Environment detection method and system for underwater multiple autonomous robotic fishes

Abstract

The invention discloses an environment detection method for underwater multiple autonomous robotic fishes, which comprises the following steps: firstly, constructing an interconnection coupling immune cooperation detection model of the underwater multiple autonomous robotic fishes and carrying out parameter initialization, acquiring surrounding obstacles and environment information of a detected area by the robotic fishes, and determining an original antigen Eo of the obstacles and an original antigen Ed of the detected area; decomposition antigens facing obstacles and detected areas are determined, so that the concentrations of all antibodies of the robotic fish are determined, and the antibodies are selected based on the concentrations of all the antibodies; and completing the next detection of the robotic fish based on the selected antibody behavior until the detection is completed. According to the scheme, the environment of the robotic fish is used as an antigen, the robotic fish is used as a B cell, the behavior of the robotic fish is used as an antibody, and an interconnection coupling immune cooperation detection network is constructed facing obstacles and a detected area, so that the balance of the robotic fish during processing of two kinds of environment information is realized, the coverage rate during environment detection is improved, the detection repetition rate is reduced, and the detection efficiency is improved. The robotic fish can be quickly and effectively transferred in an unknown complex environment, and the detection efficiency is improved.

Description

Technical Field

[0001] The present invention belongs to the field of underwater robot environment detection, and specifically relates to an underwater multi-autonomous robot fish environment detection method and system Background Art

[0002] In recent years, with the continuous development and utilization of marine resources, underwater target detection has attracted more and more attention. Bionic robot fish has the advantages of extremely high energy utilization, high speed, strong maneuverability, low noise, and low environmental disturbance, which has gotten rid of the shortcomings of traditional propeller propulsion underwater robots such as high noise, low efficiency, and high energy consumption. Compared with a single robot fish, the collaboration between multiple robot fish can significantly improve the task execution capability and efficiency.

[0003] How to improve the quality and efficiency of environmental detection in unknown and complex environments has always been a difficult and hot topic in the research of underwater autonomous robots. Although relevant research results have emerged in recent years, there are still the following shortcomings: (1) When designing environmental detection methods, robots are always regarded as particles, and the motion constraints of robots are not considered, which reduces the effectiveness and practicality of the methods; (2) The flexibility of robots' autonomous selection of detection behaviors in complex environments is not enough; (3) The collaboration mode between multiple robots is relatively simple. Therefore, it is of great significance to improve the environmental detection method by learning from the bionic intelligent information processing mechanism, so as to improve the detection quality and efficiency in unknown and complex environments, which will promote the development of my country's underwater robot industry and help my country's strategy of building a strong maritime nation. Summary of the invention

[0004] In view of the above problems, the object of the present invention is to provide an environment detection method and system for underwater multi-autonomous robotic fish.

[0005] The specific technical solution for achieving the purpose of the present invention is: An underwater multi-autonomous robotic fish environment detection method comprises the following steps: Step 1: Construct an interconnected coupled immune collaborative detection model of underwater multi-autonomous robotic fish; Step 2, initializing the parameters in the interconnected coupled immune cooperative detection model; Step 3: The robot fish obtains environmental information about surrounding obstacles and detected areas, and determines the original antigen E of the obstacle. o and the original antigen E d , and determine the decomposition antigens facing obstacles and explored areas; Step 4: Based on the information obtained in step 3, determine the antibody concentrations of the robotic fish, and select antibodies based on the antibody concentrations; Step 5: Based on the selected antibody behavior, the next step of detection of the robot fish is completed; Step 6: Determine whether the detection task is completed. If so, end; otherwise, go to step 3.

[0006] An underwater multi-autonomous robotic fish environment detection system includes the following modules: Interconnected coupling immune cooperative detection model construction module: used to build an interconnected coupling immune cooperative detection model for underwater multi-autonomous robotic fish and initialize parameters; Information acquisition module: used by the robot fish to obtain environmental information about surrounding obstacles and detected areas, and determine the original antigen E of the obstacles. o and the original antigen E d , and determine the decomposition antigens facing obstacles and explored areas; Antibody selection module: used to determine the concentration of each antibody of the robot fish based on the acquired information, and select antibodies based on each antibody concentration; Detection module: used to complete the detection of the robot fish based on the selected antibody behavior until the detection task is completed.

