BIM-based engineering site UAV collaborative inspection and dispatching method and system

By correlating construction quality attributes in the BIM three-dimensional model and using taboo search + dynamic planning algorithm to optimize the drone inspection path and speed, the problem of insufficient efficiency and accuracy of drone inspection in engineering construction is solved, and construction is realized Efficient, safe and high-quality inspections on site.

CN114706419BActive Publication Date: 2024-11-01NANJING UNIV
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202210267683.1
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2022-03-17
Publication Date
2024-11-01
Estimated Expiration
2042-03-17

AI Technical Summary

Technical Problem

During the construction process, it is difficult for the existing technology to effectively combine BIM technology and drone inspection to improve the response and accuracy of drone operations and meet the needs of high-quality inspections in complex environments on the construction site.

Method used

By associating the construction quality acceptance attributes into the BIM three-dimensional model, the target point information to be monitored is automatically generated, and a hybrid algorithm for taboo search + dynamic planning is used to build a collaborative inspection and scheduling model for multi-drone aircraft to optimize the inspection path of the drone and speed to ensure efficient completion of inspection tasks.

Benefits of technology

Real-time control of the construction site is achieved, the response and accuracy of drone inspection is improved, the operation efficiency and safety of the construction site is improved, and the construction quality is optimized.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN114706419B_ABST
    Figure CN114706419B_ABST
Patent Text Reader

Abstract

The present invention discloses a method and system for coordinated inspection and scheduling of unmanned aerial vehicles (UAVs) at engineering sites based on BIM, which associates construction quality acceptance attributes with three-dimensional entity elements, and automatically generates location information of target points to be monitored and information on inspection batches, inspection items, and inspection part attributes on the BIM three-dimensional model; based on the obtained target point information to be monitored and monitoring resource information, a multi-UAV coordinated inspection and scheduling model is constructed with the highest monitoring quality as the goal; the multi-UAV coordinated inspection and scheduling model is solved by a taboo search + dynamic programming hybrid algorithm, and finally the optimal inspection path planning and speed optimization scheme of the UAV is obtained. The construction process is controlled in real time using BIM technology, and inspection task information is accurately obtained, and decisions are made on inspection task assignment, path planning, and inspection speed control of multiple UAVs to optimize the overall inspection quality, thereby improving the operation efficiency of the construction site and improving the safety of construction.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to a construction site inspection technology, and in particular to a construction site unmanned aerial vehicle collaborative inspection scheduling method and system based on BIM. Background Art

[0002] The construction industry is a pillar industry of my country's national economy and has a critical impact on my country's economic and social development. At present, the number of newly built buildings in my country is still increasing each year. Some projects face complex geographical environment, great construction difficulty, large project volume and tight construction period. Therefore, it is even more important to ensure the quality and progress of the project during the construction phase. Real-time control of the construction process and construction quality is particularly important.

[0003] At present, most projects are controlled by introducing BIM technology. BIM technology can build a three-dimensional model of a construction project, use digital technology to record the relevant data of the model, and establish an information database. Using data in the IFC standard format in BIM, it is possible to determine the objects that need to be inspected and the inspection points required for the project through certain algorithms for different engineering projects, thereby facilitating the control of project quality.

[0004] In recent years, drone technology has continued to mature and has been widely used. Drones have the advantages of low cost, low safety risk, and ease of use. They have been gradually applied to various inspection scenarios, including quality inspections during construction. Due to the complex construction process of large-scale projects, inspection tasks change dynamically with the progress of the construction. In addition, the construction site has a large space, many quality inspection tasks, and higher quality requirements.

[0005] In view of the above characteristics, the engineering industry urgently needs to combine BIM technology to develop multi-UAV inspection coordination methods, improve the responsiveness and accuracy of UAV operations, and complete inspection tasks at all construction stages with high quality. Summary of the invention

[0006] Purpose of the invention: The purpose of the present invention is to solve the deficiencies in the prior art and to provide a BIM-based engineering site UAV collaborative inspection and scheduling method and system.

[0007] Technical solution: A BIM-based engineering site UAV collaborative inspection and dispatching method of the present invention comprises the following steps:

[0008] S1. Associate the construction quality acceptance attributes with the three-dimensional entity elements, and automatically generate the location information of the target points to be monitored and the information of the inspection batch, inspection items and inspection part attributes on the BIM three-dimensional model;

[0009] S2. Based on the obtained target point information and monitoring resource information to be monitored, a multi-UAV collaborative inspection scheduling model is constructed with the goal of achieving the highest monitoring quality;

[0010] S3. The multi-UAV collaborative inspection scheduling model is solved through the taboo search + dynamic programming hybrid algorithm, and finally the optimal inspection path planning and speed optimization solution of the UAV is obtained.

[0011] Furthermore, the construction content of the multi-UAV path optimization model in step S2 is as follows:

[0012] (1) The objective function is shown in formula (1):

[0013]

[0014] i and j are monitoring point numbers, including 1 drone airport and N inspection targets. The drone airport number is 0, i∈[0,N], j∈[0,N], and i≠j; k is the drone number, with a total of K drones; the inspection target refers to all areas that need to be inspected, and the monitoring point refers to the above areas plus the drone airport;; p i is the shortest monitoring time of inspection target i; t ik is the actual monitoring time of the UAV numbered k on the inspection target i; x ijk is the decision variable. When the UAV numbered k flies from monitoring point i to monitoring point j, x ijk =1, otherwise, x ijk =0;

[0015] (2) The constraints are shown in equations (2) to (20):

[0016]

[0017]

[0018]

[0019]

[0020]

[0021]

[0022]

[0023]

[0024] L 0k =0,k=1,...,K (30)

[0025]

[0026] L ik =A ik +t ik ,i=1,...,N,k=1,...,K (32)

[0027] A 0k ≤T,k=1,...,K (33)

[0028]

[0029]

[0030]

[0031]

[0032]

[0033]

[0034] x ijk ∈{0,1},i=0,...,N,j=0,...,N,k=1,...,K,i≠j (40)

[0035] in:

[0036] Formula (2) indicates that each UAV can reach the UAV airport at most once;

[0037] Formula (3) and Formula (4) indicate that each drone assigned with an inspection task must depart from the drone airport, and drones without inspection tasks will not take off, that is, to ensure that drones do not take off empty. M is a positive number greater than N+1.

