Air-ground combined inspection method and system

Abstract

The invention discloses an air-ground combined inspection method and system. The air-ground combined inspection method comprises the following steps: a data center sends an unmanned vehicle inspection signal and an inspection scheme to an unmanned vehicle nest controller; the unmanned vehicle nest generates a plurality of unmanned vehicle control information; the unmanned vehicle nest sends the unmanned vehicle control information to the unmanned vehicle; each unmanned vehicle arrives at a specified inspection range; the unmanned aerial vehicle nest sends an in-place signal and basic information of each unmanned vehicle to the data center after each unmanned vehicle is located in a specified inspection range; the data center sends an unmanned aerial vehicle relay information transmission scheme to the unmanned aerial vehicle nest; the unmanned aerial vehicle nest sends relay transmission signals to a plurality of unmanned aerial vehicles; all the unmanned aerial vehicles cooperatively fly, and the transmission information of all the unmanned vehicles located in the designated inspection range is collected in the flying process. Under the condition that the network signal is weak, the purpose of real-time information transmission is achieved through cooperation of the unmanned aerial vehicle and the unmanned aerial vehicle.

Description

Technical Field

[0001] The present application relates to the technical field of multi-air-ground joint inspection, and in particular to an air-ground joint inspection method and an air-ground joint inspection system. Background Art

[0002] In the prior art, when an unmanned vehicle is conducting an inspection, it usually takes photos or detects (such as air composition detection, temperature detection, etc.) at a predetermined inspection point. At this time, the unmanned vehicle cannot move and can only transmit data to the data center through the network. However, once the network is unstable or the network is not good, data transmission will become difficult. Summary of the invention

[0003] The purpose of the present invention is to provide an air-ground joint inspection method to solve at least one of the above-mentioned technical problems.

[0004] One aspect of the present invention provides an air-ground joint inspection method, the air-ground joint inspection method comprising:

[0005] The data center generates unmanned vehicle inspection signals and unmanned vehicle inspection plans;

[0006] The data center sends the unmanned vehicle inspection signal and inspection plan to the unmanned vehicle nest controller;

[0007] The unmanned vehicle nest controller generates a plurality of unmanned vehicle control information according to the inspection plan, and one unmanned vehicle control information corresponds to one unmanned vehicle;

[0008] The unmanned vehicle nest sends the unmanned vehicle control information to the corresponding unmanned vehicles respectively;

[0009] Each unmanned vehicle reaches the designated inspection range according to the unmanned vehicle control information obtained;

[0010] The drone nest obtains the location information transmitted by each unmanned vehicle, and when each unmanned vehicle is located in the designated inspection range, it sends a location signal and basic information of each unmanned vehicle to the data center;

[0011] The data center sends the drone relay information transmission plan to the drone nest based on the arrival signal and the basic information of each unmanned vehicle;

[0012] The drone nest sends relay transmission signals to multiple drones according to the drone relay information transmission scheme obtained;

[0013] Each drone performs collaborative flight based on the acquired relay transmission signal, and collects transmission information from each unmanned vehicle within the designated inspection range during the flight.

[0014] Optionally, the inspection plan includes multiple inspection locations and the number of unmanned vehicles required for each inspection location.

[0015] Optionally, the data center sends the drone relay information transmission scheme to the drone nest according to the arrival signal, including:

[0016] Divide each inspection location to generate multiple basic inspection areas;

[0017] A drone inspection strategy model and drone quantity information are generated based on the basic information of each unmanned vehicle and each basic inspection area; the drone inspection strategy model and drone quantity information are used as a drone relay information transmission scheme.

[0018] Optionally, the drone nest sending a relay transmission signal to a plurality of drones according to the acquired drone relay information transmission scheme includes:

[0019] According to the drone quantity information, a drone inspection strategy model and a relay transmission signal are sent to each drone that meets the drone quantity information.

[0020] Optionally, the air-ground joint inspection method further comprises:

[0021] After each drone collects the transmission information of each unmanned vehicle within the designated inspection range, each drone transmits the data to the data center through relay transmission.