[0007] Compared with the prior art, the present invention has the following beneficial effects: (1) The solution of the present invention uses the environment of the robot fish as an antigen, the robot fish as a B cell, and the behavior of the robot fish as an antibody, and constructs an interconnected coupled immune cooperative detection network for obstacles and detected areas, thereby achieving a balance when the robot fish processes the two types of environmental information, improving the coverage rate during environmental detection, and reducing the detection repetition rate; (2) The present invention divides the detection distance of the robot fish in the three forward directions into three levels: far, medium and near, and performs antigen encoding of environmental information based on binary, which is conducive to the accurate calculation of the coefficients of mutual stimulation and inhibition between antigens and antibodies, and antibodies, and realizes the precise selection of antibodies, which in turn helps to improve the detection quality and efficiency of the robot fish; (3) The present invention draws on the T cell effect and uses the avoidance distance and avoidance angle of the robot fish facing obstacles or detected areas as decomposition antigens. By defining the antibody concentration adjustment factor for decomposing the antigen by the avoidance distance and the potential field guidance factor for decomposing the antigen by the avoidance angle, the robot fish can be quickly and effectively transferred in an unknown and complex environment, thereby improving the detection efficiency.

[0008] The present invention is further described below in conjunction with specific implementation modes. BRIEF DESCRIPTION OF THE DRAWINGS

[0009] Figure 1 The present invention is a flow chart of an underwater multi-autonomous robotic fish environment detection method.

[0010] Figure 2 Schematic diagram of the structure of a three-joint bionic robotic fish in an embodiment of the present invention.

[0011] Figure 3 Schematic diagram of the robot fish environment detection in an embodiment of the present invention.

[0012] Figure 4 This is a schematic diagram of Jerne's unique immune network.

[0013] Figure 5 Schematic diagram of BT cell immunity model.

[0014] Figure 6 Schematic diagram of the interconnected coupled immune collaborative detection network.

[0015] Figure 7 This is a schematic diagram of the artificial potential field guidance of the underwater autonomous robotic fish of the present invention.

[0016] Figure 8 This is detection environment 1 in the embodiment of the present invention.

[0017] Fig. 9 This is detection environment 2 in the embodiment of the present invention.

[0018] Fig.10 This is detection environment three in the embodiment of the present invention.

[0019] Fig.11 This is detection environment four in the embodiment of the present invention.

[0020] Fig.12 This is a schematic diagram of the detection results of four robot fish using the CDMBIM method for environment one in an embodiment of the present invention.

[0021] Fig.13 This is a schematic diagram of the detection results of four robot fish using the ICDMBBE method for environment one in an embodiment of the present invention.

[0022] Fig.14 This is a schematic diagram of the detection results of four robot fish using the ICEDMBTCE method for environment one in an embodiment of the present invention.

[0023] Fig.15 This is a schematic diagram of the detection results of four robot fish using the CDMBIM method in environment 2 in an embodiment of the present invention.

[0024] Fig.16 This is a schematic diagram of the detection results of four robot fish using the ICDMBBE method in environment two in an embodiment of the present invention.

[0025] Fig.17 This is a schematic diagram of the detection results of four robot fish using the ICEDMBTCE method in environment two in an embodiment of the present invention.

[0026] Fig.18 This is a schematic diagram of the detection results of four robot fish using the CDMBIM method in environment three in an embodiment of the present invention.

[0027] Fig.19 This is a schematic diagram of the detection results of four robot fish using the ICDMBBE method in environment three in an embodiment of the present invention.

[0028] Fig. 20 This is a schematic diagram of the detection results of four robot fish using the ICEDMBTCE method in environment three in an embodiment of the present invention.

[0029] Fig.21 This is a schematic diagram of the detection results of four robot fish using the CDMBIM method in environment four in an embodiment of the present invention.

[0030] Fig. 22 This is a schematic diagram of the detection results of four robot fish using the ICDMBBE method in environment four in an embodiment of the present invention.

[0031] Fig.23 This is a schematic diagram of the detection results of four robot fish using the ICEDMBTCE method in environment four in an embodiment of the present invention. DETAILED DESCRIPTION

[0032] Example

[0033] Combination Figure 1 , an underwater multi-autonomous robotic fish environment detection method, comprising the following steps: Step 1: Based on the T cell effect, an interconnected coupled immune cooperative detection model of underwater multi-autonomous robotic fish is constructed: like Figure 2 As shown, the underwater robot fish for unknown environment detection of the present invention mainly includes three joints and is driven by servos respectively. The upper, lower and tail parts of the robot fish are respectively equipped with dorsal fins, pectoral fins and caudal fins. The communication module, control system and power module of the robot fish are located in the head of the robot fish. The movement of the robot fish is mainly achieved by driving three servos to realize the swing and propulsion of the fish body, so there are constraints on the turning angle when it swims in the water.

[0034] Robot fish detection mainly uses sensors to detect environmental information in the front area. In order to facilitate the design of the detection method of the present invention and improve the detection quality and efficiency of the robot fish, the robot fish environment is grid modeled. According to the motion constraints of the robot fish, its forward-looking left and right detection angles are constrained to [-45°, 45°], that is, the robot fish detection direction D = {straight ahead, 45° left, 45° right}; the detection distance of the robot fish is divided into three levels: "far", "middle" and "near", and the specific distances of the three levels here can be determined according to actual needs. The movement direction M of the robot fish is consistent with the detection direction, that is, M = {forward, 45° left, 45° right}, and is represented by l = {1, 2, 3}, such as Figure 3 As shown in the figure, black represents obstacles; the cross-hatching grid represents the detected area; and the blank grid represents the undetected area.