[0038] Formula (5) indicates that each inspection target can only be visited by one UAV;

[0039] Formula (6) represents the in-and-out balance constraint of each monitoring point;

[0040] Formula (7) indicates that the speed of each UAV assigned with inspection task from the UAV airport is equal to the speed of returning to the UAV airport;

[0041] Formulas (8) and (9) indicate that the actual monitoring time of each UAV on each inspection target cannot be shorter than its minimum monitoring time;

[0042] Formula (10) indicates that the time when each UAV assigned with inspection task departs from the UAV airport is 0;

[0043] Formula (11) represents the time when each UAV assigned with inspection task arrives at each monitoring point;

[0044] Formula (12) represents the time when each UAV assigned with inspection task leaves each monitoring point;

[0045] Formula (13) indicates that the time for each drone assigned with inspection tasks to return to the drone airport cannot exceed the maximum flight time of the drone;

[0046] Formula (14) and Formula (15) indicate that the time when each UAV assigned with an inspection task arrives at and leaves the scheduled inspection target satisfies the time window constraint;

[0047] Formula (16) represents the power consumed by each UAV assigned with inspection tasks when accelerating or decelerating in each inspection target area;

[0048] Formula (17) represents the power consumed by each UAV assigned with inspection tasks when flying at a uniform speed in each inspection target area;

[0049] Formula (18) indicates that the power consumed by each drone assigned with inspection tasks cannot exceed the maximum power of the drone;

[0050] Formula (19) indicates that the flight speed of each UAV assigned with inspection tasks satisfies the speed constraint;

[0051] Formula (20) represents the value constraints of decision variables;

[0052] Among them, d ij is the distance between monitoring point i and monitoring point j; l i is the monitoring distance length of inspection target i; [a i ,b i ] is the time period during which the inspection target i can be monitored; α is the acceleration of the UAV; -β is the acceleration of the UAV; ρ1 is the acceleration power of the UAV; ρ2 is the deceleration power of the UAV; Q is the maximum power of the UAV; T is the maximum flight time of the UAV; V min V is the slowest flight speed of the drone; max is the fastest flight speed of the drone; v ik A is the uniform flying speed of UAV k at monitoring point i; ik is the time when UAV k arrives at monitoring point i; L ik is the time when drone k leaves monitoring point i; q ik 1 is the power consumption of UAV k when accelerating or decelerating within the inspection target i; q ik 2 is the power consumption of UAV k flying at a constant speed within the inspection target i.

[0053] Furthermore, the specific process of step (3) is as follows:

[0054] S301, generating an initial inspection path planning scheme according to a conservation algorithm;

[0055] S301a, set up a separate drone for each inspection target for inspection;

[0056] S301b, calculating the flight mileage of each UAV;

[0057] S301c, according to the flight mileage, the corresponding saved mileage is calculated according to the saved mileage formula, and a saved mileage table is generated. The saved mileage formula is as follows:

[0058] S ij =d i0 +d 0j -d ij ,i=1,…,N,j=1,…,N,i≠j

[0059] S ij Indicates the number of UAV flight miles saved by removing inspection target j from the current UAV inspection path and inserting it after inspection target i; d ij is the distance between monitoring point i and monitoring point j;

[0060] S301d, if there is a non-negative number in the mileage saving meter, proceed to step S301e, if all are negative numbers, then end and obtain the initial path planning solution;

[0061] S301e, arranging the saved mileage in descending order, starting from the two paths with the largest saved mileage, and calculating the time for the combined UAV to leave each surveillance target according to the lowest flight speed;

[0062] S301f, if the latest departure time constraint formula (15) is satisfied, the two flight routes are merged and the process returns to step S301b; if not satisfied, the current solution is deleted from the mileage saving table and the process returns to step S301d;

[0063] S302, optimizing the initial inspection path planning scheme using a taboo search + dynamic programming hybrid algorithm to obtain an optimal drone inspection scheme;

[0064] S302a, dynamically planning the initial inspection path planning scheme to obtain the optimal inspection speed;

[0065] S302b, calculate the cost of the obtained inspection plan, and the cost function is:

[0066]

[0067] p i represents the shortest monitoring time of each inspection target i, ti Indicates the actual monitoring time of each inspection target i;

[0068] S302c, setting parameters m and n, where m is the maximum number of cycles of the taboo search algorithm, and n is the number of consecutive unupdated optimal solutions that need to be accumulated when implementing the disturbance, starting from the first cycle;

[0069] S302d, determine whether the current cycle exceeds the maximum number of cycles m, if so, end the cycle and output the optimal inspection plan and the maximum cost; otherwise, proceed to the next step;

[0070] S302e, using the relocate operator, the exchange operator, and the cross operator in sequence to construct a neighborhood solution space for the current solution;

[0071] S302f, extracting neighborhood solutions that meet the prohibition-breaking principle and are not tabooed in the neighborhood solution space, and selecting the inspection solution with the highest cost in the neighborhood solution as the current solution;

[0072] S302g, if the current solution is higher than the optimal solution, the optimal solution is updated as the current solution, the number of consecutive times the optimal solution has not been updated is updated to 0, the number of loops is increased by 1, and the process returns to S302d, otherwise, the process proceeds to the next step;

[0073] S302h, determine whether the disturbance condition is met, that is, whether the number of consecutive unupdated optimal solutions is not less than n. If so, randomly select a feasible neighborhood solution as the current solution, update the number of consecutive unupdated optimal solutions to 0, increase the number of loops by 1, and return to S302d. Otherwise, increase the number of consecutive unupdated optimal solutions by 1, increase the number of loops by 1, and return to S302d.

