CN102737468A - System and method for detecting fire hazard based on wireless multi-sensor information fusion - Google Patents

Abstract

The invention proposes a fire detection system and method based on wireless multi-sensor information fusion. The system is divided into sensor terminal nodes and intelligent gateways. The sensor terminal nodes use ZigBee technology to build a data transmission network for fire monitoring sites. The intelligent gateway is based on the ARM11S3C6410 chip to access the Internet and the general packet wireless service GPRS network. Users log in to the embedded server remotely. Monitoring information is available. Once a fire is detected, the alarm information is sent to the user's mobile phone through the GPRS module. Multiple sensors, including smoke, temperature, and carbon monoxide CO gas, are used to sense the state of the fire, and different trust functions are assigned to whether a fire occurs. The DS evidence theory is used to fuse the information of the three sensors to judge the state of the fire. The wireless multi-sensor information fusion fire detection system and method improve the disadvantages of the traditional fire monitoring system wired solution, and can be widely used in household and industrial environments where wiring is inconvenient and costs are reduced.

Description

Fire detection system and method based on wireless multi-sensor information fusion

technical field

The invention designs a fire detection system and method based on wireless multi-sensor information fusion, which is suitable for real-time remote monitoring of family fires.

Background technique

Fire is the most common disaster that endangers people's lives and properties. How to effectively monitor family fires and prevent disasters has become a safety issue that must be considered in family life. Traditional home fire monitoring equipment is only limited to the inside of the home. Once people leave the monitoring area, it is difficult to obtain monitoring information and take timely and effective measures to deal with dangerous situations. For this reason, the Internet network is used to connect the family fire monitoring equipment and the remote monitoring center to realize remote monitoring and expand the scope of monitoring, which is gradually replacing the traditional fire monitoring system. At present, the fire monitoring system based on the Internet often adopts a distributed design. The fire sensing unit and the data acquisition and processing workstation communicate through the CAN bus, and the workstation is connected to the remote monitoring center through the Internet. The digital sensing unit and workstation of this kind of fire monitoring system are connected by wire, which leads to high wiring cost and inflexible structure.

Wireless sensor networks are widely used in disaster monitoring systems such as landslides and forest insect disasters. They have the characteristics of small amount of transmitted data, small transmission delay and low power consumption. The sensor nodes in the wireless sensor network can be randomly distributed in the monitoring area, and form a wireless network system in a self-organizing manner, which can effectively avoid the disadvantages of high wiring costs and damage to the home environment caused by traditional solutions such as Internet-based fire monitoring, and can Realize multi-point continuous measurement of the monitoring area.

The occurrence of fire is a complex process that changes with time, and different fire characteristic signals will be produced in different stages of fire occurrence. Traditional fire detectors are often aimed at a single characteristic signal, such as point-type smoke detectors and point-type heat detectors that respond to smoke particles and heat (temperature) respectively. Single-parameter fire detectors have limited detection capabilities due to uneven response sensitivity to characteristic signals. The detectors can only be selected according to different places and the types of fires that may occur in the place. Once improperly selected, it will cause errors and missed reports. In addition, most of the data processing methods used in single-parameter fire detectors are threshold value comparison methods. This traditional fire detection is simple and easy to implement, but its environmental adaptability and anti-interference ability are poor.

The present invention is based on the fire detection system and method based on wireless multi-sensor information fusion, uses DS evidence theory to fuse multi-sensor data to judge fire, provides users with the ability to remotely monitor the environment, and has high fire monitoring accuracy and convenient and fast monitoring. the

Contents of the invention

The purpose of the present invention is to provide a fire detection system and method based on wireless multi-sensor information fusion, which combines wireless sensor ZigBee network short-distance communication and Internet remote monitoring to realize remote fire monitoring of the user's home environment.

The technical solution of the present invention to solve the above technical problems is that the fire detection hardware system includes two parts: a sensor terminal node and an intelligent gateway. The occurrence of a fire is a comprehensive process accompanied by changes in light, smoke, temperature, radiation and gas concentration, such as the release of carbon monoxide gas, sulfur dioxide and other gases, smoke such as aerosols, open flames and rapid temperature rises. The terminal node is equipped with a carbon monoxide sensor MQ-7, a broad-spectrum gas sensor MQ-2, and a temperature sensor DS18B20 to sense carbon monoxide, sulfur dioxide, smog, and temperature respectively, and the CC2430 processor is used to complete sensor data collection, preprocessing, and transmission.

