A method for optimizing safe fire-fighting paths in complex environments

Abstract

The present invention belongs to the technical field of fire-fighting equipment control, and in particular, relates to a method for optimizing safe fire-fighting paths in complex environments. The method comprises the following steps: determining feasibility indicators; regional zoning and feasible state marking; initializing path zoning information and various parameters; searching for divided areas; and repeating the search process to generate feasible fire-fighting paths for selection. The present invention provides a method for optimizing safe fire-fighting paths in complex environments for various unmanned fire-fighting equipment or systems, which can automatically search and determine the optimal fire-fighting path according to the current disaster situation, and automatically update the optimization according to the changes in the disaster situation.

Description

Technical Field

[0001] The present invention belongs to the technical field of fire-fighting equipment control, and in particular relates to a method for optimizing a safe fire-fighting path in a complex environment. Background Art

[0002] With the rapid development of unmanned equipment, in order to reduce the losses of fire incidents and avoid the harm to personnel caused by firefighting work, various types of unmanned equipment for fire monitoring and firefighting have appeared in firefighting work. Compared with flying unmanned equipment with limited carrying capacity, ground unmanned equipment has strong carrying capacity and high work efficiency. It can undertake firefighting, personnel search, rescue and transportation and other complex tasks. It is an excellent firefighting auxiliary equipment; but most of the current ground unmanned equipment mainly relies on visual information acquisition systems to obtain real-time image data, and performs indirect command control through radio signals and other methods. This method is simple but inefficient, and is only suitable for the control of a single machine or a small number of equipment, and has insufficient perception of the fire scene; it can be foreseen that the joint operation of multiple types of ground unmanned equipment and even ground-air combined multi-unmanned equipment will inevitably appear in future fire monitoring and firefighting scenarios, and the current manual control method may be difficult to meet the needs. Summary of the invention

[0003] The purpose of the present invention is to provide a method for optimizing safe fire-fighting paths in complex environments for various unmanned fire-fighting equipment or systems, based on actual needs, which can automatically search and determine the optimal fire-fighting path according to the current disaster factors, and automatically update the optimization according to changes in the disaster situation.

[0004] To achieve the above object, the present invention adopts the following technical solution.

[0005] The present invention provides a method for optimizing a safe fire-fighting path in a complex environment, comprising the following steps:

[0006] S1. Determine the type of feasible elements of the path, and formulate the feasibility index of the fire-fighting path according to the current attributes of the fire-fighting robot and the attributes of the fire area;

[0007] The path feasibility index includes several feasibility indicators {x 1,t ,x 2,t ....x i,t ...x I,t} and its growth rate {δ' 1,t ,δ' 2,t ....δ' i,t ...δ' I,t};

[0008] where x i,t represents the feasibility index of the ith search period t, δ'i,t ∈[-1,1] represents the feasibility index x i,t The rate of change within the search period t, δ' i,t =0 indicates feasibility index x i,t Keep unchanged, δ' i,t >0 indicates feasibility index x i,t Increase, δ' i,t <0 indicates feasibility index x i,t become smaller;

[0009] The feasibility index value for a certain period t in a region can be expressed as Among them, x i,t =x i,t-1 (1+δ' i,t-1 );k i represents the influence weight of the i-th feasibility indicator, f(x i,t (1+x' i,t )) represents the feasibility index x i,t The measurement function of

[0010] S2. Regional division and feasible status marking

[0011] Divide the fire protection area into zones, mark the known feasible areas and form the feasible domain Q free , mark the infeasible area and form the infeasible region Q no , where the shape of the above blocks is a concise expression of the region division; the regions where the feasible state is unknown are marked and constitute the unknown domain Q un ; For the feasible domain Q free Each region within the system generates a feasibility index value ω for each region mn , mark the infeasible area to form the infeasible region Q no , the marked position region constitutes the unknown domain Q un ;

[0012] S3, initialize the path partition information and the starting area a0 and the target area a aim ; Set the deflection rate ρ0;