[0035] In order to realize the detection of underwater multi-autonomous robot fish in unknown complex environments, the present invention draws on the following Figure 4 The unique immune network hypothesis shown in the figure takes the environment of the robot fish as an antigen, the robot fish as a B cell, and the behavior of the robot fish as an antibody. Based on the immune network dynamics model, the robot fish selects antibodies (i.e., behaviors) according to the stimulation and inhibition between antigens and antibodies, and antibodies to achieve environmental detection. Therefore, the correctness of the robot fish's antibody selection determines the quality and efficiency of the robot fish's environmental detection. In order to further improve the detection quality and efficiency of multiple robot fish in unknown complex environments, the present invention once again draws on the T cell effect in the biological immune system to improve the design of the immune cooperative detection network model and method.

[0036] In the biological immune system, T cells are mainly derived from lymphocyte stem cells in the bone marrow. After differentiation and maturation in the thymus, they are distributed to immune organs and tissues throughout the body through blood circulation. T cells do not produce antibodies themselves, but can only inform B cells of the recognition results through epitopes. When antigens invade the biological immune system, T cells assist B cells in producing antibodies, preventing B cells from producing a large number of antibodies and causing autoimmune diseases. Figure 5 As shown in the figure, when an antigen invades the biological immune system, part of it acts on phagocytes, and the other part acts directly on B cells. The phagocytes engulf the antigen, and then break it down into antigen 1 and antigen 2 under the action of enzymes. The decomposed antigens are then displayed on the surface of macrophages and passed to T cells. T cells react with antigens, notifying the immune system of antigen invasion, and assist B cells in producing antibodies in the early stages of immunity. When the antigens are eliminated, T cells inhibit B cells from continuing to produce antibodies, thereby maintaining the balance of the entire immune system.

[0037] It can be seen from the T cell effect that the antigens decomposed by phagocytes, after reacting with T cells, help assist B cells to produce antibodies. In the process of environmental detection, the selection of the optimal behavior (antibody) of the robot fish is not only directly related to the obstacle information and regional detection status information within its detection range, but also related to the optimal avoidance strategy of the robot fish relative to obstacles or detected areas. In view of this, the present invention defines the robot fish sensor as a phagocyte, and defines the avoidance distance and avoidance angle when avoiding obstacles or detected areas during the detection process of the robot fish as two decomposition antigens, and then uses the T cell effect to improve the artificial immune collaborative detection network model and detection method. In order to avoid conceptual confusion during the design of the detection method, the present invention refers to the original environmental information detected by the robot fish, i.e., the obstacle information and the detected area as original antigens.

[0038] The robot fish mainly selects detection behavior based on environmental information, namely obstacle information and detected areas. In order to facilitate the robot fish to better handle the balance relationship between the two environmental information processing, and drawing on the T cell effect, the present invention constructs the following Figure 6 The interconnected coupled immune cooperative detection network shown: The surrounding environment of the underwater robot fish is used as antigen, obstacles and detected areas are used as original antigens, the underwater robot fish is used as B cells, and the detection behavior of the underwater robot fish is used as antibodies; According to the T cell effect, the robot fish sensor is used as a phagocyte, and the avoidance distance and avoidance angle of the robot fish when avoiding obstacles or detected areas during detection are used as two decomposition antigens; Therefore, under the stimulation of the original antigens composed of obstacles and detected areas, and the decomposed antigens composed of avoidance distance and avoidance angle, and at the same time under the stimulation and inhibition between the antibodies composed of the detection behavior of the underwater robot fish, the underwater robot fish avoids obstacles and detected areas through the optimal antibody selection, approaches the undetected areas, and realizes underwater environment detection.

[0039] Step 2: Initialize the parameters in the interconnected coupled immune cooperative detection model, including the initial position of the underwater robot fish, the initial antibody stimulation value and concentration value, the concentration adjustment coefficient δ, the stimulation adjustment coefficient β, the balance coefficient γ, etc.; Step 3: The robot fish obtains environmental information about surrounding obstacles and detected areas, and determines the original antigen E of the obstacle. o and the original antigen E d , and determine the decomposition antigens facing obstacles and explored areas: The barrier original antigen E o and the original antigen E d Together they constitute the original antigen E:

[0040] Among them, the barrier original antigen E o Determined according to the detection direction and detection distance of the underwater robot fish: ; Among them, E o A 6-bit binary code is used, and each two bits represent the obstacle information in the l-th detection direction of the underwater robot fish, l∈[1, 3]; ∀ l∈[1, 3], e o (l)∈{00, 01, 10, 11}, the set elements represent the no obstacle, long distance obstacle, medium distance obstacle and short distance obstacle in the grid within the lth detection direction and detection range respectively; Detected region original antigen E d Determined according to the detection direction and detection distance of the underwater robot fish:

[0041] In the formula, E d A 6-bit binary code is used, and each two bits represent the detected area information in the l-th detection direction of the underwater robot fish, l∈[1, 3]; ∀ l∈[1, 3], e d (l)∈{00, 01, 10, 11}, the set elements represent the lth detection direction and the detection range from far to near grids without detection, 1 detection, 2 detections and 3 detections respectively.