[0022] Optionally, the air-ground joint inspection method further comprises:

[0023] After receiving the preset number of rounds of data, the data center generates a zone change inspection signal and an unmanned vehicle zone change inspection plan;

[0024] The data center sends the zone change inspection signal and the unmanned vehicle zone change inspection plan to the unmanned vehicle nest controller;

[0025] The unmanned vehicle nest controller determines whether the number of unmanned vehicles currently inspected meets the zone change requirements according to the unmanned vehicle zone change inspection plan. If so,

[0026] The unmanned vehicle nest controller generates a plurality of zone-changing unmanned vehicle control information according to the unmanned vehicle zone-changing inspection scheme, wherein one zone-changing unmanned vehicle control information corresponds to an unmanned vehicle that is already inspecting;

[0027] The unmanned vehicle nest sends the control information of each unmanned vehicle in the change zone to the corresponding unmanned vehicle;

[0028] Each unmanned vehicle arrives at a new designated inspection range based on the unmanned vehicle control information obtained;

[0029] The drone nest obtains the location information transmitted by each unmanned vehicle, and sends a location signal to the data center when each unmanned vehicle is located in the designated inspection range;

[0030] The data center sends a parameter update plan to the drone nest based on the arrival signal;

[0031] The drone nest sends updated parameters to multiple drones according to the parameter update scheme;

[0032] Each UAV performs collaborative flight according to the updated parameters and collects transmission information from each unmanned vehicle within the new designated inspection range during the flight.

[0033] Optionally, each basic inspection area is obtained by a clustering algorithm.

[0034] The present application also provides an air-ground joint inspection system, which includes a data center, a drone nest, and an unmanned vehicle nest. A plurality of drones are arranged in the drone nest, and a plurality of unmanned vehicles are arranged in the unmanned vehicle nest; wherein:

[0035] The data center, drone nest, and unmanned vehicle nest cooperate to implement the air-ground joint inspection method as described above.

[0036] Beneficial Effects

[0037] The air-ground joint inspection method of the present application can achieve the purpose of real-time information transmission through the cooperation of drones and unmanned vehicles when the network signal is weak. That is, the unmanned vehicles are used to collect information, and the network composed of drones forms a transmission chain, thereby solving the problem that the existing technology cannot transmit distant data when there is no network or the network is weak. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] Figure 1 It is a flowchart of an air-ground joint inspection method according to an embodiment of the present application. DETAILED DESCRIPTION

[0039] In order to make the purpose, technical scheme and advantages of the implementation of this application clearer, the technical scheme in the embodiment of this application will be described in more detail below in conjunction with the drawings in the embodiment of this application. In the drawings, the same or similar reference numerals throughout represent the same or similar elements or elements with the same or similar functions. The described embodiments are part of the embodiments of this application, not all of them. The embodiments described below with reference to the drawings are exemplary and are intended to be used to explain this application, and should not be construed as limitations on this application. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of this application. The embodiments of this application are described in detail below in conjunction with the drawings.

[0040] Figure 1It is a flowchart of an air-ground joint inspection method according to an embodiment of the present application.

[0041] like Figure 1 The air-ground joint inspection method shown includes:

[0042] Step 1: The data center generates an unmanned vehicle inspection signal and an unmanned vehicle inspection plan;

[0043] Step 2: The data center sends the unmanned vehicle inspection signal and inspection plan to the unmanned vehicle nest controller;

[0044] Step 3: The unmanned vehicle nest controller generates a plurality of unmanned vehicle control information according to the inspection plan, and one unmanned vehicle control information corresponds to one unmanned vehicle;

[0045] Step 4: The unmanned vehicle nest sends the unmanned vehicle control information to the corresponding unmanned vehicles;

[0046] Step 5: Each unmanned vehicle arrives at the designated inspection range according to the unmanned vehicle control information obtained;

[0047] Step 6: The drone nest obtains the location information transmitted by each unmanned vehicle, and when each unmanned vehicle is located in the designated inspection range, sends a location signal and basic information of each unmanned vehicle to the data center;

[0048] Step 7: The data center sends the drone relay information transmission plan to the drone nest based on the arrival signal and the basic information of each unmanned vehicle;

[0049] Step 8: The drone nest sends a relay transmission signal to multiple drones according to the drone relay information transmission scheme obtained;

[0050] Step 9: Each UAV performs cooperative flight based on the acquired relay transmission signal, and collects transmission information of each unmanned vehicle within the specified inspection range during the flight.