[0013] From the starting area a0 to the target area a aim To search, follow these steps:

[0014] Arrival area a m , according to area a m Status update areaa m feasibility indicators and their growth rates;

[0015] Filter areaa m The n regions in the adjacent regions that belong to the feasible domain or the unknown domain constitute the domain set {a n}; Judgment domain set {a n} whether there is a m With target area a aim Area a in the direction of the path n ; If it exists, select the area according to the deflection rate ρ0, and search the area to determine its feasible state;

[0016] If area a n If it is selected and is in a feasible state, it will be added to the feasible domain Q free , and update the feasibility index and growth rate of the region according to the regional status; if region a n If it is selected and is in an infeasible state, it will be added to the infeasible region Q no ;

[0017] S4. If area a n If not selected, then from the domain set {a n} to delete area a n Then we get a new domain set {a j}, calculate the area set {a j Each area a j The cumulative feasibility index value ρ j,t =ω 0,t +ω 1,t +ω 2,t +ω 3,t ... j,t ; Among them, a j Represents the domain set {a j} in the jth region; ω o,t ,ω 1,t ,ω 2,t ,ω 3,t ... j,t Indicates the departure area a0 to area a j The feasibility index of all the areas passed in the current search cycle;

[0018] According to the calculation results, the direction with the largest cumulative feasibility index is selected for single-direction search, or the search resources are allocated in descending order according to the cumulative feasibility index for multi-directional search;

[0019] If a new feasible area is found, it will be added to the feasible domain Q free , update the feasibility index and growth rate of the region according to the region status; otherwise, add all searched regions to the infeasible region Q no ;

[0020] If the domain set {a j All regions in} are divided into the infeasible region Q no , then the area a mAlso divided into the infeasible region Q no , and returns area a m , remove region a from the domain set of that region m Then search again;

[0021] S5. Repeat the above search process until a number of regional chains formed by connecting multiple feasible regions are obtained, and a feasible fire extinguishing path is generated from the regional chain for selection.

[0022] The further improvement or specific implementation of the above-mentioned method for optimizing the safe fire-fighting path in a complex environment also includes a step for optimizing the regional chain path, specifically referring to

[0023] S6. For a certain region chain {c1,c2,c3...c y ,c y+1 ...c Y},c1=a0,c Y =a aim , according to the order of the region chain from front to back, one of the regions c y As the starting point, from its subsequent non-adjacent area {c y+2 ,c y+3 ...} selects regions as end points in order to connect them. If the path between the two regions is in the feasible domain Q free If the end point is changed to the next area for testing, the path between the end points passes through the infeasible region Q. no ;

[0024] The last one that does not pass through the infeasible region Q no The area between the starting point and the end point of the path L3 is deleted from the area chain and the areas passed by the path are added to the area chain in order. The search test is performed backwards again from the area where the end point of L3 is located; the above search steps are repeated until the end point is the target area a aim .

[0025] The further improvement or specific implementation of the above-mentioned method for optimizing the safe fire-fighting path in a complex environment also includes a step for optimizing the further optimization of the regional chain, specifically referring to

[0026] S7. During the process of forming regional connections, the travel costs of each region are recorded;

[0027] For the obtained P regional chains {L1, L2, L3...L p ...L P}, where L p represents the p-th regional chain;

[0028] Analyze whether there are regional chain fragments with the same starting point and end point in any two or more regional chains, cut the aforementioned regional chain fragments from the regional chain for comparison, and calculate the cumulative travel cost of each area in each regional chain fragment;

[0029] In order, the regional chain fragment with the lowest travel cost is selected, and the regional chain fragment is used to replace the corresponding regional chain fragment in each regional chain; one or more regional chain fragments with the second lowest travel cost are selected as redundant candidate storage data.

