programs from personal computers. These boards can be programmed using programming languages such as C and C++.

FPGA
The Filed Programming Gate Arrays (FPGA) are semiconductor-baseddevices equipped with interconnection circuits and programmable logic blocks. The FPGA provides features such as the high speed with real-time processing. It can be programmed to meet the requirement of functionality even after its manufacturing.The emergency evacuation system for multi-storey buildings using FPGA is proposed which can operate at the very clocking rate and can handle multiple inputs and multiple outputs simultaneously. This feature of FPGA's is very useful when there is multi-storey building as it needs many sensors and local control boards for every floor.

DSP
DSP is a kind of processor for processing real-time signals. During fire accidents, fire signals need to be processed in real-time with high precision and high speed. DSP not only controls the sampling procedure but deals with the sampling data in advance also. Table 2 provides the comparison of all hardware platforms used which readily displays the response time with the kind of language supported, thus highlighting that FGPA and DSP have faster response time as compared with Raspberry Pi and Arduino board.

Simulation tools 3.3.1. BIM
Building Information Modeling (BIM) provides the facility to makea layout in three-dimension (3D) and carry out disaster prevention activities. Specifically, it is used in the design of fire sprinklers and an alarming system. e.g. BIM 360, AutoCAD, Revit, etc. A 3-dimensional building evacuation and guidance system overcomes the problem of poor visibility in 2 dimensional based systems. The basic use of BIM technology is the exchanging and processing of building information, so the architecture of the fire evacuation system can be designed. BIM-based 3-dimensional model is very close to a real-time scenario and it is proven that this model helps to provide information such as rescue conditions, evacuation paths, etc [35].

GIS
Spatial or geographical information can be achieved through a software tool such as a geographic information system (GIS) [36]. It contains activities such as capture, store, manipulate, analyzes, manage, and present. GIS and CAD integration helps to visualize geospatial information from various locations of the building during a fire emergency.The GIS is used to simulate evacuation environment and to visualize evacuation process. In [37], the author presented a multi-agent-based fire evacuation model, in which GIS is used to capture the geospatial data of agents and trace its location and provide a safe exit path.

FDS
Fire Dynamics Simulator (FDS) is a fire-driven fluid flow model with computational fluid dynamics. It is a sort of a simulation tool developed by NIST's Building Fire Research Laboratory. FDS can analyze three-dimensional fires very quickly [38]. This enables simulation of smoke and heat transfer from fire, besides, FDS provides a fire detector and water sprinkler analysis.PyroSim is a kind of simulation tool can be used as a Fire Dynamic Simulator (FDS). It can also be used to simulate the motion of people.

MATLAB
MATLAB based image processing can be used to detect a fire accident. Any CCTV camera mounted in the building premises will take the input image for the device from video in real-time. Initially, we need to separate the fire and background from the received image using the color segmentation method. Further, fire growth is estimated by comparing two consecutive fire images. If fire growth continues over certain images, the emergency alarm is switched on and the computer screen displays the real-time video with the fire indication in red color.MATLAB robotics package has been used to perform path planning. MATLAB AHP toolbox can be used to calculate available evacuation paths and avoid high-risk ones based on the information of the properties, people density, and evacuation distance. Figure 3 displays color segmentation with a five frame difference in the image processing methodology adopted for fire detection based on MATLAB. Detail comparison of software tools is mentioned in Table 3.  Figure 3. MATLAB based image processing for fire detection [31] Table 3. Comparison of simulation tools used to design smart fire detection and safe evacuation system Source Software Tool Use cases in Fire safety [7] Mobix3D To show a 3D view of the exit path [19] CAD 3D Floor design [29] FDS To determine the flow of fire [34] BIM 3D dimensional building visualization [37] GIS For designing routing service [31] MATLAB Image processing for fire detection

Communication technology
Sensors are the end devices that are meant to collect various environmental parameters for early detection of fire. These sensors are connected in a suitable type of network for smooth communication among them. Following are the communication technology preferred by researchers:

Zigbee
It is a wireless communication technology based on IEEE 802.15.4, used for remote low-power digital radios and automation. It is used for various applications like monitoring hydrant, fire alarming system, predicting safe exit path, etc., in fire detection and safe evacuation system. The physical range of ZigBee wireless communication is about 10 to 20 meters and it varies according to environmental conditions. Figure 4 displays the flow diagram of a wireless sensor network. It uses small and low powered radio devices and it consumes less amount of power to transmit and receive data.

RFID
Radio-Frequency Identification tag (RFID) uses radio waves to read and collect tag information attached to any object. It can detect and read from several feet of distance and it not needs to be within LoS. In many research works RFID tags are used to find the accurate evacuation path and in some cases, it is used to know the density of people in a particular region. The RFID-based module used to calculate the density of individuals in each room and to predict an algorithm-based evacuation path. This RFID reader receives signals from all tags and decides where each tag is located using a unique number.

Bluetooth
Bluetooth is a wireless technology used for the exchange of data between fixed and mobile devices for short-range communication. It's very suitable for IoT-based applications because it's power-efficient and low cost.It is preferred where distance is not a matter like within a room, communication between sensor nodes. A detailed comparison is mentioned in Table 4.