[0074] Furthermore, the specific process of obtaining the optimal inspection speed solution by using the dynamic programming algorithm in step S302a is as follows:

[0075] S302a-1. Simplify the flight speed of the drone into integers, ranging from V min To V max , simplify the total power consumption of the drone to a certain inspection target into an integer ranging from 0 to Q, and simplify the total flight time of the drone to a certain inspection target into an integer ranging from 0 to T;

[0076] S302a-2, assuming that the speed at which the drone starts and ends inspection is V min ;

[0077] S302a-3, starting from the last inspection target and ending at the first inspection target, traverse each inspection target of each drone in the initial inspection path planning scheme;

[0078] S302a-4. For the last inspection target, all possible inspection speeds, power consumption and flight time of the previous inspection target are enumerated, and the best inspection benefit of the last inspection target is calculated according to the state transition formula. The state transition formula is as follows:

[0079]

[0080] (v0,Q K-1 ,T K-1 ,v K-1 ) represents the state, that is, the starting inspection speed is v0, and the total power consumption at the end of the K-1th inspection target is Q K-1 , the total flight time at the end of the K-1th inspection target is T K-1 , the inspection speed of the K-1th inspection target is v K-1 ; V K (v0,Q K-1 ,T K-1 ,v K-1 ) represents the total monitoring benefit from the Kth inspection target to the end of the inspection in this state; p K is the shortest monitoring time of the Kth inspection target; t K is the actual monitoring time of the Kth inspection target; p K (t K -p K ) represents the monitoring benefit of the Kth inspection target;

[0081] S302a-5. For other inspection targets, all possible inspection speeds, power consumption and flight time of the previous inspection target are enumerated, and the optimal inspection benefit of each inspection target is calculated according to the state transition formula. The state transition formula is as follows:

[0082]

[0083] S302a-6, increase the speed of the drone at the start and end of the inspection by 1, and return to S302a-3 until the drone speed exceeds V max Finally, the inspection benefits at each speed are compared, and the best inspection benefit and inspection plan are output. The total inspection benefit calculation formula is as follows:

[0084]

[0085] Furthermore, when constructing the neighborhood solution space in step S302d, the Relocate operator is first used to delete an inspection target from the current path and insert it into any position of any inspection path other than the current path; then the Exchange operator is used to exchange the positions of any two inspection targets in an inspection path; finally, the Cross operator is used to select any two inspection paths, select any inspection target in the path respectively, and exchange the inspection paths after the two targets.

[0086] Furthermore, the relocate operator executes the following steps:

[0087] (a) Starting from the first inspection target, delete it from the currently scheduled drone team and insert it into each inspection position of each other drone inspection team to generate a new path planning solution;

[0088] (b) Each time a new path planning scheme is generated, the maximum flight speed of the drone is used as the inspection speed to determine whether the inspection start time of each inspection target is earlier than the specified inspection start time. If it is earlier, the scheme is deleted; the minimum flight speed of the drone is used as the inspection speed to determine whether the inspection end time of each inspection target is later than the specified inspection end time. If it is later, the scheme is deleted;

[0089] (c) Using dynamic programming algorithm to find the optimal inspection speed for all generated solutions and calculate the cost of the solution;

[0090] (d) Compare the cost of each solution and select the relocate solution with the highest cost.

[0091] Furthermore, the exchange operator executes the following steps:

[0092] (a) For each UAV in the inspection plan generated by the relocate operator, select two inspection targets in its inspection path and exchange the inspection order to generate a new path planning plan;

[0093] (b) Each time a new path planning scheme is generated, the maximum flight speed of the drone is used as the inspection speed to determine whether the inspection start time of each inspection target is earlier than the specified inspection start time. If it is earlier, the scheme is deleted; the minimum flight speed of the drone is used as the inspection speed to determine whether the inspection end time of each inspection target is later than the specified inspection end time. If it is later, the scheme is deleted;

[0094] (c) Using dynamic programming algorithm to find the optimal inspection speed for all generated solutions and calculate the cost of the solution;

[0095] (d) Compare the cost of each solution and select the exchange solution with the highest cost.

[0096] Furthermore, the cross operator performs the following steps:

[0097] (a) For the inspection plan generated by the exchange operator, select any two drones, choose two inspection targets in the two inspection routes as exchange points, and exchange the inspection routes after the exchange points to generate a new path planning plan;

[0098] (b) Each time a new path planning scheme is generated, the maximum flight speed of the drone is used as the inspection speed to determine whether the inspection start time of each inspection target is earlier than the specified inspection start time. If it is earlier, the scheme is deleted; the minimum flight speed of the drone is used as the inspection speed to determine whether the inspection end time of each inspection target is later than the specified inspection end time. If it is later, the scheme is deleted;

[0099] (c) Using dynamic programming algorithm to find the optimal inspection speed for all generated solutions and calculate the cost of the solution;

[0100] (d) Compare the exchange costs of each pair of drones choosing each exchange point and select the optimal solution for each pair of drones to exchange.

[0101] The present invention also discloses a system for a coordinated inspection and dispatching method of unmanned aerial vehicles (UAVs) at engineering sites based on BIM, comprising an environmental monitoring device, a BIM three-dimensional model and a plurality of unmanned aerial vehicles, wherein the unmanned aerial vehicles are equipped with a data collector, a data sensor, a three-dimensional engine camera, a GPS positioning system and a camera;

[0102] BIM Revit builds a 3D model of the project site, uses Timeliner to generate a standard model with a construction schedule, and uses Navisworks for animation simulation. Finally, BIM Serve.org is used to generate quality checkpoint information, which is then used to optimize the inspection route.

[0103] The data in the BIM 3D model is converted into coordinate data, and the corresponding relationship between the BIM 3D model and the drone is established through the 3D engine camera, and the inspection information is input into the drone;

[0104] The drone receives the BIM data information and the relevant information of the environmental monitoring equipment on site through the 5G device network, and then determines whether to start the inspection. If the conditions are met, the inspection will start;

[0105] During inspections, on-site data is recorded through data collectors and data sensors, and the inspection data is transmitted back to the BIM platform in real time through 5G network equipment. Finally, data comparison is performed to determine whether manual on-site review is required.

[0106] Beneficial effects: The present invention uses BIM technology to control the construction process in real time, accurately obtains inspection task information, and further considers various realistic constraints such as inspection requirements, endurance, and spatial distance, and makes overall decisions on the inspection task assignment, path planning, and inspection speed control of multiple drones to optimize the overall inspection quality, thereby improving the operation efficiency of the construction site and improving the construction safety. Compared with the prior art, the present invention has the following advantages:

[0107] 1) Based on BIM, the present invention combines the time dimension, the three-dimensional space of the engineering construction process and the monitoring demand information, and takes advantage of drones to conduct comprehensive real-time monitoring and feedback information for managers to make timely judgments. It can achieve the coordination of engineering physical entity construction and computer virtual entity construction, and guide the construction process to be carried out effectively, safely and with high quality.