[0030] The further improvement or specific implementation method of the above-mentioned method for optimizing the safe fire-fighting path under complex environment also includes a step for comprehensively evaluating the regional chain and determining the optimal regional chain to determine the fire-fighting path; specifically,

[0031] S8. Comprehensive evaluation of multi-factor regional chains

[0032] According to various expected needs of regional fire protection, a two-layer evaluation index set is determined, the upper layer is the decision layer, and the lower layer is the impact layer. Each evaluation index in the determination layer index set is affected by the comprehensive influence of several evaluation indicators in the impact layer index set;

[0033] Based on the current fire protection demand, the impact layer evaluation index set {z1,z2...z N} for the evaluation index set {z'1,z'2...z' M The influence of each evaluation index in} mn , where γ mn represents the impact layer evaluation index set {z1,z2...z N The nth evaluation index z in n Relative to the decision layer evaluation index set {z'1,z'2...z' M}The mth evaluation index z' m Relative importance;

[0034] For each path plan, its comprehensive evaluation index can be expressed as

[0035] Among them, R n It refers to the impact layer evaluation index set {z1,z2...z N The nth evaluation index z in n The value of N represents the impact layer evaluation index set {z1,z2...z N}; M represents the evaluation index set {z'1,z'2...z' M} in the evaluation index number; represents k' m The evaluation index set {z'1,z'2...z' MThe evaluation index z' in m weight in this evaluation.

[0036] For the further improvement or specific implementation of the above-mentioned safe fire-fighting path optimization method under complex environment, each evaluation index in the evaluation index set is used to analyze the quality of the determined fire-fighting path. The commonly used decision-layer evaluation indexes include expected fire-fighting cost and expected fire-fighting results. The commonly used influence-layer evaluation indexes that affect the decision-layer evaluation indexes include path length, number of path corners, average passing time, and average path capacity.

[0037] Among them, the path length refers to the actual length of the fire-fighting path, which directly affects the cost and risk probability of fire-fighting materials and personnel entering and exiting the fire scene; the number of path corners refers to the number of path segment angles in the fire-fighting path that are less than the specified angle. When the path segment angle is less than the specified angle (determined by the actual path state), the passing speed of fire-fighting equipment or personnel will be rapidly reduced when passing through, affecting the overall efficiency; the average passing time is the total time measured or estimated for personnel or equipment to pass through the fire-fighting path; the average path capacity refers to the total amount of fire-fighting materials or personnel that can pass through the path at one time during a single pass.

[0038] For further improvement or specific implementation of the above-mentioned safe fire extinguishing path optimization method in complex environment, the importance of the two evaluation indicators is determined by the division method, and the divisions are as follows

[0039] BRIEF DESCRIPTION OF THE DRAWINGS

[0040] Figure 1 It is a flowchart of the method for optimizing safe fire-fighting paths in complex environments;

[0041] Figure 2 It is a schematic diagram of path optimization search;

[0042] Figure 3 It is a schematic diagram of the optimal solution. DETAILED DESCRIPTION

[0043] The present invention is described in detail below in conjunction with specific embodiments.

[0044] The present invention provides a method for optimizing a safe fire-fighting path in a complex environment, which is applicable to various types of ground unmanned equipment, improves its path-finding efficiency in a complex and unknown environment with continuously changing states, and enables it to continuously optimize and update the path based on the implementation environment, thereby achieving dynamic and continuous updating.

[0045] The present invention is mainly used to assist ground fire-fighting drones to achieve autonomous pathfinding from the current position to the target position in a complex and unknown environment. Compared with the joint pathfinding method of drone swarms, considering the characteristics of future unmanned equipment with multiple manufacturers, types and specifications, as well as the rapidly changing disaster situation at the fire scene, this application can enable the unmanned equipment to rely on its own capabilities to determine the fire-fighting path without the assistance of other equipment information, providing a basis for building flexible autonomous firefighting operations with multiple devices.

