A multi-sensor cooperative alarm fire monitoring method

Abstract

The application discloses a kind of fire monitoring methods of multi-sensor cooperative alarm, it is related to fire monitoring technical field, it solves the technical problem that some smoke sensors can occur false alarm in the existing fire monitoring method, leading to overall recognition accuracy is low, difficult in later maintenance;The application can significantly improve the recognition accuracy of system for fire, reduce system maintenance cost by considering the cooperative alarm signal of multiple smoke sensors in target area to realize the effective differentiation of sensor false alarm situation, compared with only considering the alarm of single node;The application simplifies fire smoke diffusion process by establishing space topology model, only needs to determine the Euclidean distance of adjacent nodes in space to calculate the shortest path distance between any two sensor nodes, and then solve the cumulative alarm confidence from source node.

Description

Technical Field

[0001] This invention belongs to the field of fire monitoring and relates to a fire monitoring technology with multi-sensor collaborative alarm, specifically a fire monitoring method with multi-sensor collaborative alarm. Background Technology

[0002] Fire is a common disaster that threatens public safety and causes loss of life and property. Especially in high-rise buildings, the smoke produced by a fire can impair a person's mobility when inhaled, leading to entrapment and injury. Therefore, timely monitoring and alarm systems for fires are crucial.

[0003] Current fire monitoring technology primarily utilizes photoelectric smoke sensors. These sensors contain an optical labyrinth with infrared pairs: an infrared emitter and an infrared receiver. In the absence of smoke, the receiver does not receive the infrared light emitted by the emitter. When smoke enters the optical labyrinth, smoke particles scatter some of the light beam onto the receiver, which then receives the infrared light. The denser the smoke, the more light is scattered onto the receiver. When a certain level is reached, the sensor triggers an alarm and transmits a signal to the central control system, indicating a potential fire.

[0004] For large buildings, there are often dozens or even hundreds of smoke sensors installed inside. Due to potential design flaws in some sensors, false alarms may occur at certain sensor nodes throughout the building. In such cases, effectively distinguishing between false alarm signals and genuine fire alarm signals is crucial for accurate and effective fire monitoring. Summary of the Invention

[0005] The present invention aims to solve at least one of the technical problems existing in the prior art; to this end, the present invention proposes a fire monitoring method with multi-sensor collaborative alarm, which is used to solve the technical problem in the existing fire monitoring method that some smoke sensors may give false alarms, resulting in low overall recognition accuracy and difficulty in later maintenance.

[0006] To achieve the above objectives, a first aspect of the present invention provides a fire monitoring method with multi-sensor collaborative alarm, comprising:

[0007] Record all sensor nodes deployed within the target area and construct access nodes related to the spread of fire smoke; the target area is the fire monitoring area.

[0008] A spatial topological model of the fire smoke diffusion process is constructed based on several sensor nodes and channel entrance nodes, and the Euclidean distance between adjacent nodes in the space is measured.

[0009] Given an alarm node, calculate the shortest path length between the alarm node and all other sensor nodes using a path calculation algorithm; the path calculation algorithm includes Dijkstra's algorithm.

[0010] The shortest path length between the alarm node and other sensor nodes is mapped to the alarm confidence increment, and fire monitoring is achieved by combining it with the set alarm confidence threshold.

[0011] Preferably, the identification and marking of the plurality of sensor nodes and the plurality of channel port nodes deployed within the target area includes:

[0012] Record the positions of the smoke sensors in the target area as sensor nodes, and label them as S = [S1, S2, ..., Sn]; where n is an integer greater than 0; and

[0013] Based on the internal structure of the building, the access nodes related to the spread of fire smoke are constructed and labeled as G = [G1, G2, ..., Gm], where m is an integer greater than 0.

[0014] Preferably, the spatial topology model related to the smoke diffusion process is constructed based on the spatial relative positions of the sensor nodes and the channel entrance nodes, including:

[0015] Draw the spatial relative positions of the sensor nodes and the channel entrance nodes; simultaneously determine the alarm nodes, isolated nodes, and ordinary nodes.