[0051] The air-ground joint inspection method of the present application can achieve the purpose of real-time information transmission through the cooperation of drones and unmanned vehicles when the network signal is weak, thereby solving the problem that the prior art cannot transmit distant data when there is no network or the network is weak.

[0052] In this embodiment, the inspection plan includes multiple inspection locations and the number of unmanned vehicles required for each inspection location.

[0053] In this embodiment, the data center sends the drone relay information transmission scheme to the drone nest according to the arrival signal and the basic information of each unmanned vehicle, including:

[0054] Divide each inspection location to generate multiple basic inspection areas;

[0055] A drone inspection strategy model and drone quantity information are generated based on the basic information of each unmanned vehicle and each basic inspection area; the drone inspection strategy model and drone quantity information are used as a drone relay information transmission scheme.

[0056] In this embodiment, each basic inspection area can be obtained by using a clustering algorithm.

[0057] For example, assuming there are 50 inspection locations, a set of 50 inspection locations (0, 1, 2, 3...50) can be generated;

[0058] For this set, calculate the sample density ρ(k) of all inspection locations and take That is, C1 is a center point.

[0059] The other inspection locations are judged as follows:

[0060] Determine whether the distance of the center point minus the distance of any inspection location is less than the average distance. If so, the inspection location with a distance less than the average distance belongs to the same class as the center point.

[0061] After judging all the inspection locations in the set of 50 inspection locations, each inspection location belonging to the same category as the central point can be obtained, and each inspection location belonging to the same category and the central point constitute the first basic inspection area.

[0062] When the first basic inspection area is formed, the inspection locations after the inspection locations that have formed the first basic inspection area are discarded from the set of inspection locations. For example, if there are 10 inspection locations in the first basic inspection area, there are 40 inspection locations left.

[0063] The remaining 40 inspection locations are clustered using the above method, and this is repeated until all inspection locations have corresponding basic inspection areas.

[0064] In this embodiment, the drone nest sends a relay transmission signal to multiple drones according to the acquired drone relay information transmission scheme, including:

[0065] According to the drone quantity information, a drone inspection strategy model and a relay transmission signal are sent to each drone that meets the drone quantity information.

[0066] In this embodiment, the air-ground joint inspection method further includes:

[0067] After each drone collects the transmission information of each unmanned vehicle within the designated inspection range, each drone transmits the data to the data center through relay transmission.

[0068] In this embodiment, each drone will perform the following process when performing a mission:

[0069] 1. The drones depart from the drone nest and head to their respective target areas;

[0070] 2. The drones interact with the unmanned vehicles according to their own stored drone inspection strategy models to complete the data transmission task;

[0071] 3. The drone returns to the drone nest.

[0072] In this embodiment, the drone inspection strategy model of the present application is a trained POMDP model framework.

[0073] In this embodiment, the POMDP model framework is trained using the following method:

[0074] Randomly initialize the parameters ω, ε1, ε2 of the Actor network and the two Critic networks;

[0075] Randomly initialize the embedding table parameters ζ and encoder parameters χ e , decoder parameters χ d ;

[0076] Initialize the total number of data in the experience pool D;

[0077] Start an interaction round, initialize the environment, and obtain the initial state set S;

[0078] According to the state information s of each drone in the state set S obtained at this moment i Input the observed state information to the actor network corresponding to each drone as the input of the Actor to obtain the discrete latent action z k and continuous hidden actions

[0079] The discrete latent action is sent as input to the embedding table, and the discrete action with the highest matching degree is obtained by using the similarity principle;

[0080] Discrete latent action z k It is concatenated with the input observation state information into a conditional vector and combined with the continuous hidden action Multiply as the input of the decoder to get the continuous action x k ;

[0081] Combine discrete actions k with continuous actions x k Input into the environment and get the next state s -, immediate return r and end mark Done;

[0082] Will Deposit into experience replay pool D;

[0083] After a certain number of steps, the neural network is trained until the maximum number of steps in a training round is reached;

[0084] The conditional variational autoencoder is trained after a certain number of rounds until the maximum number of training rounds is reached.

[0085] In this embodiment, the drone inspection strategy model of the present application includes an encoder, a decoder and an embedding table, and the trainable parameters are respectively the embedding table parameter ζ, the encoder parameter χ e , decoder parameters χ d .