[0046] like Figure 1 As shown, the method for optimizing safe fire-fighting paths in a complex environment of the present invention mainly includes the following steps:

[0047] S1. Determine the type of feasible elements of the path, and formulate the feasibility index of the fire-fighting path according to the current attributes of the fire-fighting robot and the attributes of the fire area;

[0048] This step is mainly used to determine the optimal path-finding parameters for different types of fire areas or fire scenes. In actual implementation, the fire protection system determines the optimal path-finding parameters based on historical disaster information and the corresponding fire area's own attributes. The path feasibility index refers to several feasibility indicators {x 1,t ,x 2, t ....x i,t ...x I,t} and its growth rate {δ' 1,t ,δ' 2,t ....δ' i,t ...δ' I,t};

[0049] where x i,t represents the feasibility index of the ith search period t, δ' i,t ∈[-1,1] represents the feasibility index x i,t The rate of change within the search period t, δ' i,t =0 indicates feasibility index x i,t Keep unchanged, δ' i,t >0 indicates feasibility index x i,t Increase, δ' i,t <0 indicates feasibility index x i,t become smaller;

[0050] For the feasibility index value ω within a period t in the region t It can be expressed as:

[0051]

[0052] x i,t =x i,t-1 (1+δ' i,t-1 )

[0053] where k i represents the influence weight of the i-th feasibility indicator, f(x i,t (1+x' i,t )) represents the feasibility index x i,t The measurement function of

[0054] Among them, the feasibility index is formulated by the fire protection system according to the characteristics of the fire area and the types of fires that may be involved. It is mainly composed of various factors that determine whether the road is connected and whether personnel or equipment can pass through, including but not limited to the state of insurmountable obstacles in the path, the height and width of the passage path, the bearing capacity of the bridge structure, and the air quality and temperature of the path involving the safety of personnel or equipment. Various visible and measurable or deducible influencing parameters. The above-mentioned influencing parameters determine whether the path can be used as an effective path for fire-fighting materials or personnel to safely enter and exit the fire area when a fire occurs. When implementing it specifically, it should be flexibly formulated according to the actual situation.

[0055] For example, in conventional inland firefighting scenarios, passability is mainly determined by road conditions (bearing capacity, equipment passability, etc.), current area temperature (to avoid damage to equipment or personnel), etc., while for water platform areas, it is necessary to consider wind speed and direction, stability of the path structure (whether it will be damaged under high temperatures, etc.) and other attributes. Therefore, according to the actual situation and the purpose of firefighting, there are various possible combinations of feasibility indicators. The specific indicators and combinations need to be formulated by the firefighting system according to the purposes of various firefighting tasks and the characteristics of the firefighting areas. This part is mainly designed by the firefighting system based on historical experience data, or extracted based on the existing firefighting indicator parameter system, which will not be elaborated in this application.

[0056] S2. Regional division and feasible status marking

[0057] The zone paths are marked according to the current fire status in the fire protection area. The zone marking is based on the following methods:

[0058] like Figure 2 As shown, the fire protection area is divided into zones, the known feasible areas are marked and the feasible domain Q is constructed. free (As shown in the white blocks 01 and 02 in the figure), mark the infeasible areas and form the infeasible region Q no (black blocks 03, 04, 05, etc. in the figure), and the uncertain areas are not marked, such as the gray areas in the figure. It should be made clear that the shapes of the above blocks are concise expressions of regional divisions, not the actual regional topography, etc.;

[0059] Mark the areas where the feasible state is unknown and form the unknown domain Q un ; For the feasible domain Q freeEach region within the system generates a feasibility index value ω for each region mn , mark the infeasible area to form the infeasible region Q no , the marked position region constitutes the unknown domain Q un ;

[0060] S3, initialize the path partition information and the starting area a0 and the target area a aim ; Set the deflection rate ρ0,

[0061] From the starting area a0 to the target area a aim To search, follow these steps:

[0062] Arrival area a m , according to area a m Status update areaa m feasibility indicators and their growth rates;

[0063] Filter areaa m The n regions in the adjacent regions that belong to the feasible domain or the unknown domain constitute the domain set {a n}; Judgment domain set {a n} whether there is a m With target area a aim Area a in the direction of the path n ; If it exists, select the area according to the deflection rate ρ0, and search the area to determine its feasible state;