[0108] 2) The present invention comprehensively considers multiple factors such as inspection needs, spatial location, monitoring quality, etc., and proposes a two-layer algorithm for multi-UAV collaborative inspection paths, which efficiently generates UAV inspection task allocation plans, inspection path planning plans, and inspection time and speed control methods to optimize the quality of on-site inspections. BRIEF DESCRIPTION OF THE DRAWINGS

[0109] Figure 1 It is a schematic diagram of the overall scheduling process of the present invention;

[0110] Figure 2 It is a schematic diagram of the process flow of the taboo search-dynamic programming algorithm in the present invention;

[0111] Figure 3 (a) Schematic diagram of generating a feasible neighborhood space relocate operator;

[0112] Figure 3 (b) Schematic diagram of generating feasible neighborhood space exchange operator;

[0113] Figure 3 (c) is a schematic diagram of generating a feasible neighborhood space cross operator;

[0114] Figure 4 FIG. 4 is a system structure diagram of an embodiment of the present invention. DETAILED DESCRIPTION

[0115] The technical solution of the present invention is described in detail below, but the protection scope of the present invention is not limited to the embodiments.

[0116] like Figure 4 As shown, the system of the BIM-based engineering site UAV collaborative inspection and scheduling method of the present invention includes environmental monitoring equipment, BIM three-dimensional model and multiple UAVs, wherein the UAVs are equipped with data collectors, data sensors, three-dimensional engine cameras, GPS positioning and cameras; BIM Revit establishes a three-dimensional model of the engineering site, uses Timeliner to generate a standard model with a construction schedule, and performs animation simulation through Navisworks, and finally uses BIMServe.org to generate quality inspection point related information, and uses the quality inspection point related information to optimize the inspection path; the data in the BIM three-dimensional model is converted into coordinate data, and the corresponding relationship between the BIM three-dimensional model and the UAV is established through the three-dimensional engine camera, and the inspection information is input into the UAV; the UAV receives the BIM data information and the relevant information of the environmental monitoring equipment on site received through the 5G device network, and then determines whether to start the inspection, and starts the inspection if the conditions are met; during the inspection, the on-site data is recorded through the data collector and the data sensor, and the inspection data is transmitted back to the BIM platform in real time through the 5G network equipment, and finally the data is compared to determine whether manual on-site review is required.

[0117] The BIM-based engineering site UAV collaborative inspection and scheduling method of this embodiment includes the following steps:

[0118] Step S1, associating the construction quality acceptance attributes with the three-dimensional entity elements, and automatically generating the location information of the target points to be monitored and the information of the inspection batch, inspection items and inspection part attributes on the BIM three-dimensional model;

[0119] Step S2: Based on the obtained target point information and monitoring resource information to be monitored, a multi-UAV collaborative inspection scheduling model is constructed with the goal of achieving the highest monitoring quality;

[0120] (1) The objective function is shown in formula (1):

[0121]

[0122] i and j are monitoring point numbers, including 1 drone airport and N inspection targets. The drone airport number is 0, i∈[0,N], j∈[0,N], and i≠j; k is the drone number, with a total of K drones; p i is the shortest monitoring time of inspection target i; t ik is the actual monitoring time of the UAV numbered k on the inspection target i; x ijk is the decision variable. When the UAV numbered k flies from monitoring point i to monitoring point j, x ijk =1, otherwise, x ijk =0;

[0123] (2) The constraints are shown in equations (2) to (20):

[0124]

[0125]

[0126]

[0127]

[0128]

[0129]

[0130]

[0131]

[0132] L 0k =0,k=1,...,K (50)

[0133]

[0134] L ik =A ik +t ik ,i=1,...,N,k=1,...,K (52)

[0135] A 0k ≤T,k=1,...,K (53)

[0136]

[0137]

[0138]

[0139]

[0140]

[0141]

[0142] x ijk ∈{0,1},i=0,...,N,j=0,...,N,k=1,...,K,i≠j (60)

[0143] in:

[0144] Formula (2) indicates that each UAV can reach the UAV airport at most once;

[0145] Formula (3) and Formula (4) indicate that each drone assigned with an inspection task must depart from the drone airport to ensure that the drone does not take off empty. M is a positive number greater than N+1.

[0146] Formula (5) indicates that each inspection target can only be visited by one UAV;

[0147] Formula (6) represents the in-and-out balance constraint of each monitoring point;

[0148] Formula (7) indicates that the speed of each UAV assigned with inspection task from the UAV airport is equal to the speed of returning to the UAV airport;

[0149] Formulas (8) and (9) indicate that the actual monitoring time of each UAV on each inspection target cannot be shorter than its minimum monitoring time;

[0150] Formula (10) indicates that the time when each UAV assigned with inspection task departs from the UAV airport is 0;

[0151] Formula (11) represents the time when each UAV assigned with inspection task arrives at each monitoring point;

[0152] Formula (12) represents the time when each UAV assigned with inspection task leaves each monitoring point;

[0153] Formula (13) indicates that the time for each drone assigned with inspection tasks to return to the drone airport cannot exceed the maximum flight time of the drone;

[0154] Formula (14) and Formula (15) indicate that the time when each UAV assigned with an inspection task arrives at and leaves the scheduled inspection target satisfies the time window constraint;

[0155] Formula (16) represents the power consumed by each UAV assigned with inspection tasks when accelerating or decelerating in each inspection target area;

[0156] Formula (17) represents the power consumed by each UAV assigned with inspection tasks when flying at a uniform speed in each inspection target area;

[0157] Formula (18) indicates that the power consumed by each drone assigned with inspection tasks cannot exceed the maximum power of the drone;

[0158] Formula (19) indicates that the flight speed of each UAV assigned with inspection tasks satisfies the speed constraint;

[0159] Formula (20) represents the value constraints of decision variables;

[0160] Among them, d ijis the distance between monitoring point i and monitoring point j; l i is the monitoring distance length of inspection target i; [a i ,b i ] is the time period during which the inspection target i can be monitored; α is the acceleration of the UAV; -β is the acceleration of the UAV; ρ1 is the acceleration power of the UAV; ρ2 is the deceleration power of the UAV; Q is the maximum power of the UAV; T is the maximum flight time of the UAV; V min V is the slowest flight speed of the drone; max is the fastest flight speed of the drone; v ik A is the uniform flight speed of UAV k at monitoring point i; ik is the time when UAV k arrives at monitoring point i; L ik is the time when drone k leaves monitoring point i; q ik 1 is the power consumption of UAV k when accelerating or decelerating within the inspection target i; q ik 2 is the power consumption of UAV k flying at a constant speed within the inspection target i;

[0161] Step S3: solve the multi-UAV collaborative inspection scheduling model through the taboo search + dynamic programming hybrid algorithm, and finally obtain the optimal inspection path planning and speed optimization solution of the UAV.