User interface
When fire accidents happen, people stuck inside the building should be evacuated through a path, and information ona safe evacuation route needs to be conveyed properly. Following are the available options for guiding people about safe exit path.

Mobile terminal with app and/or web browser
Withthe rapid evolution in the mobile manufacturing industry, Smartphone is easily available at a decent price. In general, each person carries his/her own Smartphone. We can develop mobile phone-based apps and/web app which can alert individuals during an emergency and also guide them by proving safe exit paths. The mobile-based app can provide a 3D view of the evacuation path using tools such as MobiX3D viewer. According to some researchers, based on the number of smartphones carrying by peoples in a geographical area it is also possible to predict the density of people in that area.

Digital signage and audio visualization
Digital signage is placed on the walls of the building, which provides information related to the environment [39]. During an emergency, these modules receive an emergency signal from the controlling unit. Along with digital signage, audio signals assist people to exit safely with proper guidance.

RESEARCH GAP AND SUGGESTIONS 4.1. Long range communication technology
Connected devices have so far running on traditional protocols like Bluetooth, Wi-Fi, and Cellular Network, etc. But embedded sensors and other IoT devices have different network requirements like low power, long-range, etc. The basic architecture of LoRa can be seen in Figure 5. To meet these requirements one can go for LPWAN, LoRa. LoRa is a Long Range, low power wireless, low data rate communication protocol built to construct an IoT network [40]40 [41]. LoRa uses sub-gigahertz radio frequency bands which are license-free. It provides long-range transmissions (1-15 km) with low power consumption.

Intelligent digital display
During fire accidents, people caught inside building needs to be evacuated through a safe exit path. Traditional fire exit paths may lead to the wrong destination as fire spreads in any direction. Hence, the intelligent digital display needs to be developed which can guide people based on real-time situations. Figure 6 displays the sign conventions for the intelligent and dynamic LED display for right and left turn with no and safe entry [42]2, [43].

Hardware selection & customization
The selection of hardware is a key parameter while designing fire detection and evacuation system. The parameters like accuracy, quick response, cost, size, robustness, and reliability need to be considered while selecting hardware [44]. Instead of considering ready boards for the designing system, it is recommended to develop customized hardware that assists to reduce cost and size. Also, it provides more scalability for future expansion.

Density of people
Secure evacuation of people from the building is of utmost importance during a fire emergency. People density is a key parameter in rescue operation during a fire event and it is partially addressed till now [45]5 [46]. We need to install an effective system that can measure several persons stuck in a fire. Motion sensors, RFID, IP Based camera, etc, are the solution for measuring people's density.

NOVEL ARCHITECTURE FOR FIRE DETECTION AND EVACUATION
Early detection of fire helps to reduce major casualties and saves the life of human beings and property. In the view of this general architecture of smart fire detection and evacuation system is suggested in Figure 7. It is constituted by end devices, Controlling unit, Central Controlling Unit, and End-user Interface. Popular communication technologies such as Bluetooth, wifi, LoRA, etc. are used in the proposed system.

End devices
The key component of our system is end devices, which play the role of collecting valuable data. The collected data like temperature, humidity can be used to decide the fire accident and take further action. These end devices involve sensors such as temperature sensor, smoke sensor, humidity sensors, actuators such as sprinkler system for controlling fire, IP based CCTV for finding the flow of fire and detecting number people inside the building and led display for guiding people to safe exit [47]7 [48].

Controlling unit
These devices are controlling the end devices through Wi-Fi and or Bluetooth. They are capable to collect data from sensors and cameras and further based on programming it will actuate the sprinkler system and LED display accordingly. The controlling unit can be of any microcontroller/Microprocessor-based system such as Arduino, Raspberry Pi, ESP8266, etc. It is enabled with a wireless protocol such as ZigBee [49].

Central controlling unit
It acts as a mediator between the controlling unit and the end-user application. In this case, to communicate with end-user such a fire department we need long-range communication technology. LoRa is a long-range communication technology that contains Long Range communication capability hardware.

End-user interface
End-user such people inside the building should know the status of fire accidents on their mobile phones. It is possible through a mobile-based Android app [32]. It is also necessary to inform the authorized fire department regarding fire accident with live status and location. Mobile or desktop-based applications will act as a user interface for the evacuee and fire department [50].

CONCLUSION
With the rapid growth in technology, the life of human beings is changing enormously as things are becoming smarter. In the era of modern technology and lifestyle, we should think about human safety and security in an emergency condition. Fire accident is one of the dangerous events which need to be addressed effectively with the help smart technology. Because of this, we have surveyed the available methods and systems for fire detection and evacuation systems. Detail analysis is done according to parameters such as sensors, hardware platforms, software tools, and user interface for safe exit during an emergency.
We have identified some areas where still improvement is possible, and it can be done with advanced methods such as (i) LoRa for long-range communication, (ii) customized hardware for more reliability,(iii) Dynamic display guide, and (iv)people density for safe evacuation are suggested. To overcome the issues with the existing system we have proposed a novel architecture for smart fire detection and evacuation. The proposed architecture involves accurate and early sensing of a fire event, IP cameras for tracing people inside building under fire, LoRa for long communication, Zigbee for inter-node communication, actuators to enable sprinkler system and effective user interface.