[0016] Draw the flow edges between alarm nodes, isolated nodes, and normal nodes to generate the spatial topology model; wherein, the flow edges include unidirectional edges and bidirectional edges.

[0017] Preferably, after generating the spatial topology model, measuring the Euclidean distance between adjacent nodes in the space includes:

[0018] Determine the original node, and designate the node adjacent to the original node as the target node;

[0019] The original node emits a pulsed laser beam towards the target node via a laser rangefinder, and then the pulsed laser beam reflected back from the target node is received by an optoelectronic element;

[0020] The time difference between transmitting and receiving pulsed laser beams is used as the Euclidean distance between the original node and the target node; the Euclidean distance between adjacent nodes is measured and stored as a two-dimensional array.

[0021] Preferably, the shortest path length between the alarm node and each of the other sensor nodes is calculated based on Dijkstra's algorithm, including:

[0022] Create and initialize the array flag[n+m] to mark whether each sensor node has found the shortest path; and dist[n+m] to record the shortest path length from the alarm node to other sensor nodes;

[0023] Traverse each node and find the sensor node i whose shortest path length is not determined, and set flag[i] to true; check all nodes adjacent to sensor node i. If the flag value of sensor node j is false, update the dist information from the alarm node to the sensor node.

[0024] The process is iterated until the shortest path length from the alarm node to each node is determined; the calculated shortest path length from the alarm node to each other sensor node is dist[n] = [d1, d2, ..., dn].

[0025] Preferably, mapping the shortest path length to an alarm confidence increment includes:

[0026] The shortest path length is standardized by using the standard installation spacing Δd of the smoke sensor and the standardization formula xi=di / Δd to obtain the standardized shortest path length X=[x1,x2,…,xn]; where i=1,2,…,n;

[0027] The alarm confidence increment pi is obtained according to the formula pi=exp(-xi) / (exp(-x1)+exp(-x2)+…+exp(-xn)).

[0028] Compared with the prior art, the beneficial effects of the present invention are:

[0029] 1. This invention effectively distinguishes between false alarms by considering the coordinated alarm signals of multiple smoke sensors within the target area. Compared with considering only the alarm of a single node, this invention can significantly improve the system's accuracy in identifying fires and reduce system maintenance costs.

[0030] 2. This invention simplifies the fire smoke diffusion process by establishing a spatial topology model. It only needs to measure the Euclidean distance between adjacent nodes in the space to calculate the shortest path distance between any two sensor nodes, and then solve for the cumulative alarm confidence level starting from the source node. Attached Figure Description

[0031] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0032] Figure 1 This is a schematic diagram illustrating the working steps of the fire monitoring method with multi-sensor collaborative alarm according to the present invention;

[0033] Figure 2 This is a schematic diagram illustrating the deployment of smoke sensors in a simulated real-world scenario.

[0034] Figure 3 A schematic diagram of the spatial topological model of the fire smoke diffusion process constructed in this invention;

[0035] Figure 4 This is a schematic diagram showing the alarm confidence increment result after a single node alarm occurs according to the present invention. Detailed Implementation

[0036] The technical solution of the present invention will be clearly and completely described below with reference to the embodiments. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0037] Please see Figure 1 The first aspect of this invention provides a fire monitoring method with multi-sensor collaborative alarm, comprising:

[0038] S101: Record all sensor nodes deployed within the target area and construct access nodes related to the spread of fire smoke; where the target area is the fire monitoring area.

[0039] S102: Construct a spatial topology model of the fire smoke diffusion process based on several sensor nodes and channel entrance nodes, and measure the Euclidean distance between adjacent nodes in the space;

[0040] S103: Given an alarm node, calculate the shortest path length between the alarm node and all other sensor nodes using a path calculation algorithm; the path calculation algorithm includes Dijkstra's algorithm.

[0041] S104: Map the shortest path length between the alarm node and other sensor nodes to the alarm confidence increment, and combine it with the set alarm confidence threshold to achieve fire monitoring.