According to the distance range measured by the sensor, the family room is divided into multiple monitoring areas, and sensor terminal nodes are installed in each area. The ZigBee protocol is used between the sensor terminal and the intelligent gateway to establish a wireless sensor network for wireless communication, which meets the needs of multiple fire sensor terminal nodes for network communication and transmission of information. At the same time, several routing nodes are installed inside the home for data exchange to increase transmission distance and ensure network reliability.

The gateway is responsible for the tasks of organizing the network, processing multi-sensor data and user interaction, and has high computational complexity and intensity. The gateway is composed of ARM11 processor, RF transceiver module CC2430, power supply and reset module, camera, GPRS module (GTM900-C), Ethernet interface (DM9000AEP) module and memory module. The gateway is equipped with an embedded Linux system as the software platform of the gateway, and the Boa embedded Web server is transplanted on the ARM embedded system, with SQLite as the embedded database, and the CGI interface is used to realize the dynamic page between the embedded Web server and the user browser interaction. The gateway receives the remote query request from the browser, displays the current sensor information status in real time, fuses multi-sensor data to make decisions about the fire situation, and controls GPRS to send alarm information. The camera is used for video surveillance so that users can further confirm the fire status through video.

The workflow of fire information fusion and detection based on ZigBee wireless sensor network is as follows: the gateway first starts to establish a new network; when the network is successfully established, it accepts multiple terminal nodes and routing node network access requests and assigns short addresses; the gateway starts monitoring commands , regularly receive multiple sensor data on the terminal node, and perform D-S evidence information fusion and reasoning to decide whether a fire has occurred; during this period, the routing node forwards the information of the terminal node.

The present invention also proposes a fire detection method based on multi-sensor information fusion, which uses the D-S evidence theory to perform multi-sensor data fusion on the carbon monoxide concentration, temperature and smoke collected by the terminal nodes, and characterizes the reliability function of fire occurrence based on each information. Whether the fire occurred.

According to the sensor output response characteristics and expert knowledge, the Gaussian function is selected as the confidence function of smoke, temperature, and carbon monoxide CO gas sensor outputs to represent the occurrence of fire, where the abscissa is the sensor output value, and the ordinate is the fire probability. The trust degree function has three segments, from left to right: no fire, uncertain, and fire. When the smoke concentration is greater than 2.1%/ft, the carbon monoxide CO output is greater than 150ppm or the temperature exceeds , it is determined that there is a fire, and the confidence level increases with the increase of the sensor output value.

 和,其中

，和

分别表示烟雾浓度判断有火，无火和不确定的概率；

，

和

分别表示温度判断有火，无火和不确定的概率；

，

和

分别表示一氧化碳CO气体判断有火，无火和不确定的概率。 After obtaining the output of the smoke, temperature, and carbon monoxide CO gas sensors, the confidence degree of the current fire occurrence can be obtained according to the confidence function of the fire occurrence, which are recorded as ,

and ,in

, and

Respectively represent the smoke concentration to judge the probability of fire, no fire and uncertainty;

,

and

Respectively represent the probability of temperature judging fire, no fire and uncertainty;

,

and

Respectively represent the probability of carbon monoxide CO gas judging fire, no fire and uncertainty.

，其中

表示完全冲突假设

和所有信任度乘积之和，且满足

。 The three sensor data are fused in a sequential manner, that is, the fire occurrence trust function is obtained by fusing smoke and temperature first, and then the trust function is fused with carbon monoxide CO gas. The order of fusion does not affect the final result. According to the DS theory, the calculation formula of the fusion trust function is

,in

represents a complete conflict hypothesis

and The sum of all trust products, and satisfies

.

Thus, the smoke, temperature and carbon monoxide gas have fire, no fire and uncertain probability of the fire, and then judge whether the fire occurs.

The software technical scheme of the present invention mainly includes two processes, that is, the construction process of the ZigBee wireless sensor network and the process of fusion of multi-sensor data by the gateway D-S evidence theory.

The construction process of ZigBee wireless sensor network is as follows:

1) System initialization, including initialization of the coordination node, routing node, and terminal node device type; the LED indicators of the coordination node, routing node, and terminal node are on, indicating that the system initialization work is completed;

2) The coordinating node forms a network, and the terminal node and routing node send out an application to join the network; the coordinating node waits for the node to join the network;

3) The terminal node and the routing node are connected to the network successfully, proceed to the next step 4), otherwise return to step 2) and continue to try to join the network;

4) The coordinating node allocates short addresses for the terminal nodes and routing nodes, and the network construction is completed;

5) The coordinating node sends monitoring commands to the terminal node and waits for the terminal to return data;

6) The terminal node regularly collects sensor data and sends it to the coordination node; the routing node forwards the data of the terminal node;

7) The coordinating node receives the data sent by the terminal node and submits the data to the gateway for processing.