[0042] The decomposition antigen of the avoidance distance is:

[0043]

[0044] in, is the concentration adjustment factor of antibody i that decomposes antigens facing obstacles, δ o is the antibody concentration adjustment coefficient towards the obstacle, is the obstacle antigen stimulus value at time t, that is, the reciprocal of the distance between the underwater robot fish and the obstacle in the direction of antibody behavior, is the incentive value of antibody i at time t-1 in the obstacle-oriented immune detection network; is the concentration adjustment factor of antibody i for the decomposed antigen in the detected area, δ d is the antibody concentration adjustment factor for the detected area, is the antigen stimulation value of the detected area, and is the reciprocal of the distance between the underwater robot fish and the detected area in the direction of antibody behavior; is the incentive value of antibody i at time t-1 in the immune detection network facing the detected area.

[0045] It can be seen from the above formula that the closer the robot fish is to the obstacle, the smaller the antibody concentration adjustment factor is, thereby reducing the concentration of antibodies to avoid the probability of collision between the robot fish and the obstacle; conversely, it helps to increase the concentration of antibodies and improve the antibody execution efficiency when encountering long-distance obstacles.

[0046] Similarly, the closer the robot fish is to the detected area, the smaller the antibody concentration adjustment factor will be, thereby reducing the concentration of the antibody to avoid the probability of the robot fish repeatedly detecting the detected area; conversely, it will help to increase the concentration of the antibody and improve the antibody execution efficiency in the detected area.

[0047] The decomposed antigens at the avoidance angle are also divided into two categories: facing obstacles and detected areas, and the stimulating effect of the decomposed antigens on antibodies is mainly reflected by defining the potential field guidance factor. Figure 7 As shown in the figure, X r , X obs , X dt , X udt The coordinates of the robot fish, obstacle, detected area and undetected area respectively. The robot fish is subject to the virtual repulsion of the obstacle (detected area) ( ), and the virtual gravity F of the unexplored area closest to the robot fish att The virtual gravitational and repulsive forces act on the surface of the object, and under the combined force of virtual gravity and repulsion, the antibody behavior that matches the direction of the combined force is preferentially selected to achieve rapid detection. Specifically: According to the artificial potential field method: The virtual gravity of the robot fish is:

[0048] According to the artificial potential field method, facing the obstacles within the detection range, the underwater robot fish exerts a virtual repulsive force F on the obstacles within the detection range. o rep , obstacle avoidance angle λ o and the potential field guidance factor σ o They are: ; ;

[0049] Among them, F att is the virtual gravity of the robot fish, X r , X udt are the coordinates of the robot fish and the undetected area, ║·║ is the Euclidean distance, k att is the gravitational proportional factor, X obs is the obstacle coordinate, ρ o (X r , X obs )=║Xr -X obs ║,ρ r is the detection distance of the robot fish, k o is the obstacle repulsion proportional factor, η o is the angle between the direction of the robot fish's antibody behavior facing the obstacle and the y-axis, ν o is the obstacle potential field guidance adjustment coefficient; Similarly, the underwater robot fish faces the virtual repulsion F of the detected area within the detection range. d rep , obstacle avoidance angle λ d and the potential field guidance factor σ d They are:

[0050] ;

[0051]

[0052] Among them, F att is the virtual gravity of the robot fish, X dt is the coordinate of the detected area, ρ d (X r , X dt )=║X r – X dt ║,k d is the repulsion proportional factor of the detected area, η d is the angle between the behavior direction of a certain antibody facing the detected area and the y-axis, ν d is the potential field guidance adjustment coefficient of the detected area.

[0053] Step 4: Based on the information obtained in step 3, determine the antibody concentrations of the robotic fish, and select antibodies based on the antibody concentrations; First, the antibody alignment facing the obstacle and the antibody alignment facing the probed area are defined.

[0054] The antibody represents the behavior of the robot fish in processing environmental information, and its position is mainly composed of the prerequisites and behaviors of obstacle avoidance (detected area) during the detection process. According to the interconnected coupled immune cooperative detection network, the antibody of the present invention also includes two types: facing obstacles and facing detected areas.