[0042] Please see Figure 2 This invention application identifies and marks several sensor nodes and several channel port nodes deployed within the target area, including:

[0043] Record the location of the smoke sensor in the target area as a sensor node and label it as S = [S1, S2, ..., Sn]; and construct the access point node related to the spread of fire smoke according to the internal structure of the building and label it as G = [G1, G2, ..., Gm].

[0044] Since large buildings often deploy dozens or even hundreds of smoke sensor nodes, in order to effectively monitor the spread of fire smoke, it is necessary to record the approximate location of each sensor node in the building within a relatively enclosed area and number them as S = [S1, S2, ..., Sn].

[0045] For any two sensor nodes, the ease or difficulty of fire smoke diffusion needs to be considered. Due to the diversity of building internal structures, there may be some kind of obstacle between two adjacent sensor nodes that prevents smoke from passing directly. Therefore, it is necessary to construct access nodes related to fire smoke diffusion in the relevant area and number them as G = [G1, G2, ..., Gm].

[0046] Figure 2 This is a schematic diagram of the deployment of smoke sensors in a simulated real-world scenario for the present invention application. The gray nodes are smoke sensor nodes deployed in the area, and the black nodes are channel entrance nodes constructed in relation to the spread of fire smoke.

[0047] Please see Figure 3 A spatial topology model related to the smoke diffusion process is constructed based on the spatial relative positions of sensor nodes and channel entrance nodes, including:

[0048] Draw the spatial relative positions of sensor nodes and channel entrance nodes; simultaneously determine alarm nodes, isolated nodes, and ordinary nodes; draw the flow edges between alarm nodes, isolated nodes, and ordinary nodes to generate a spatial topology model; where flow edges include unidirectional edges and bidirectional edges.

[0049] Given all sensor nodes, all access point nodes, and alarm nodes, a spatial topology model of the fire smoke diffusion process is constructed. The specific steps are as follows:

[0050] 1. Draw the spatial relative positions of all sensor nodes and channel port nodes;

[0051] 2. For two directly adjacent nodes in space, there are three possible scenarios:

[0052] 1) One of the two nodes is an alarm sensor node, such as node S1 and node G1. Draw a one-way edge with alarm node S1 as the source and the other node G1 as the destination;

[0053] 2) If one of the two nodes has no other directly adjacent nodes, such as nodes S2 and G2, draw a one-way edge with isolated node S2 as the endpoint and node G2 as the source.

[0054] 3) Other cases, such as nodes G1 and G2. Both nodes can serve as the source and destination, and a bidirectional edge can be drawn between the nodes.

[0055] In this invention application, after generating the spatial topology model, measuring the Euclidean distance between adjacent nodes in the space includes:

[0056] The original node is determined, and the nodes adjacent to the original node are designated as target nodes. A pulsed laser beam is emitted from the original node to the target node using a laser rangefinder, and the pulsed laser beam reflected back from the target node is received by a photoelectric element. The time difference between the emission and reception of the pulsed laser beam is taken as the Euclidean distance between the original node and the target node. The Euclidean distance between adjacent nodes is measured and stored as a two-dimensional array.

[0057] A laser rangefinder is an instrument that accurately measures the distance to a target using a modulated laser with a specific parameter. In practical operation, it emits a pulsed laser beam towards the target, and then a photoelectric element receives the pulsed laser beam reflected back from the target. The Euclidean distance between the rangefinder and the target is then accurately measured by calculating the time difference between emission and reception.

[0058] against Figure 3 In a spatial topology model, the distance between two adjacent nodes is determined and the result is stored in a two-dimensional array, denoted as:

[0059]

[0060] In this table, each row represents the source node, each column represents the destination node, and |G1S1| represents the Euclidean distance from node G1 to node S1. If the value is 0, it means that there is no directly connected propagation path.

[0061] This invention application calculates the shortest path length between the alarm node and other sensor nodes based on Dijkstra's algorithm, including:

[0062] Create and initialize the array flag[n+m] to mark whether each sensor node has found a shortest path; and dist[n+m] to record the shortest path length from the alarm node to other sensor nodes; traverse each node, find sensor node i whose shortest path length has not been determined, and set flag[i] to true; check all nodes adjacent to sensor node i, and if the flag value of sensor node j is false, update the dist information from the alarm node to the sensor node; repeat the traversal process until the shortest path length from the alarm node to each node is determined; the calculated shortest path length from the alarm node to all other sensor nodes is dist[n] = [d1, d2, ..., dn]. Note that there are n smoke sensor locations in the target area, and m access points related to the fire smoke diffusion within the building's internal structure.