[0162] The specific process is as follows Figure 2 As shown: S301, generating an initial inspection path planning scheme according to a saving algorithm;

[0163] S301a, set up a separate drone for each inspection target for inspection;

[0164] S301b, calculating the flight mileage of each UAV;

[0165] S301c. According to the flight mileage, the corresponding saved mileage is calculated according to the saved mileage formula. The saved mileage formula is as follows:

[0166] S ij =d i0 +d 0j -d ij ,i=1,…,N,j=1,…,N,i≠j

[0167] S ij It means that the monitoring target j is deleted from the current drone inspection path and inserted after the monitoring target i for inspection, and the flight mileage of the drone is saved; d ij is the distance between surveillance target i and surveillance target j;

[0168] S301d, if there is a non-negative number in the mileage saving meter, proceed to step S301e, if all are negative numbers, then end and obtain the initial path planning solution;

[0169] S301e, arranging the saved mileage in descending order, starting from the two paths with the largest saved mileage, and calculating the time for the combined UAV to leave each surveillance target according to the lowest flight speed;

[0170] S301f, if the latest departure time constraint is satisfied, that is, formula (15) in the model, the two flight routes are merged and the process returns to step S301b; if not, the current solution is deleted from the mileage saving table and the process returns to step S301d;

[0171] S302, optimizing the initial inspection path planning scheme using a taboo search + dynamic programming hybrid algorithm to obtain an optimal drone inspection scheme;

[0172] S302a, dynamically planning the initial inspection path planning scheme to obtain the optimal inspection speed;

[0173] S302b, calculate the cost of the obtained inspection plan, and the cost function is:

[0174]

[0175] p i represents the shortest monitoring time of each inspection target i, t i Indicates the actual monitoring time of each inspection target i;

[0176] S302c, set parameters m and n, where m is the maximum number of cycles of the taboo search algorithm and n is the number of consecutive optimal solutions that need to be accumulated when implementing the disturbance. Start from the first cycle.

[0177] S302d, determine whether the current cycle exceeds the maximum number of cycles m, if so, the algorithm ends and outputs the optimal inspection plan and the maximum cost; otherwise, proceed to the next step;

[0178] S302e, using the relocate operator, the exchange operator, and the cross operator in sequence to construct a neighborhood solution space for the current solution;

[0179] S302f, extracting neighborhood solutions that meet the prohibition-breaking principle and are not tabooed in the neighborhood solution space, and selecting the inspection solution with the highest cost in the neighborhood solution as the current solution;

[0180] S302g, if the current solution is higher than the optimal solution, the optimal solution is updated as the current solution, the number of consecutive times the optimal solution has not been updated is updated to 0, the number of loops is increased by 1, and the process returns to S302d, otherwise, the process proceeds to the next step;

[0181] S302h, determine whether the disturbance condition is met, that is, whether the number of consecutive unupdated optimal solutions is not less than n. If so, randomly select a neighborhood solution as the current solution, update the number of consecutive unupdated optimal solutions to 0, increase the number of loops by 1, and return to S302d. Otherwise, increase the number of consecutive unupdated optimal solutions by 1, increase the number of loops by 1, and return to S302d.

[0182] Furthermore, the specific process of obtaining the optimal inspection speed solution by using the dynamic programming algorithm in step S302a is as follows:

[0183] S302a-1. Simplify the flight speed of the drone into integers, ranging from V min To V max , simplify the total power consumption of the drone to a certain inspection target into an integer ranging from 0 to Q, and simplify the total flight time of the drone to a certain inspection target into an integer ranging from 0 to T;

[0184] S302a-2, assuming that the speed at which the drone starts and ends inspection is V min ;

[0185] S302a-3, starting from the last inspection target and ending at the first inspection target, traverse each inspection target of each drone in the initial inspection path planning scheme;

[0186] S302a-4. For the last inspection target, all possible inspection speeds, power consumption and flight time of the previous inspection target are enumerated, and the best inspection benefit of the last inspection target is calculated according to the state transition formula. The state transition formula is as follows:

[0187]

[0188] (v0,Q K-1 ,T K-1 ,v K-1 ) represents the state, that is, the starting inspection speed is v0, and the total power consumption at the end of the K-1th inspection target is Q K-1 , the total flight time at the end of the K-1th inspection target is T K-1 , the inspection speed of the K-1th inspection target is v K-1 ; V K (v0,Q K-1 ,T K-1 ,v K-1 ) represents the total monitoring benefit from the Kth inspection target to the end of the inspection in this state; p K is the shortest monitoring time of the Kth inspection target; t K is the actual monitoring time of the Kth inspection target; p K (tK -p K ) represents the monitoring benefit of the Kth inspection target;

[0189] S302a-5. For other inspection targets, all possible inspection speeds, power consumption and flight time of the previous inspection target are enumerated, and the optimal inspection benefit of each inspection target is calculated according to the state transition formula. The state transition formula is as follows:

[0190]

[0191] S302a-6, increase the speed of the drone at the start and end of the inspection by 1, and return to S302a-3 until the drone speed exceeds V max Finally, the inspection benefits at each speed are compared, and the best inspection benefit and inspection plan are output. The total inspection benefit calculation formula is as follows:

[0192]

[0193] like Figure 3 As shown, in step S302d of this embodiment, when constructing the neighborhood solution space, the Relocate operator is first used to delete an inspection target from the current path and insert it into any position of any inspection path other than the current path; then the Exchange operator is used to exchange the positions of any two inspection targets in an inspection path; finally, the Cross operator is used to select any two inspection paths, select any inspection target in the path respectively, and exchange the inspection paths after the two targets.