[0086] The embedding table takes the unique hot encoding k of the discrete action as input and obtains the discrete latent action z k . It is concatenated with the state s as the conditional vector of the drone inspection strategy model and combined with the continuous action x k Multiply as the input of the encoder, and the output is the continuous hidden action mean μ and the continuous hidden action standard deviation σ of the same dimension. Re-parameterize the hidden action mean and standard deviation to get the hidden action z k , and multiplied by the conditional vector of CVAE as the input of the decoder, and finally the reconstructed continuous action x is obtained k- and the state transfer residual Δs - .

[0087] In this embodiment, the drone inspection strategy model of the present application includes an Actor network and its target network Actor', two Critic networks and their respective target networks Critic'.

[0088] In this embodiment, the basic information required for this application is as follows:

[0089] The number of unmanned vehicles, the number of drones, the time of each mission or the preset number of rounds, and the basic information of drones (for example, the drone's sensing range, altitude, safe flight distance, maximum flight speed, and transmission power). The data transmission power of the unmanned vehicle, the Poisson distribution parameters and the Shota distribution parameters in the process of sensing data packets, and the energy consumption of sensing data packets. The channel gain at a reference distance of 1m, the drone's receiving noise, the transmission bandwidth, and the energy transmission efficiency. The drone's flight energy consumption parameters.

[0090] In this embodiment, the air-ground joint inspection method further includes:

[0091] After receiving the preset number of rounds of data, the data center generates a zone change inspection signal and an unmanned vehicle zone change inspection plan;

[0092] The data center sends the zone change inspection signal and the unmanned vehicle zone change inspection plan to the unmanned vehicle nest controller;

[0093] The unmanned vehicle nest controller determines whether the number of unmanned vehicles currently inspected meets the zone change requirements according to the unmanned vehicle zone change inspection plan. If so,

[0094] The unmanned vehicle nest controller generates a plurality of zone-changing unmanned vehicle control information according to the unmanned vehicle zone-changing inspection scheme, wherein one zone-changing unmanned vehicle control information corresponds to an unmanned vehicle that is already inspecting;

[0095] The unmanned vehicle nest sends the control information of each unmanned vehicle in the change zone to the corresponding unmanned vehicle;

[0096] Each unmanned vehicle arrives at a new designated inspection range based on the unmanned vehicle control information obtained;

[0097] The drone nest obtains the location information transmitted by each unmanned vehicle, and sends a location signal to the data center when each unmanned vehicle is located in the designated inspection range;

[0098] The data center sends a parameter update plan to the drone nest based on the arrival signal;

[0099] The drone nest sends updated parameters to multiple drones according to the parameter update scheme;

[0100] Each UAV performs collaborative flight according to the updated parameters and collects transmission information from each unmanned vehicle within the new designated inspection range during the flight.

[0101] In this embodiment, the data center sends a parameter update scheme to the drone nest according to the arrival signal, including:

[0102] The data center obtains a parameter database, wherein the parameter database includes at least one preset unmanned vehicle zone change inspection scheme and a parameter update scheme corresponding to each preset unmanned vehicle zone change inspection scheme;

[0103] Obtain the parameter update plan corresponding to the preset unmanned vehicle zone change inspection plan that is the same as the current unmanned vehicle zone change inspection plan.

[0104] In actual application, each unmanned vehicle may have to perform multiple tasks. If the method of this application needs to be repeated after completing a task, the unmanned vehicle will need to return to the unmanned vehicle nest and the drone will need to return to the drone nest to re-apply the plan of this application. By pre-setting the possible unmanned vehicle zone change inspection plan to the parameter database, the parameters in the corresponding drone inspection strategy model of each drone can be directly replaced when changing the plan. On the one hand, there is no need for the drone and unmanned vehicle to return to the nest for updating. On the other hand, by pre-setting the parameters, there is no need to return the drone to the drone nest to retrain the model every time.

[0105] The present application also provides an air-ground joint inspection system, which includes a data center, a drone nest, and an unmanned vehicle nest. A plurality of drones are arranged in the drone nest, and a plurality of unmanned vehicles are arranged in the unmanned vehicle nest; wherein:

[0106] The data center, drone nest, and unmanned vehicle nest cooperate to implement the air-ground joint inspection method as described above.

[0107] Although the present invention has been described in detail above with general descriptions and specific embodiments, it is obvious to those skilled in the art that some modifications or improvements may be made thereto based on the present invention. Therefore, these modifications or improvements made without departing from the spirit of the present invention all fall within the scope of protection claimed by the present invention.