[0064] If area a n If it is selected and is in a feasible state, it will be added to the feasible domain Q free (As shown in Figure 06, 07, 08), and update the feasibility index and growth rate of the region according to the regional status; if region a n If it is selected and is in an infeasible state, it will be added to the infeasible region Q no (As shown in Figure 09, 10); Figure 2 In the figure, for the area determined on the search path, the feasible area is marked in white, the infeasible area is marked in black, and the area on the search path is gray;

[0065] S4. If area a n If not selected, then from the domain set {a n} to delete area a n Then we get a new domain set {a j}, calculate the area set {a j Each area a j The cumulative feasibility index value ρ j,t =ω 0,t +ω 1,t +ω 2,t+ω 3,t ... j,t ;

[0066] Among them, a j Represents the domain set {a j} in the jth region; ω o,t ,ω 1,t ,ω 2,t ,ω 3,t ... j,t Indicates the departure area a0 to area a j The feasibility index of all the areas passed in the current search cycle;

[0067] According to the calculation results, the direction with the largest cumulative feasibility index is selected for single-direction search, or the search resources are allocated in descending order according to the cumulative feasibility index for multi-directional search;

[0068] If a new feasible area is found, it will be added to the feasible domain Q free , update the feasibility index and growth rate of the region according to the region status; otherwise, add all searched regions to the infeasible region Q no ;

[0069] If the domain set {a j All regions in} are divided into the infeasible region Q no , then the area a m Also divided into the infeasible region Q no , and returns area a m , remove region a from the domain set of that region m Then search again;

[0070] S5, repeat the above search process until a number of regional chains formed by connecting multiple feasible regions (such as abcdefgh in the figure) are obtained, and feasible fire extinguishing paths are generated from the regional chains for selection;

[0071] S6. For a certain region chain {c1,c2,c3...c y ,c y+1 ...c Y}, where c1 = a0, c Y =a aim ;

[0072] like Figure 2 As shown, according to the order of the regional chain from front to back, one of the regions c y As the starting point (as shown in Figure a), from its subsequent non-adjacent area {c y+2 ,c y+3...} (as shown in the figure cd...) select the area in order as the end point (as shown in the figure c) to connect, if the path between the two areas is in the feasible domain Q free If the path between the endpoints passes through the infeasible region Q, the end point is changed to the next area for testing (such as L1 in the figure); until the path between the endpoints passes through the infeasible region Q no (L2 in the figure), the last one that does not pass through the infeasible region Q no The area between the starting point and the end point of the path (L3 in the figure) is deleted from the area chain (as shown in the figure b, c, d), and the areas passed by the path are added to the area chain in order, and the search test is restarted from the area where the end point L3 is located, and the above search steps are repeated until the end point is the target area a aim .

[0073] S7. During the process of forming regional connections, the travel costs of each region are recorded;

[0074] For the obtained P regional chains {L1, L2, L3...L p ...L P}, where L p represents the p-th regional chain;

[0075] Analyze whether there are regional chain fragments with the same starting point and end point in any two or more regional chains, cut the aforementioned regional chain fragments from the regional chain for comparison, and calculate the cumulative travel cost of each area in each regional chain fragment;

[0076] In order, select the regional chain fragment with the lowest traffic cost, and use the regional chain fragment to replace the corresponding regional chain fragment in each regional chain; select one or more regional chain fragments with the second lowest traffic cost as redundant candidate storage data;

[0077] S8. Comprehensive evaluation of multi-factor regional chains

[0078] In order to determine the optimal fire-fighting path, a comprehensive evaluation model is determined based on the current actual fire-fighting situation. Flexible weight adjustment and multi-level indicator evaluation are used to simulate the current fire-fighting demand's expectations or requirements for the fire-fighting path, so as to determine the optimal regional chain and its corresponding fire-fighting path.