The gateway uses the D-S evidence theory to fuse multi-sensor data as follows:

1) The gateway receives the data collected by the sensor, and assigns the trust function of the fire occurrence according to the sensor output response characteristics and expert knowledge;

2) Calculate the combined evidence of smoke sensor and temperature sensor evidence;

3) Calculate the inconsistency factor of the two evidences in step 2), and then obtain the fusion result of smoke and temperature sensors;

4) Repeat step 2) and step 3) to fuse the first result and carbon monoxide CO gas sensor data to obtain the final fusion result;

5) Continue to receive sensor data for the next fusion.

Description of drawings

Fig. 1 is a topological structure diagram of the present invention, including the logical connection schematic diagram of each module;

Fig. 2 is a block diagram of the terminal node structure of the present invention. On the SOC chip with CC2430 as the core, smoke sensor MQ-2, temperature sensor DS18B20, and carbon monoxide CO gas sensor MQ-7 are connected;

Fig. 3 is the construction flowchart of ZigBee wireless sensor network;

Figure 4 is a flow chart of gateway DS evidence theory fusion of multi-sensor data;

Figure 5 is the confidence function of the smoke sensor output representing the occurrence of a fire;

Fig. 6 is the confidence function that the output of the temperature sensor represents the occurrence of a fire;

Figure 7 is the confidence function of the carbon monoxide CO gas sensor output representing the occurrence of a fire.

Detailed ways

In order to better understand the technical solutions of the present invention, the implementation manners of the present invention will be further described below in conjunction with the accompanying drawings.

Figure 1 shows the topology of the fire detection system and method based on wireless multi-sensor information fusion in the present invention. A variety of sensors are installed on the terminal node with CC2430 as the core, and are arranged in the monitoring area; in order to expand the range of data transmission, several routing nodes are set; the coordinating node of the ZigBee wireless sensor network is installed on the gateway; one end of the gateway receives the coordinating node through the serial port The other end is connected to the GPRS network through the GPRS/GSM communication module GTM900-C; it is connected to the Internet network through the DM9000AEP and the network port.

FIG. 2 is a structural block diagram of terminal nodes of the fire detection system and method based on wireless multi-sensor information fusion of the present invention. On the SOC chip with CC2430 as the core, connect smoke sensor MQ-2, temperature sensor DS18B20, carbon monoxide CO gas sensor MQ-7. The CC2430 chip not only contains ZigBee wireless RF front end, but also integrates an 8051 core as the processor of the terminal node.

As shown in Figure 3, it is the construction flowchart of the ZigBee wireless network based on the fire detection system and method of the wireless multi-sensor information fusion of the present invention, and the flow process is as follows:

1) System initialization, including initializing the coordination node, routing node, and terminal node device type; after initialization, the LED indicators of the coordination node, routing node, and terminal node are on, indicating that the system initialization work is completed;

2) The coordinating node forms a network, and the terminal node and routing node send out an application to join the network; the coordinating node waits for the node to join the network;

3) The terminal node and the routing node are connected to the network successfully, proceed to the next step 4), otherwise return to step 2) and continue to try to join the network;

4) The coordinating node allocates short addresses for the terminal nodes and routing nodes, and the network construction is completed;

5) The coordinating node sends monitoring commands to the terminal node and waits for the terminal to return data;

6) The terminal node regularly collects sensor data and sends it to the coordination node; the routing node forwards the data of the terminal node;

7) The coordinating node receives the data sent by the terminal node and submits the data to the gateway for processing.

As shown in Figure 4, it is a gateway DS evidence theory fusion multi-sensor data flow chart of the fire detection system and method based on wireless multi-sensor information fusion in the present invention, and the process is as follows:

1) The gateway receives the data collected by the sensor, and assigns the trust function of the fire occurrence according to the sensor output response characteristics and expert knowledge;

2) Calculate the combined evidence of smoke sensor and temperature sensor evidence;

3) Calculate the inconsistency factor of the two evidences in step 2), and then obtain the fusion result of smoke and temperature sensors;

4) Repeat step 2) and step 3) to fuse the first result and carbon monoxide CO gas sensor data to obtain the final fusion result;

5) Continue to receive sensor data for the next fusion.

As shown in FIG. 5 , the fire detection system and method based on wireless multi-sensor information fusion of the present invention represents the confidence function of smoke representing the occurrence of fire, where the abscissa is the output value of the smoke sensor, and the ordinate is the fire probability. The trust degree function has three segments, from left to right: no fire, uncertain, and fire. When the smoke concentration is greater than 2.1%/ft, it is determined that there is a fire, and the confidence level increases with the increase of the sensor output value.