[0055] In this embodiment, we define nine behaviors, corresponding to nine antibodies, and determine the antibody concentration of each antibody one by one in this step; The antibody alignment coding for the obstacle is determined according to the following table: Antibody ID <![CDATA[Preconditions for obstacle avoidance (P o )]]> Behavior 45° left straight ahead Right 45° 1 ## 00 ## go ahead 2 ## 01 ## go ahead 3 11 10 11 go ahead 4 00 10 ## Turn left 45° 5 ## 10 00 Turn right 45° 6 00 11 ## Turn left 45° 7 ## 11 11 Turn left 45° 8 ## 11 00 Turn right 45° 9 11 11 ## Turn right 45° Among them, each antibody number corresponds to an action of the robot fish, and each action corresponds to an obstacle avoidance prerequisite P o , that is, its corresponding bit encoding, the # in the encoding means that 0 or 1 can be selected; P o A 6-bit binary code is used, and each two bits represent the obstacle information in the l-th detection direction of the underwater robot fish, l∈[1, 3]; ∀ l∈[1, 3], e o (l)∈{00, 01,10, 11}, the set elements represent the no obstacle, long distance obstacle, medium distance obstacle and short distance obstacle in the grid within the lth detection direction and detection range respectively; The antibody position coding for the detected region is determined according to the following table: Antibody ID <![CDATA[Prerequisite for Avoiding Detected Areas (P d )]]> Behavior 45° left straight ahead Right 45° 1 ## 00 ## go ahead 2 ## 01 ## go ahead 3 1# 10 1# go ahead 4 00 10 ## Turn left 45° 5 ## 10 00 Turn right 45° 6 0# 11 ## Turn left 45° 7 00 11 1# Turn left 45° 8 ## 11 0# Turn right 45° 9 1# 11 00 Turn right 45° Among them, each antibody number corresponds to an action of the robot fish, and each action corresponds to a prerequisite P for avoiding the detected area. d , that is, its corresponding bit encoding, the # in the encoding means that 0 or 1 can be selected; P d A 6-bit binary code is used, and each two bits represent the detected area information in the l-th detection direction of the underwater robot fish, l∈[1, 3]; ∀ l∈[1, 3], e d (l)∈{00, 01, 10, 11}, the set elements represent the lth detection direction and the detection range from far to near grids without detection, 1 detection, 2 detections and 3 detections respectively.

[0056] According to the principle of the interconnected coupled immune cooperative detection network model, the antibodies of the present invention include two types: facing obstacles and detected areas. Therefore, when calculating the antibody concentration, it is necessary to first calculate the antibody concentration in the two cases separately, and then perform a weighted sum. When calculating the antibody concentration, an antibody concentration adjustment factor based on the avoidance distance decomposition of antigens and a potential field-oriented adjustment factor based on the avoidance angle decomposition of antigens are further introduced; The concentrations for antibody i are:

[0057]

[0058]

[0059]

[0060]

[0061] ; ; ; .

[0062] .

[0063] .

[0064] Among them, γ is the balance coefficient of the two immune networks, , are the incentive values ​​of antibody i facing the obstacle at time t and t-1 respectively; N is the number of antibodies; , and are the stimulation coefficient of antibody j facing the obstacle to antibody i, the inhibition coefficient of antibody i to antibody j, and the stimulation coefficient of antigen to antibody i; β o is the regulatory coefficient of antigen on antibody stimulation, is the natural mortality rate of antibody i facing the obstacle antigen, ⊕ is the XOR operation; and are the positional encodings of antibodies j and i in the immune cooperative detection network facing obstacles; E o (s) is the original antigen code of the obstacle; , are the stimulus values ​​of antibody i facing the detected area at time t and t-1 respectively; N is the number of antibodies; , and are the stimulation coefficient of antibody j to antibody i, the inhibition coefficient of antibody i to antibody j, and the stimulation coefficient of antigen to antibody i, respectively. d is the regulatory coefficient of antigen on antibody stimulation, is the natural mortality rate of antibodies against antigens in the detected region, and E is the positional encoding of antibody j and i in the immune cooperative detection network facing the detected area; d (s) is the original antigen code of the detected region; After determining the concentration of each antibody (antibody number 1 to antibody number 9), antibody selection is performed based on the roulette algorithm, completing the behavioral selection of the underwater robotic fish in this step.

[0065] Step 5: Based on the selected antibody behavior, the next step of detection of the robot fish is completed; Step 6: Determine whether the detection task is completed. If so, end; otherwise, go to step 3.

[0066] An underwater multi-autonomous robotic fish environment detection system includes the following modules: Interconnected coupling immune cooperative detection model construction module: used to build an interconnected coupling immune cooperative detection model for underwater multi-autonomous robotic fish and initialize parameters; Information acquisition module: used by the robot fish to obtain environmental information about surrounding obstacles and detected areas, and determine the original antigen E of the obstacles. o and the original antigen E d , and determine the decomposition antigens facing obstacles and explored areas; Antibody selection module: used to determine the concentration of each antibody of the robot fish based on the acquired information, and select antibodies based on each antibody concentration; Detection module: used to complete the detection of the robot fish based on the selected antibody behavior until the detection task is completed.