[0079] Firstly, a two-layer evaluation index set is determined according to various expected requirements for regional fire protection. The upper layer is the decision layer, and the lower layer is the impact layer. Each evaluation index in the decision layer index set is subject to the comprehensive influence of several evaluation indexes in the impact layer index set. Each evaluation index in the evaluation index set is used to conduct quality analysis on the determined fire extinguishing path. Common evaluation indicators in the decision layer include expected fire protection cost, expected fire protection results, etc. Common evaluation indicators in the impact layer that affect the evaluation indicators in the decision layer include path length, number of path corners, average passing time, average path capacity, etc.

[0080] Path length refers to the actual length of the fire-fighting path, which directly affects the cost and risk probability of fire-fighting materials and personnel entering and exiting the fire scene; the number of path corners refers to the number of path segment angles in the fire-fighting path that are less than the specified angle. When the path segment angle is less than the specified angle (determined by the actual path state), the passing speed of fire-fighting equipment or personnel will be rapidly reduced when passing through, affecting the overall efficiency; the average passing time is the total time measured or estimated for personnel or equipment to pass through the fire-fighting path; the average path capacity refers to the total amount of fire-fighting materials or personnel that can pass through the path at one time during a single pass;

[0081] The above evaluation indicators have different importance according to the actual situation of the fire scene. For example, when it is necessary to quickly reduce the number of fire sources in the early stage of the fire, it is necessary to prioritize reaching the fire scene in the shortest time. When it is necessary to effectively extinguish a large fire after it occurs, it is necessary to prioritize ensuring that the fire extinguishing path has enough capacity to sustain a large number of firefighting materials or personnel.

[0082] To this end, the relative importance of each evaluation indicator γ mn To express;

[0083] like Figure 3 As shown in the figure, based on the current fire protection demand, the impact layer evaluation index set {z1,z2...z N} for the evaluation index set {z'1,z'2...z' M The influence of each evaluation index in} mn , where γ mn represents the impact layer evaluation index set {z1,z2...z N The nth evaluation index z in n Relative to the decision layer evaluation index set {z'1,z'2...z' M}The mth evaluation index z' m relative importance;

[0084] In actual implementation, it is often difficult to determine the relative weight γ by quantifying the value mn The specific size of the evaluation index can be analyzed and compared by the division method, that is, the importance of the evaluation indexes is determined by a number of proportional scales, for example, the commonly used nine-division method is shown in Table 1:

[0085] <![CDATA[Relative weight quantization value γ mn (graduation)]]> meaning 1 <![CDATA[The nth evaluation index z n with respect to the mth evaluation index z' m has a very low impact]]> 3 <![CDATA[The nth evaluation index z n relative to the mth evaluation index z' m has a lower impact]]> 5 <![CDATA[The nth evaluation index z n with respect to the mth evaluation index z' m has a general impact]]> 7 <![CDATA[The nth evaluation index z n with respect to the mth evaluation index z' m has a greater impact]]> 9 <![CDATA[The nth evaluation index z n relative to the mth evaluation index z' m has a great impact <!-- 6 -->]]> 2、4、6、8 Between the above impact levels

[0086] The impact degree is a description of the weight of the impact layer indicator on the decision layer indicator. Since the impact weight is usually difficult to quantify, the nine-point method is used to express and control it, which is conducive to simplifying the analysis process and reducing the difficulty of data processing.

[0087] The degree of impact is a fuzzy evaluation indicator, which is generally given through expert review.

[0088] in

[0089] For each path plan, its comprehensive evaluation index can be expressed as

[0090] Among them, R n It refers to the impact layer evaluation index set {z1,z2...z N The nth evaluation index z in n The value of N represents the impact layer evaluation index set {z1,z2...z N}; M represents the evaluation index set {z'1,z'2...z' M} in the evaluation index number; represents k' m The evaluation index set {z'1,z'2...z' M The evaluation index z' in m weight in this evaluation.

[0091] Relevant indicator parameters are obtained, and one or more optimal solutions are determined after calculation through the influencing layer and the determining layer indicators to obtain the best one or several path optimal solutions.

[0092] Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present invention, rather than to limit the scope of protection of the present invention. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solution of the present invention can be modified or replaced by equivalents without departing from the essence and scope of the technical solution of the present invention.