[0063] Through the above steps, we can determine the Euclidean distance Dij from a given node i to all its neighboring nodes j. The next step is to calculate the shortest path length from the alarm node to all other sensor nodes, using Dijkstra's algorithm.

[0064] 1. Create and initialize two types of array information:

[0065] 1) flag[n+m]: Used to mark whether each sensor node has found the shortest path. Initialize the flag of the alarm node to true, and set the flag of all other nodes to false;

[0066] 2) dist[n+m]: Used to record the known shortest path lengths from the alarm node to all other nodes. Initialize the shortest path length of the alarm node to 0, the shortest path length of the nodes directly connected to the alarm node to be the Euclidean distance to the alarm node, and the shortest path length of all other nodes to be infinity.

[0067] 2. Traverse each node and find the node i whose shortest path length has not yet been determined (i.e., whose flag value is false and whose dist value is the smallest), and set flag[i] to true;

[0068] 3. Check all nodes adjacent to node i. If the flag value of any node j is false, update the dist information from the alarm node to that node, as follows:

[0069] If the known shortest path length from the alarm node to node i plus the Euclidean distance from node i to node j is less than the known shortest path length from the alarm node to node j, it indicates that a shorter path from the alarm node to node j has been found. The dist of node j is updated to the known shortest path length from the alarm node to node i plus the Euclidean distance from node i to node j. Otherwise, there is no need to update the dist information of node j, that is: dist[j] = min{dist[i] + Dij, dist[j]};

[0070] 4. Repeat steps 2 and 3 until the shortest path length to each node is determined.

[0071] Let the shortest path length from the alarm node to each of the other sensor nodes be calculated as: dist[n] = [d1, d2, ..., dn].

[0072] This invention application maps the shortest path length to an alarm confidence increment, including:

[0073] The shortest path length is standardized by using the standard installation spacing Δd of the smoke sensor and the standardized formula xi=di / Δd to obtain the standardized shortest path length X=[x1,x2,…,xn]; where i=1,2,…,n; and the alarm confidence increment pi is obtained according to the formula pi=exp(-xi) / (exp(-x1)+exp(-x2)+…+exp(-xn)).

[0074] Based on the previous step, the shortest path length dist[n] = [d1, d2, ..., dn] can be obtained from the alarm node to each of the other sensor nodes.

[0075] Based on the standard installation spacing of smoke sensors Δd = 7.5m, the shortest path length from the alarm node to other sensor nodes is standardized as follows: xi = di / Δd; the standardized shortest path length from the alarm node to other sensor nodes is X = [x1, x2, ..., xn].

[0076] The shortest path length to the corresponding node is mapped to the alarm confidence increment after the alarm occurs at that node. The specific calculation formula is: pi=exp(-xi) / (exp(-x1)+exp(-x2)+…+exp(-xn)).

[0077] Considering the case where the simplified topology model has only two nodes, the relationship between the standardized shortest path length and the alarm confidence increment after a single node alarm is as follows: Figure 4 As shown.

[0078] Tests showed that when two nodes are close together, the system alarm confidence increment is small after a single source node sensor alarms. Only when multiple nodes alarm simultaneously can the system alarm confidence exceed the set threshold, after which the system will alarm. This avoids the situation where the overall system recognition accuracy is low due to false alarms from a single sensor.

[0079] When two nodes are far apart, the system alarm confidence increment is large after a single source node sensor alarms, exceeding the set alarm threshold. The system will then immediately issue an alarm, ensuring timely alerts even when only a single isolated sensor exists in the event of a fire in a certain area. The drawback is that if this isolated sensor node issues a false alarm, it becomes difficult to determine whether a fire has actually occurred in the corresponding area.