[0067] A computer device comprises a memory, a processor and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the following steps are implemented: Step 1: Construct an interconnected coupled immune collaborative detection model of underwater multi-autonomous robotic fish; Step 2, initializing the parameters in the interconnected coupled immune cooperative detection model; Step 3: The robot fish obtains environmental information about surrounding obstacles and detected areas, and determines the original antigen E of the obstacle. o and the original antigen E d , and determine the decomposition antigens facing obstacles and explored areas; Step 4: Based on the information obtained in step 3, determine the antibody concentrations of the robotic fish, and select antibodies based on the antibody concentrations; Step 5: Based on the selected antibody behavior, the next step of detection of the robot fish is completed; Step 6: Determine whether the detection task is completed. If so, end; otherwise, go to step 3.

[0068] A computer storable medium having a computer program stored thereon, wherein when the computer program is executed by a processor, the following steps are implemented: Step 1: Construct an interconnected coupled immune collaborative detection model of underwater multi-autonomous robotic fish; Step 2, initializing the parameters in the interconnected coupled immune cooperative detection model; Step 3: The robot fish obtains environmental information about surrounding obstacles and detected areas, and determines the original antigen E of the obstacle. o and the original antigen E d , and determine the decomposition antigens facing obstacles and explored areas; Step 4: Based on the information obtained in step 3, determine the antibody concentrations of the robotic fish, and select antibodies based on the antibody concentrations; Step 5: Based on the selected antibody behavior, the next step of detection of the robot fish is completed; Step 6: Determine whether the detection task is completed. If so, end; otherwise, go to step 3.

[0069] In order to verify the effectiveness and superiority of the underwater multi-autonomous robotic fish environment detection method (Immune Collaborative Environment Detection Method Based on T Cell Effect, ICEDMBTCE) of the present invention, for example Figures 8 to 11 The numerical detection tests of different numbers of robotic fish were carried out in the four environments shown in the figure. The test results were compared with the test results of the Complete Detection Method Based on Immune Mechanism [1] (Complete Detection Method Based on Immune Mechanism, CDMBIM) and the Immune Cooperative Detection Method Based on Biological Entropy [2] (Immune Cooperative Detection Method Based on Biological Entropy, ICDMBBE) based on the performance indicators of average detection steps, average repeated detection steps, coverage and detection repetition rate. The detection comparison results of the three methods for four different environments and different numbers of robotic fish are shown in Tables 3 and 4.

[0070] Table 3

[0071] Table 4:

[0072] It can be seen from the contents of Tables 3 and 4 that when two robot fish are used to detect four environments, the coverage rate of the ICEDMBTCE method of the present invention is increased by 2.9%, the repeated detection rate is reduced by 15.5%, the average number of detection steps is reduced by 6.1% on average, and the average number of repeated detection steps is reduced by 12.8% on average; when three robot fish are used to detect four environments, the coverage rate of the ICEDMBTCE method of the present invention is increased by 7.7%, the repeated detection rate is reduced by 33.4%, the average number of detection steps is reduced by 13.1% on average, and the average number of repeated detection steps is reduced by 27.7% on average; and when four robot fish are used to detect four environments, the coverage rate of the ICEDMBTCE method of the present invention is increased by 12.3%, the repeated detection rate is reduced by 49.7%, the average number of detection steps is reduced by 15.6% on average, and the average number of repeated detection steps is reduced by 33.8% on average, which not only verifies the effectiveness of the method of the present invention, but also shows that its detection performance is significantly better than the other two methods.

[0073] Further comparison Figures 8 to 11 It can be seen that Figures 8 to 11 The complexity of the four environments increases gradually in sequence, especially in environment four, there are not only numerous obstacles, but also two concave obstacles, which increases the detection difficulty of the robot fish. Based on this, further comparison of the data in Table 3 and Table 4 shows that with the increase in the complexity of the detection environment and the increase in the number of robot fish used for detection, the performance advantage of the ICEDMBTCE method of the present invention is further highlighted, which shows that it is better at handling complex unknown environment detection and can better play the collaboration between multiple robot fish. This is firstly due to the interconnected coupled immune cooperative detection network constructed by the present invention for obstacles and detected areas, which effectively balances the processing of environmental information of both during environmental detection, improves the coverage rate during environmental detection, reduces the detection repetition rate, and improves the detection quality; secondly, thanks to the design of the avoidance distance and avoidance angle decomposition antigen regulation factors for obstacles / detected areas, the obstacle avoidance efficiency of the robot fish is improved, thereby reducing the average number of detection steps and the average number of repeated detection steps, and improving the detection efficiency. In addition, the robot fish's division of detection distance into long, medium and short distances in the three forward directions and the antigen encoding of the detected environmental information not only improve the accuracy of environmental detection, but also realize the precise calculation of antibody concentration. While helping the robot fish to achieve accurate antibody selection, it also improves the detection quality and efficiency of multiple robot fish.