Claims

1. A method for optimizing safe fire-fighting paths in complex environments, characterized in that: The steps include: S1. Determine the type of feasible elements of the path, and formulate the feasibility index of the fire-fighting path according to the current attributes of the fire-fighting robot and the attributes of the fire area; The path feasibility index includes several feasibility indexes {x 1,t ,x 2,t ....x i,t ...x I,t } and its growth rate {δ1' ,t ,δ2' ,t ....δ i ' ,t ...δ I ' ,t }; The feasibility indicators include the status of insurmountable obstacles in the path, the height and width of the passage path, the bearing capacity of the bridge structure, and the air quality and temperature of the path related to the safety of personnel or equipment; The growth rate refers to the rate of change of the feasibility indicator during the search cycle; where x i,t represents the feasibility index of the ith search period t, δ i ' ,t ∈[-1,1] represents the feasibility index x i,t The rate of change within the search period t, δ i ' ,t =0 indicates feasibility index x i,t Keep unchanged, δ i ' ,t >0 indicates feasibility index x i,t Get bigger, δ i ' ,t <0 indicates feasibility index x i,t become smaller; The feasibility index value for a certain period t in a region can be expressed as Among them, x i,t =x i,t-1 (1+δ i ' ,t-1 );k i represents the influence weight of the i-th feasibility indicator, f(x i,t (1+x i ' ,t )) represents the feasibility index x i,t The measurement function of S2. Regional division and feasible status marking Divide the fire protection area into zones, mark the known feasible areas and form the feasible domain Q free , mark the infeasible area and form the infeasible region Q no , mark the area where the feasible state is unknown and form the unknown domain Q un ; For the feasible domain Q free Each region within the system generates a feasibility index value ω for each region mn , mark the infeasible area to form the infeasible region Q no , marking the unknown region to form the unknown domain Q un ; S3, initialize the path partition information and the starting area a0 and the target area a aim ; Set the deflection rate ρ0; From the starting area a0 to the target area a aim To search, follow these steps: Arrival area a m , according to area a m Status update areaa m feasibility indicators and their growth rates; Filter areaa m The n regions in the adjacent regions that belong to the feasible domain or the unknown domain constitute the domain set {a n }; Judgment domain set {a n } whether there is a m With target area a aim Area a in the direction of the path n ; If it exists, select the area according to the deflection rate ρ0, and search the area to determine its feasible state; If area a n If it is selected and is in a feasible state, it will be added to the feasible domain Q free , and update the feasibility index and growth rate of the region according to the regional status; if region a n If it is selected and is in an infeasible state, it will be added to the infeasible region Q no ; S4. If area a n If not selected, then from the domain set {a n } to delete area a n Then we get a new domain set {a j }, calculate the area set {a j Each area a j The cumulative feasibility index value ρ j,t =ω 0,t +ω 1,t +ω 2,t +ω 3,t ... j,t ; Among them, a j Represents the domain set {a j } in the jth region; ω o,t ,ω 1,t ,ω 2,t ,ω 3,t ... j,t Indicates the departure area a0 to area a j The feasibility index of all the areas passed in the current search cycle; According to the calculation results, the direction with the largest cumulative feasibility index is selected for single-direction search, or the search resources are allocated in descending order according to the cumulative feasibility index for multi-directional search; If a new feasible area is found, it will be added to the feasible domain Q free , update the feasibility index and growth rate of the region according to the region status; otherwise, add all searched regions to the infeasible region Q no ; If the domain set {a j All regions in} are divided into the infeasible region Q no , then the area a m Also divided into the infeasible region Q no , and returns area a m , remove region a from the domain set of that region m Then search again; S5. Repeat the above search process until a number of regional chains formed by connecting multiple feasible regions are obtained, and a feasible fire extinguishing path is generated from the regional chain for selection.