[0194] like Figure 3 As shown in (a), the relocate operator of this embodiment executes the following steps:

[0195] (a) Starting from the first inspection target, delete it from the currently scheduled drone team and insert it into each inspection position of each other drone inspection team to generate a new path planning solution;

[0196] (b) Each time a new path planning scheme is generated, the maximum flight speed of the drone is used as the inspection speed to determine whether the inspection start time of each inspection target is earlier than the specified inspection start time. If it is earlier, the scheme is deleted; the minimum flight speed of the drone is used as the inspection speed to determine whether the inspection end time of each inspection target is later than the specified inspection end time. If it is later, the scheme is deleted;

[0197] (c) Using dynamic programming algorithm to find the optimal inspection speed for all generated solutions and calculate the cost of the solution;

[0198] (d) Compare the cost of each solution and select the relocate solution with the highest cost.

[0199] like Figure 3 As shown in (b), the execution steps of the exchange operator in this embodiment are:

[0200] (a) For each UAV in the inspection plan generated by the relocate operator, select two inspection targets in its inspection path and exchange the inspection order to generate a new path planning plan;

[0201] (b) Each time a new path planning scheme is generated, the maximum flight speed of the drone is used as the inspection speed to determine whether the inspection start time of each inspection target is earlier than the specified inspection start time. If it is earlier, the scheme is deleted; the minimum flight speed of the drone is used as the inspection speed to determine whether the inspection end time of each inspection target is later than the specified inspection end time. If it is later, the scheme is deleted;

[0202] (c) Using dynamic programming algorithm to find the optimal inspection speed for all generated solutions and calculate the cost of the solution;

[0203] (d) Compare the cost of each solution and select the exchange solution with the highest cost.

[0204] like Figure 3 As shown in (c), the cross operator execution steps of this embodiment are:

[0205] (a) For the inspection plan generated by the exchange operator, select any two drones, choose two inspection targets in the two inspection routes as exchange points, and exchange the inspection routes after the exchange points to generate a new path planning plan;

[0206] (b) Each time a new path planning scheme is generated, the maximum flight speed of the drone is used as the inspection speed to determine whether the inspection start time of each inspection target is earlier than the specified inspection start time. If it is earlier, the scheme is deleted; the minimum flight speed of the drone is used as the inspection speed to determine whether the inspection end time of each inspection target is later than the specified inspection end time. If it is later, the scheme is deleted;

[0207] (c) Using dynamic programming algorithm to find the optimal inspection speed for all generated solutions and calculate the cost of the solution;

[0208] (d) Compare the exchange costs of each pair of drones choosing each exchange point and select the optimal solution for each pair of drones to exchange.

[0209] Example:

[0210] according to Figure 4 and Figure 1 As shown in FIG, the BIM multiple UAV aircraft engineering site collaborative inspection and dispatching system includes environmental monitoring equipment, BIM three-dimensional models and multiple UAVs, and the UAVs are equipped with data collectors, data sensors, three-dimensional engine cameras, GPS positioning and cameras. Specifically, it can be divided into the construction management stage, the quality management stage and the UAV dispatching stage.

[0211] During the construction management stage, BIM Revit is used to generate a BIM 3D model of the engineering entity. The model is divided according to the project engineering, and the various professional models of the project are disassembled according to the components. The Timeliner function is used to simulate the construction progress, generate a Gantt chart, see the construction parts required for each time period, and generate a standard model with a construction progress plan.

[0212] During the quality management stage, Navisworks is used to animate the situation on the construction site, and the data in the BIM is exported in the form of IFC data. The IFC data is associated with the construction quality acceptance information, that is, the construction quality acceptance attributes are associated with the three-dimensional entity elements. BIMServe.org is used to automatically generate inspection batches and checkpoints for spot check plans, and the acceptance data is entered. After the inspection batches, inspection items and checkpoints are automatically generated, the current checkpoints can be displayed on the BIM three-dimensional model, and the generated relevant data is saved in the database.

[0213] In the UAV scheduling stage, based on the location information of the target points to be monitored and the monitoring resource information obtained from the BIM three-dimensional model, a multi-UAV collaborative inspection scheduling model is constructed with the goal of maximizing monitoring quality. The multi-UAV collaborative inspection scheduling model is solved by a hybrid algorithm of taboo search + dynamic programming, and finally the optimal inspection path planning and speed optimization solution of the UAV are obtained.

[0214] The environmental monitoring device of this embodiment includes a wind speed monitoring module, a wind direction monitoring module, a temperature monitoring module and a rainfall monitoring module. The environmental monitoring device transmits relevant data parameters to the drone setting module and the data acquisition module through the 5G network device.

[0215] The drone of this embodiment is equipped with a GPS positioning system, a three-dimensional engine camera, and related camera devices. The inspection route and related data determined therein are imported into the three-dimensional engine through BIM technology, and the corresponding relationship between the drone and the three-dimensional engine camera is established through the method of coordinate system conversion, and the inspection route of the drone is set. The drone conducts inspections according to the inspection route through the GPS positioning system and camera devices; the drone data acquisition module includes a data collector and a data sensor installed on the drone, which conducts a comprehensive analysis of the data parameters transmitted from the environmental monitor and the inspection information data to determine whether the cruising conditions are met. If the cruising conditions are not met (such as weather factors, etc.), the data acquisition module will return to the relevant BIM platform; if the cruising conditions are met, the drone automatically starts the inspection work according to the cruising route.

[0216] like Figure 3 As shown, Figure 3 (a) Schematic diagram of generating a feasible neighborhood space relocate operator; Figure 3 (b) Schematic diagram of generating feasible neighborhood space exchange operator; Figure 3 (c) is a schematic diagram of generating a feasible neighborhood space cross operator;

[0217] Figure 3 The solid line and dotted line in the figure represent two different inspection paths; the left figure shows the inspection path before the change, and the right figure shows the inspection path after the operator operation.