[0074] Figures 12 to 23 The results of detection by four robot fish using three immune detection methods for four environments are given. As can be seen from the figure, although the three methods all achieve detection in different environments, the detection coverage of the method of the present invention (Immune Collaborative Environment Detection Method Based on T Cell Effect, ICEDMBTCE) is significantly higher, while the immune cooperative detection method based on biological entropy (Immune Cooperative Detection Method Based on Biological Entropy, ICDMBBE) is second.

[0075] The above-mentioned embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the invention patent. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the patent of the present application shall be subject to the attached claims.

Claims

1. An underwater multi-autonomous robotic fish environment detection method, characterized in that: The following steps are involved: Step 1: Construct an interconnected coupled immune collaborative detection model of underwater multi-autonomous robotic fish; Step 2, initializing the parameters in the interconnected coupled immune cooperative detection model; Step 3: The robot fish obtains environmental information about surrounding obstacles and detected areas, and determines the original antigen E of the obstacle. o and the original antigen E d , and determine the decomposition antigens facing obstacles and explored areas; Step 4: Based on the information obtained in step 3, determine the antibody concentrations of the robotic fish, and select antibodies based on the antibody concentrations; Step 5: Based on the selected antibody behavior, the next step of detection of the robot fish is completed; Step 6: Determine whether the detection task is completed. If so, end; otherwise, go to step 3.

2. The method for underwater multi-autonomous robotic fish environment detection according to claim 1, characterized in that: The interconnected coupled immune cooperative detection model of underwater multiple autonomous robotic fish in step 1 is specifically as follows: The surrounding environment of the underwater robot fish is used as antigen, obstacles and detected areas are used as original antigens, the underwater robot fish is used as B cells, and the detection behavior of the underwater robot fish is used as antibodies; According to the T cell effect, the robot fish sensor is used as a phagocyte, and the avoidance distance and avoidance angle of the robot fish when avoiding obstacles or detected areas during detection are used as two decomposition antigens; Therefore, under the stimulation of the original antigens composed of obstacles and detected areas, and the decomposed antigens composed of avoidance distance and avoidance angle, and at the same time under the stimulation and inhibition between the antibodies composed of the detection behavior of the underwater robot fish, the underwater robot fish avoids obstacles and detected areas through the optimal antibody selection, approaches the undetected areas, and realizes underwater environment detection.

3. The method for underwater multi-autonomous robotic fish environment detection according to claim 2, characterized in that: The original antigen E in step 3 o and the original antigen E d Together they constitute the original antigen E: ； Among them, the barrier original antigen E o Determined according to the detection direction and detection distance of the underwater robot fish: ； Among them, E o A 6-bit binary code is used, and each two bits represent the obstacle information in the l-th detection direction of the underwater robot fish, l∈[1, 3]; ∀ l∈[1, 3], e o (l)∈{00, 01, 10, 11}, the set elements represent the no obstacle, long distance obstacle, medium distance obstacle and short distance obstacle in the grid within the lth detection direction and detection range respectively; Detected region original antigen E d Determined according to the detection direction and detection distance of the underwater robot fish: ； In the formula, E d A 6-bit binary code is used, and each two bits represent the detected area information in the l-th detection direction of the underwater robot fish, l∈[1, 3]; ∀ l∈[1, 3], e d (l)∈{00, 01, 10, 11}, the set elements represent the lth detection direction and the detection range from far to near grids without detection, 1 detection, 2 detections and 3 detections respectively.

4. The method for underwater multi-autonomous robotic fish environment detection according to claim 3, characterized in that: The decomposition antigen of the avoidance distance is: ， ； in, is the concentration adjustment factor of antibody i that decomposes antigens facing obstacles, δ o is the antibody concentration adjustment coefficient towards the obstacle, is the obstacle antigen stimulus value at time t, that is, the reciprocal of the distance between the underwater robot fish and the obstacle in the direction of antibody behavior, is the incentive value of antibody i at time t-1 in the obstacle-oriented immune detection network; is the concentration adjustment factor of antibody i for the decomposed antigen in the detected area, δ d is the antibody concentration adjustment factor for the detected area, is the antigen stimulation value of the detected area, and is the reciprocal of the distance between the underwater robot fish and the detected area in the direction of antibody behavior; is the incentive value of antibody i at time t-1 in the immune detection network facing the detected area.