2. The method for optimizing safe fire-fighting paths in complex environments according to claim 1 is characterized in that: It also includes steps for optimizing the regional chain path, specifically: S6. For a certain region chain {c1,c2,c3...c y ,c y+1 ...c Y },c1=a0,c Y =a aim , according to the order of the region chain from front to back, one of the regions c y As the starting point, from its subsequent non-adjacent area {c y+2 ,c y+3 ...} selects regions as end points in order to connect them. If the path between the two regions is in the feasible domain Q free If the end point is changed to the next area for testing, the path between the end points passes through the infeasible region Q. no ; The last one that does not pass through the infeasible region Q no The area between the starting point and the end point of the path L3 is deleted from the area chain and the areas passed by the path are added to the area chain in order, and the search test is performed backwards again from the area where the end point of L3 is located; step S6 is repeated until the end point is the target area a aim .

3. The method for optimizing safe fire-fighting paths in complex environments according to claim 1 is characterized in that: Also includes a step for further optimizing the regional chain, specifically referring to S7. During the process of forming regional connections, the travel costs of each region are recorded; For the obtained P regional chains {L1, L2, L3...L p ...L P }, where L p represents the p-th regional chain; Analyze whether there are regional chain fragments with the same starting point and end point in any two or more regional chains, cut the aforementioned regional chain fragments from the regional chain for comparison, and calculate the cumulative travel cost of each area in each regional chain fragment; In order, the regional chain fragment with the lowest travel cost is selected, and the regional chain fragment is used to replace the corresponding regional chain fragment in each regional chain; one or more regional chain fragments with the second lowest travel cost are selected as redundant candidate storage data.

4. The method for optimizing safe fire-fighting paths in complex environments according to claim 1 is characterized in that: It also includes steps for comprehensively evaluating the regional chains and determining the optimal regional chain to determine the fire extinguishing path; specifically, S8. Comprehensive evaluation of multi-factor regional chains According to various expected needs of regional fire protection, a two-layer evaluation index set is determined, the upper layer is the decision layer, and the lower layer is the impact layer. Each evaluation index in the determination layer index set is affected by the comprehensive influence of several evaluation indicators in the impact layer index set; Based on the current fire protection demand, the impact layer evaluation index set {z1,z2...z N } for the evaluation index set {z1',z'2...z' M The influence of each evaluation index in} mn , where γ mn represents the impact layer evaluation index set {z1,z2...z N The nth evaluation index z in n Relative to the decision layer evaluation index set {z1',z'2...z' M }The mth evaluation index z' m Relative importance; For each path plan, its comprehensive evaluation index can be expressed as Among them, R n It refers to the impact layer evaluation index set {z1,z2...z N The nth evaluation index z in n The value of N represents the impact layer evaluation index set {z1,z2...z N }; M represents the evaluation index set {z1',z'2...z' M } in the evaluation index number; represents k' m The evaluation index set {z1',z'2...z' M The evaluation index z' in m weight in this evaluation.

5. The method for optimizing safe fire-fighting paths in complex environments according to claim 4 is characterized in that: Each evaluation index in the evaluation index set is used to conduct quality analysis on the determined fire-fighting path. The evaluation indexes at the decision layer include expected fire-fighting cost and expected fire-fighting results. The evaluation indexes at the influence layer that affect the evaluation indexes at the decision layer include path length, number of path corners, average passing time and average path capacity. Among them, path length refers to the actual length of the fire-fighting path, which directly affects the cost and risk probability of fire-fighting materials and personnel entering and exiting the fire scene; the number of path corners refers to the number of path segment angles in the fire-fighting path that are less than the specified angle. When the actual path segment angle is less than the specified angle, the passing speed of fire-fighting equipment or personnel will be rapidly reduced when passing through, affecting the overall efficiency; the average passing time is the total time measured or estimated for personnel or equipment to pass through the fire-fighting path; the average path capacity refers to the total amount of fire-fighting materials or personnel that can pass through the path at one time during a single pass.

6. The method for optimizing safe fire-fighting paths in complex environments according to claim 4 is characterized in that: The importance of each pair of evaluation indicators is determined by the graduation method. The graduations are as follows

Images