Claims

1. A method for joint air-ground inspection, characterized in that: The air-ground joint inspection method comprises: The data center generates unmanned vehicle inspection signals and unmanned vehicle inspection plans; The data center sends the unmanned vehicle inspection signal and inspection plan to the unmanned vehicle nest controller; The unmanned vehicle nest controller generates a plurality of unmanned vehicle control information according to the inspection plan, and one unmanned vehicle control information corresponds to one unmanned vehicle; The unmanned vehicle nest sends the unmanned vehicle control information to the corresponding unmanned vehicles respectively; Each unmanned vehicle reaches the designated inspection range according to the unmanned vehicle control information obtained; The drone nest obtains the location information transmitted by each unmanned vehicle, and when each unmanned vehicle is located in the designated inspection range, it sends a location signal and basic information of each unmanned vehicle to the data center; The data center sends the drone relay information transmission plan to the drone nest based on the arrival signal and the basic information of each unmanned vehicle; The drone nest sends relay transmission signals to multiple drones according to the drone relay information transmission scheme obtained; Each drone performs collaborative flight based on the acquired relay transmission signal, and collects transmission information from each unmanned vehicle within the designated inspection range during the flight.

2. The air-ground joint inspection method according to claim 1, characterized in that: The inspection plan includes multiple inspection locations and the number of unmanned vehicles required for each inspection location.

3. The air-ground joint inspection method according to claim 2, characterized in that: The data center sends the drone relay information transmission scheme to the drone nest according to the arrival signal and the basic information of each unmanned vehicle, including: Divide each inspection location to generate multiple basic inspection areas; A drone inspection strategy model and drone quantity information are generated based on the basic information of each unmanned vehicle and each basic inspection area; the drone inspection strategy model and drone quantity information are used as a drone relay information transmission scheme.

4. The air-ground joint inspection method according to claim 3, characterized in that: The drone nest sends a relay transmission signal to multiple drones according to the drone relay information transmission scheme obtained, including: According to the drone quantity information, a drone inspection strategy model and a relay transmission signal are sent to each drone that meets the drone quantity information.

5. The air-ground joint inspection method according to claim 4, characterized in that: The air-ground joint inspection method further comprises: After each drone collects the transmission information of each unmanned vehicle within the designated inspection range, each drone transmits the data to the data center through relay transmission.

6. The air-ground joint inspection method according to claim 5, characterized in that: The air-ground joint inspection method further comprises: After receiving the preset number of rounds of data, the data center generates a zone change inspection signal and an unmanned vehicle zone change inspection plan; The data center sends the zone change inspection signal and the unmanned vehicle zone change inspection plan to the unmanned vehicle nest controller; The unmanned vehicle nest controller determines whether the number of unmanned vehicles currently inspected meets the zone change requirements according to the unmanned vehicle zone change inspection plan. If so, The unmanned vehicle nest controller generates a plurality of zone-changing unmanned vehicle control information according to the unmanned vehicle zone-changing inspection scheme, wherein one zone-changing unmanned vehicle control information corresponds to an unmanned vehicle that is already inspecting; The unmanned vehicle nest sends the control information of each unmanned vehicle in the change zone to the corresponding unmanned vehicle; Each unmanned vehicle arrives at a new designated inspection range based on the unmanned vehicle control information obtained; The drone nest obtains the location information transmitted by each unmanned vehicle, and sends a location signal to the data center when each unmanned vehicle is located in the designated inspection range; The data center sends a parameter update plan to the drone nest based on the arrival signal; The drone nest sends updated parameters to multiple drones according to the parameter update scheme; Each UAV performs collaborative flight according to the updated parameters and collects transmission information from each unmanned vehicle within the new designated inspection range during the flight.

7. The air-ground joint inspection method according to claim 6, characterized in that: Each basic inspection area is obtained through clustering algorithm.

8. An air-ground joint inspection system, characterized in that: The air-ground joint inspection system includes a data center, a drone nest, and an unmanned vehicle nest. The drone nest is provided with a plurality of drones, and the unmanned vehicle nest is provided with a plurality of unmanned vehicles; wherein, The data center, drone nest, and unmanned vehicle nest cooperate to implement the air-ground joint inspection method as described in any one of claims 1 to 7.

Images