As shown in FIG. 6 , the fire detection system and method based on wireless multi-sensor information fusion of the present invention represents the confidence function of fire occurrence by smoke, where the abscissa is the output value of the temperature sensor, and the ordinate is the fire probability. The trust degree function has three segments, from left to right: no fire, uncertain, and fire. When the temperature exceeds , it is determined that there is a fire, and the confidence level increases with the increase of the sensor output value.

As shown in FIG. 7 , the fire detection system and method based on wireless multi-sensor information fusion of the present invention represents the confidence function of carbon monoxide CO to represent the occurrence of fire, where the abscissa is the output value of the carbon monoxide sensor, and the ordinate is the fire probability. The trust degree function has three segments, from left to right: no fire, uncertain, and fire. When the carbon monoxide CO output is greater than 150ppm, it is determined that there is a fire, and the confidence level increases with the increase of the sensor output value.

Claims (8)

1. based on the fire detection system and the method for wireless multi-sensor information fusion; It is characterized in that fire detection system utilizes wireless monitoring terminal node detection condition of a fire environmental parameter, make up intelligent gateway with embedded system; Realize the interconnected of wireless senser ZigBee network, Internet network and general packet radio service GPRS network; The user is long-range to obtain real-time monitor message, in case when fire took place, system sent warning message to user mobile phone through GPRS.

2. fire detection system and method based on wireless multi-sensor information fusion according to claim 1; It is characterized in that; Configuration fire detection terminal node, this sense terminals node is a core with ZigBee SOC(system on a chip) CC2430, the multiple sensors of peripheral configuration detection fire; Comprise smoke transducer MQ-2, temperature DS18B20 and carbon monoxide CO gas sensor MQ-7, and electric power system on the plate.

3. fire detection system and method based on wireless multi-sensor information fusion according to claim 1; It is characterized in that; The configuration intelligent gateway; This gateway is a core with embedded chip S3C6410, and network interface HR911105A, camera, SD card and draw-in groove thereof, USB2.0 interface, ZigBee wireless module, power supply chip AMS1117, power light, the GPRS/GSM network-in-dialing that periphery design comprises Flash, SDRAM, GPRS module GTM900-C, SIM mobile phone card and draw-in groove thereof, GPRS radio-frequency antenna, network card chip DM9000AEP, band transformer be pilot lamp, a hummer, a reset key whether.

4. fire detection system and method based on wireless multi-sensor information fusion according to claim 1; It is characterized in that; After the sense terminals node powered on, the available networking in the search radio-frequency region added after the ZigBee network of gateway configuration; Be in the gateway commands state that receives, afterwards timing acquiring environmental parameter and send to gateway.

5. fire detection system and method based on wireless multi-sensor information fusion according to claim 1 is characterized in that gateway at first starts the foundation of ZigBee network; The transmission terminal node carries out the order of data acquisition; Be in the wait accepting state afterwards, gateway carries out the DS evidence theory to the data that receive merge to be judged, the whether breaking out of fire of making a strategic decision; In case think that dangerous situation takes place, start the GPRS module and send warning message to user mobile phone.

6. fire detection system and method based on wireless multi-sensor information fusion according to claim 1; It is characterized in that; Gateway carries the software platform of embedded Linux system as the gateway part, on the ARM embedded system, transplants the Boa embedded web server, with SQLite as embedded database; Use the CGI interface and realize dynamic page mutual between embedded web server and the user browser; Receive the request of browser remote inquiry, on webpage, show current various sensor information situations in real time, merge multi-sensor data decision-making fire condition.

7. fire detection system and method based on wireless multi-sensor information fusion according to claim 5; It is characterized in that; Select Gaussian function to characterize the degree of belief function that fire takes place as smog, temperature, the output of carbon monoxide CO gas sensor; The degree of belief function has three sections; From left to right be followed successively by no fire, uncertain, fire arranged; When smokescope greater than 2.1%/foot, carbon monoxide CO output is greater than 150ppm or temperature surpasses

; Be judged to be fire, and degree of belief increases with sensor output value.

8. fire detection system and method based on wireless multi-sensor information fusion according to claim 5; It is characterized in that; Adopt the D-S evidence theory to merge smog, temperature, carbon monoxide CO gas sensor data with sequential system; Be that first D-S evidence merges smog and temperature obtains fire generation belief function, and then this belief function and carbon monoxide CO gas are carried out the fusion of D-S evidence, obtain smog, temperature and CO gas the fire that has thus the condition of a fire; Do not have fire and uncertain probability, and then judge whether fire takes place.

Images