Claims

1. A BIM-based engineering site UAV collaborative inspection and dispatching method, characterized by: The following steps are involved: S1. Associate the construction quality acceptance attributes with the three-dimensional entity elements, and automatically generate the location information of the target points to be monitored and the information of the inspection batch, inspection items and inspection part attributes on the BIM three-dimensional model; S2. Based on the obtained target point information and monitoring resource information to be monitored, a multi-UAV collaborative inspection scheduling model is constructed with the goal of achieving the highest monitoring quality; S3. Solve the multi-UAV collaborative inspection scheduling model through the taboo search + dynamic programming hybrid algorithm, and finally obtain the optimal inspection path planning and speed optimization solution for the UAV; The construction content of the multi-UAV path optimization model in step S2 is as follows: (1) The objective function is shown in formula (1): i and j are monitoring point numbers, including 1 drone airport and N inspection targets. The drone airport number is 0, i∈[0,N], j∈[0,N], and i≠j; k is the drone number, with a total of K drones; p i is the shortest monitoring time of inspection target i; t ik is the actual monitoring time of the UAV numbered k on the inspection target i; x ijk is the decision variable. When the UAV numbered k flies from monitoring point i to monitoring point j, x ijk =1, otherwise, x ijk =0; (2) The constraints are shown in equations (2) to (20): L 0k =0,k=1,...,K (10) L ik =A ik +t ik ,i=1,...,N,k=1,...,K (12) A 0k ≤T,k=1,...,K (13) x ijk ∈{0,1},i=0,...,N,j=0,...,N,k=1,...,K,i≠j (20) in: Formula (2) indicates that each UAV can reach the UAV airport at most once; Formula (3) and Formula (4) indicate that each drone assigned with an inspection task must depart from the drone airport, and drones without inspection tasks will not take off, that is, to ensure that drones do not take off empty. M is a positive number greater than N+1. Formula (5) indicates that each inspection target can only be visited by one UAV; Formula (6) represents the in-and-out balance constraint of each monitoring point; Formula (7) indicates that the speed of each UAV assigned with inspection task from the UAV airport is equal to the speed of returning to the UAV airport; Formulas (8) and (9) indicate that the actual monitoring time of each UAV on each inspection target cannot be shorter than its minimum monitoring time; Formula (10) indicates that the time when each UAV assigned with inspection task departs from the UAV airport is 0; Formula (11) represents the time when each UAV assigned with inspection task arrives at each monitoring point; Formula (12) represents the time when each UAV assigned with inspection task leaves each monitoring point; Formula (13) indicates that the time for each UAV assigned with inspection mission to return to the UAV airport cannot exceed the maximum flight time of the UAV; Formula (14) and Formula (15) indicate that the time when each UAV assigned with an inspection task arrives at and leaves the scheduled inspection target satisfies the time window constraint; Formula (16) represents the power consumed by each UAV assigned with inspection tasks when accelerating or decelerating in each inspection target area; Formula (17) represents the power consumed by each UAV assigned with inspection tasks when flying at a uniform speed in each inspection target area; Formula (18) indicates that the power consumed by each drone assigned with inspection tasks cannot exceed the maximum power of the drone; Formula (19) indicates that the flight speed of each UAV assigned with inspection tasks satisfies the speed constraint; Formula (20) represents the value constraints of decision variables; Among them, d ij is the distance between monitoring point i and monitoring point j; l i is the monitoring distance length of inspection target i; [a i ,b i ] is the time period during which the inspection target i can be monitored; α is the acceleration of the UAV; -β is the acceleration of the UAV; ρ1 is the acceleration power of the UAV; ρ2 is the deceleration power of the UAV; Q is the maximum power of the UAV; T is the maximum flight time of the UAV; V min V is the slowest flight speed of the drone; max is the fastest flight speed of the drone; v ik A is the uniform flight speed of UAV k at monitoring point i; ik is the time when UAV k arrives at monitoring point i; L ik is the time when drone k leaves monitoring point i; q ik 1 is the power consumption of UAV k when accelerating or decelerating within the inspection target i; q ik 2 is the power consumption of UAV k flying at a constant speed within the inspection target i; The specific process of step S3 is: S301, generating an initial inspection path planning scheme according to a conservation algorithm; S301a, set up a separate drone for each inspection target for inspection; S301b, calculating the flight mileage of each UAV; S301c, according to the flight mileage, the corresponding saved mileage is calculated according to the saved mileage formula, and a saved mileage table is generated. The saved mileage formula is as follows: S ij =d i0 +d 0j -d ij ,i=1,…,N,j=1,…,N,i≠j S ij Indicates the number of UAV flight miles saved by removing inspection target j from the current UAV inspection path and inserting it after inspection target i; d ij is the distance between monitoring point i and monitoring point j; S301d, if there is a non-negative number in the mileage saving meter, proceed to step S301e, if all are negative numbers, end and obtain the initial path planning solution; S301e, arranging the saved mileage in descending order, starting from the two paths with the largest saved mileage, and calculating the time for the combined UAV to leave each surveillance target according to the lowest flight speed; S301f, if the latest departure time constraint formula (15) is satisfied, the two flight routes are merged and the process returns to step S301b; if not satisfied, the current solution is deleted from the mileage saving table and the process returns to step S301d; S302, optimizing the initial inspection path planning scheme by using a taboo search + dynamic programming hybrid algorithm to obtain an optimal drone inspection scheme; S302a, dynamically planning the initial inspection path planning scheme to obtain the optimal inspection speed; S302b, calculate the cost of the obtained inspection plan, and the cost function is: p i represents the shortest monitoring time of each inspection target i, t i Indicates the actual monitoring time of each inspection target i; S302c, setting parameters m and n, where m is the maximum number of cycles of the taboo search algorithm, and n is the number of consecutive unupdated optimal solutions that need to be accumulated when implementing the disturbance, starting from the first cycle; S302d, determine whether the current cycle exceeds the maximum number of cycles m, if so, end the cycle and output the optimal inspection plan and the maximum cost; otherwise, proceed to the next step; S302e, using the relocate operator, the exchange operator, and the cross operator in sequence to construct a neighborhood solution space for the current solution; S302f, extracting neighborhood solutions that meet the prohibition-breaking principle and are not tabooed in the neighborhood solution space, and selecting the inspection solution with the highest cost in the neighborhood solution as the current solution; S302g, if the current solution is higher than the optimal solution, the optimal solution is updated as the current solution, the number of consecutive times the optimal solution has not been updated is updated to 0, the number of loops is increased by 1, and the process returns to S302d, otherwise, the process proceeds to the next step; S302h, determine whether the disturbance condition is met, that is, whether the number of consecutive unupdated optimal solutions is not less than n. If so, randomly select a feasible neighborhood solution as the current solution, update the number of consecutive unupdated optimal solutions to 0, increase the number of loops by 1, and return to S302d. Otherwise, increase the number of consecutive unupdated optimal solutions by 1, increase the number of loops by 1, and return to S302d.