5. The method for underwater multi-autonomous robotic fish environment detection according to claim 3, characterized in that: The decomposition antigens of the avoidance angle are: Virtual repulsion F of underwater robot fish facing obstacles within the detection range o rep , obstacle avoidance angle λ o and the potential field guidance factor σ o They are: ， ， ， ； Among them, F att is the virtual gravity of the robot fish, X r , X udt are the coordinates of the robot fish and the undetected area, ║·║ is the Euclidean distance, k att is the gravitational proportional factor, X obs is the obstacle coordinate, ρ o (X r , X obs )=║X r -X obs ║,ρ r is the detection distance of the robot fish, k o is the obstacle repulsion proportional factor, η o is the angle between the direction of the robot fish's antibody behavior facing the obstacle and the y-axis, ν o is the obstacle potential field guidance adjustment coefficient; The virtual repulsive force F of the underwater robot fish facing the detected area within the detection range d rep , obstacle avoidance angle λ d and the potential field guidance factor σ d They are: ， ， ， ； Among them, F att is the virtual gravity of the robot fish, X dt is the coordinate of the detected area, ρ d (X r , X dt )=║X r – X dt ║,k d is the repulsion proportional factor of the detected area, η d is the angle between the behavior direction of a certain antibody facing the detected area and the y-axis, ν d is the potential field guidance adjustment coefficient of the detected area.

6. The method for underwater multi-autonomous robotic fish environment detection according to claim 2, characterized in that: The antibody concentration in step 4 is: ， ， ， ， ， ， ， ， ， ， ； Among them, γ is the balance coefficient, , are the incentive values ​​of antibody i facing the obstacle at time t and t-1 respectively; N is the number of antibodies; , and are the stimulation coefficient of antibody j facing the obstacle to antibody i, the inhibition coefficient of antibody i to antibody j, and the stimulation coefficient of antigen to antibody i; β o is the regulatory coefficient of antigen on antibody stimulation, is the natural mortality rate of antibody i facing the obstacle antigen, ⊕ is the XOR operation; and are the positional encodings of antibodies j and i in the immune cooperative detection network facing obstacles; E o (s) is the original antigen code of the obstacle; , are the stimulus values ​​of antibody i facing the detected area at time t and t-1 respectively; N is the number of antibodies; , and are the stimulation coefficient of antibody j to antibody i, the inhibition coefficient of antibody i to antibody j, and the stimulation coefficient of antigen to antibody i, respectively. d is the regulatory coefficient of antigen on antibody stimulation, is the natural mortality rate of antibodies against antigens in the detected region, and E is the positional encoding of antibody j and i in the immune cooperative detection network facing the detected area; d (s) is the original antigen code of the detected region; After determining the concentration of each antibody, antibody selection is performed based on the roulette algorithm, which is the behavioral selection of the underwater robotic fish.

7. The method for underwater multi-autonomous robotic fish environment detection according to claim 6, characterized in that: The antibody alignment coding for the obstacle is determined according to the following table: Among them, each antibody number corresponds to an action of the robot fish, and each action corresponds to an obstacle avoidance prerequisite P o , that is, its corresponding bit encoding, the # in the encoding means that 0 or 1 can be selected; P o A 6-bit binary code is used, and each two bits represent the obstacle information in the l-th detection direction of the underwater robot fish, l∈[1, 3]; ∀ l∈[1, 3], e o (l)∈{00, 01, 10,11}, the set elements represent the no obstacle, long distance obstacle, medium distance obstacle and short distance obstacle in the grid within the lth detection direction and detection range respectively; The antibody position coding for the detected region is determined according to the following table: Among them, each antibody number corresponds to an action of the robot fish, and each action corresponds to a prerequisite P for avoiding the detected area. d , that is, its corresponding bit encoding, the # in the encoding means that 0 or 1 can be selected; P d A 6-bit binary code is used, and each two bits represent the detected area information in the l-th detection direction of the underwater robot fish, l∈[1, 3]; ∀ l∈[1, 3], e d (l)∈{00, 01, 10, 11}, the set elements represent the lth detection direction and the detection range from far to near grids without detection, 1 detection, 2 detections and 3 detections respectively.

8. An underwater multi-autonomous robotic fish environment detection system, characterized in that: Includes the following modules: Interconnected coupling immune cooperative detection model construction module: used to build an interconnected coupling immune cooperative detection model for underwater multi-autonomous robotic fish and initialize parameters; Information acquisition module: used by the robot fish to obtain environmental information about surrounding obstacles and detected areas, and determine the original antigen E of the obstacles o and the original antigen E d , and determine the decomposition antigens facing obstacles and explored areas; Antibody selection module: used to determine the concentration of each antibody of the robot fish based on the acquired information, and select antibodies based on each antibody concentration; Detection module: used to complete the detection of the robot fish based on the selected antibody behavior until the detection task is completed.

9. A computer device comprising a memory, a processor and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 7 are implemented.

10. A computer storable medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.

Images