[0080] Based on the above analysis, in actual production practice, the situation where adjacent smoke sensors are deployed too far apart should be minimized. This ensures that the system can determine whether a fire has actually occurred in the corresponding area through multi-sensor collaborative alarm methods, thereby significantly improving the accuracy of the system alarm and reducing the cost of manual system maintenance in the later stages.

[0081] The data in the above formula are all calculated by removing the dimensions and taking the numerical values. The formula is the closest to the real situation obtained by software simulation of a large amount of collected data. The preset parameters and preset thresholds in the formula are set by those skilled in the art according to the actual situation or obtained through simulation of a large amount of data.

[0082] The above embodiments are only used to illustrate the technical methods of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical methods of the present invention without departing from the spirit and scope of the technical methods of the present invention.

Claims

1. A fire monitoring method of multi-sensor cooperative alarm, characterized by, include: Record all sensor nodes deployed within the target area and construct access nodes related to the spread of fire smoke; the target area is the fire monitoring area. A spatial topological model of the fire smoke diffusion process is constructed based on several sensor nodes and channel entrance nodes, and the Euclidean distance between adjacent nodes in the space is measured. Given an alarm node, calculate the shortest path length between the alarm node and all other sensor nodes using a path calculation algorithm; the path calculation algorithm includes Dijkstra's algorithm. The shortest path length between the alarm node and other sensor nodes is mapped to the alarm confidence increment, and fire monitoring is achieved by combining it with the set alarm confidence threshold. Identifying and marking a plurality of sensor nodes and a plurality of channel port nodes deployed within the target area, including: Record the positions of the smoke sensors in the target area as the sensor nodes, and label them as S=[S1, S2, ..., Sn]; where n is an integer greater than 0; and Based on the internal structure of the building, construct the access nodes related to the spread of fire smoke, and label them as G=[G1, G2, ..., Gm]; where m is an integer greater than 0; Based on the spatial relative positions of the sensor nodes and the channel entrance nodes, a spatial topology model related to the smoke diffusion process is constructed, including: Draw the spatial relative positions of the sensor nodes and the channel entrance nodes; simultaneously determine the alarm nodes, isolated nodes, and ordinary nodes. Draw the flow edges between alarm nodes, isolated nodes, and normal nodes to generate the spatial topology model; wherein, the flow edges include unidirectional edges and bidirectional edges; Mapping the shortest path length to an alarm confidence increment includes: The shortest path length is standardized by using the standard installation spacing Δd of the smoke sensor and the standardization formula xi=di / Δd to obtain the standardized shortest path length X=[x1,x2,…,xn]; where i=1,2,…,n, and di is the shortest path length from the alarm node to the i-th sensor node. The alarm confidence increment pi is obtained according to the formula pi=exp(-xi) / (exp(-x1)+exp(-x2)+…+exp(-xn)).

2. The fire monitoring method with multi-sensor collaborative alarm according to claim 1, characterized in that, After generating the spatial topology model, the Euclidean distance between adjacent nodes in the space is measured, including: Determine the original node, and designate the node adjacent to the original node as the target node; The original node emits a pulsed laser beam towards the target node via a laser rangefinder, and then the pulsed laser beam reflected back from the target node is received by an optoelectronic element; The time difference between transmitting and receiving pulsed laser beams is used as the Euclidean distance between the original node and the target node; the Euclidean distance between adjacent nodes is measured and stored as a two-dimensional array.

3. The fire monitoring method with multi-sensor collaborative alarm according to claim 2, characterized in that, The shortest path length between the alarm node and each of the other sensor nodes is calculated based on Dijkstra's algorithm, including: Create and initialize the array flag[n+m] to mark whether each sensor node has found the shortest path; and dist[n+m] to record the shortest path length from the alarm node to other sensor nodes; Traverse each node and find the sensor node i whose shortest path length is not determined, and set flag[i] to true; check all nodes adjacent to sensor node i. If the flag value of sensor node j is false, update the dist information from the alarm node to the sensor node. The process is iterated until the shortest path length from the alarm node to each node is determined; the calculated shortest path length from the alarm node to each other sensor node is dist[n]=[d1,d2,…,dn].

Images