2. The method for coordinated inspection and dispatching of unmanned aerial vehicles on engineering sites based on BIM according to claim 1 is characterized in that: The specific process of using the dynamic programming algorithm to obtain the optimal inspection speed solution in step S302a is as follows: S302a-1. Simplify the flight speed of the drone into integers, ranging from V min To V max , simplify the total power consumption of the drone to a certain inspection target into an integer ranging from 0 to Q, and simplify the total flight time of the drone to a certain inspection target into an integer ranging from 0 to T; S302a-2, assuming that the speed at which the drone starts and ends inspection is V min ; S302a-3, starting from the last inspection target and ending at the first inspection target, traverse each inspection target of each drone in the initial inspection path planning scheme; S302a-4. For the last inspection target, all possible inspection speeds, power consumption and flight time of the previous inspection target are enumerated, and the best inspection benefit of the last inspection target is calculated according to the state transition formula. The state transition formula is as follows: V K (v0,Q K-1 ,T K-1 ,v K-1 )=max vK p K (t K -p K ) (v0,Q K-1 ,T K-1 ,v K-1 ) represents the state, that is, the starting inspection speed is v0, and the total power consumption at the end of the K-1th inspection target is Q K-1 , the total flight time at the end of the K-1th inspection target is T K-1 , the inspection speed of the K-1th inspection target is v K-1 ; V K (v0,Q K-1 ,T K-1 ,v K-1 ) represents the total monitoring benefit from the Kth inspection target to the end of the inspection in this state; p K is the shortest monitoring time of the Kth inspection target; t K is the actual monitoring time of the Kth inspection target; p K (t K -p K ) represents the monitoring benefit of the Kth inspection target; S302a-5. For other inspection targets, all possible inspection speeds, power consumption and flight time of the previous inspection target are enumerated, and the optimal inspection benefit of each inspection target is calculated according to the state transition formula. The state transition formula is as follows: S302a-6, increase the speed of the drone at the start and end of the inspection by 1, and return to S302a-3 until the drone speed exceeds V max Finally, the inspection benefits at each speed are compared, and the best inspection benefit and inspection plan are output. The total inspection benefit calculation formula is as follows:

3. The method for coordinated inspection and dispatching of unmanned aerial vehicles on engineering sites based on BIM according to claim 1 is characterized in that: When constructing the neighborhood solution space in step S302d, the relocate operator is first used to delete an inspection target from the current path and insert it into any position of any inspection path other than the current path; then the exchange operator is used to exchange the positions of any two inspection targets in an inspection path; finally, the cross operator is used to select any two inspection paths, select any inspection target in the path respectively, and exchange the inspection paths after the two targets.

4. The method for coordinated inspection and dispatching of unmanned aerial vehicles at engineering sites based on BIM according to claim 3 is characterized in that: The relocate operator executes the following steps: (a) Starting from the first inspection target, delete it from the currently scheduled drone team and insert it into each inspection position of each other drone inspection team to generate a new path planning solution; (b) Each time a new path planning scheme is generated, the maximum flight speed of the drone is used as the inspection speed to determine whether the inspection start time of each inspection target is earlier than the specified inspection start time. If it is earlier, the scheme is deleted; the minimum flight speed of the drone is used as the inspection speed to determine whether the inspection end time of each inspection target is later than the specified inspection end time. If it is later, the scheme is deleted; (c) Using dynamic programming algorithm to find the optimal inspection speed for all generated solutions and calculate the cost of the solution; (d) Compare the cost of each solution and select the relocate solution with the highest cost.

5. The method for coordinated inspection and dispatching of unmanned aerial vehicles at engineering sites based on BIM according to claim 3 is characterized in that: The exchange operator executes the following steps: (a) For each UAV in the inspection plan generated by the relocate operator, select two inspection targets in its inspection path and exchange the inspection order to generate a new path planning plan; (b) Each time a new path planning scheme is generated, the maximum flight speed of the drone is used as the inspection speed to determine whether the inspection start time of each inspection target is earlier than the specified inspection start time. If it is earlier, the scheme is deleted; the minimum flight speed of the drone is used as the inspection speed to determine whether the inspection end time of each inspection target is later than the specified inspection end time. If it is later, the scheme is deleted; (c) Using dynamic programming algorithm to find the optimal inspection speed for all generated solutions and calculate the cost of the solution; (d) Compare the cost of each solution and select the exchange solution with the highest cost.

6. The BIM-based engineering site UAV collaborative inspection and dispatching method according to claim 3 is characterized by: The cross operator performs the following steps: (a) For the inspection plan generated by the exchange operator, select any two drones, choose two inspection targets in the two inspection routes as exchange points, and exchange the inspection routes after the exchange points to generate a new path planning plan; (b) Each time a new path planning scheme is generated, the maximum flight speed of the drone is used as the inspection speed to determine whether the inspection start time of each inspection target is earlier than the specified inspection start time. If it is earlier, the scheme is deleted; the minimum flight speed of the drone is used as the inspection speed to determine whether the inspection end time of each inspection target is later than the specified inspection end time. If it is later, the scheme is deleted; (c) Using dynamic programming algorithm to find the optimal inspection speed for all generated solutions and calculate the cost of the solution; (d) Compare the exchange costs of each pair of drones choosing each exchange point and select the optimal solution for each pair of drones to exchange.

7. A system for implementing the BIM-based engineering site UAV collaborative inspection and scheduling method according to any one of claims 1 to 6, characterized in that: It includes environmental monitoring equipment, BIM three-dimensional models and multiple drones, each of which is equipped with a data collector, a data sensor, a three-dimensional engine camera, a GPS positioning system and a camera; BIM Revit builds a 3D model of the project site, uses Timeliner to generate a standard model with a construction schedule, and uses Navisworks for animation simulation. Finally, BIM Serve.org is used to generate quality checkpoint information, which is then used to optimize the inspection route. The data in the BIM 3D model is converted into coordinate data, and the corresponding relationship between the BIM 3D model and the drone is established through the 3D engine camera, and the inspection information is input into the drone; The drone receives the BIM data information and the relevant information of the environmental monitoring equipment on site through the 5G device network, and then determines whether to start the inspection. If the conditions are met, the inspection will start; During inspections, on-site data is recorded through data collectors and data sensors, and the inspection data is transmitted back to the BIM platform in real time through 5G network equipment. Finally, data comparison is performed to determine whether manual on-site review is required.