Underwater robot networking environment sensing method based on immune mechanism

Abstract

The invention discloses an underwater robot networking environment sensing method based on an immune mechanism, and the method comprises the steps: distributing a recognition main body according to the environment sensing task demands of an underwater robot, constructing an environment database, adding a dynamic updating mechanism, and updating the environment database, a sensor is arranged for an underwater robot to collect original data, the original data serve as antigens to be matched with recognition subjects, the recognition subjects are screened and copied according to the matching result, the complexity degree of the current underwater environment is obtained to determine the number of the copied recognition subjects, and the copied recognition subjects are evaluated and optimized. And executing an environment perception task according to the optimized copy identification main body and an initial identification main body of the copy identification main body, sharing information through networking, setting a task expected result, calculating a difference degree between an actual result and the task expected result, and performing adjustment according to difference degree output. Therefore, the dynamic adaptability and the cooperation efficiency of the underwater robot networking in the underwater environment sensing process are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of underwater robot collaborative control, and in particular to an underwater robot networking environment perception method based on an immune mechanism. Background Art

[0002] In the field of underwater robot collaborative control technology, underwater environmental perception is the core link to achieve efficient and precise underwater robot operations. However, the underwater environment is extremely complex, which brings many severe challenges to environmental perception.

[0003] The underwater terrain is complex and diverse, with obstacles such as reefs, gullies, and caves widely distributed. Traditional environmental perception technologies are often difficult to adapt quickly and cannot be adjusted in real time to cope with dynamic changes in the environment. This increases the risk of collision for underwater robots and affects their operational safety and efficiency.

[0004] When multiple underwater robots work together, low collaborative efficiency is a pressing issue that needs to be addressed. Existing collaborative control strategies are inadequate and fail to fully leverage the advantages of multiple robots. This can lead to irrational task allocation, duplication of tasks, and reduced overall operational efficiency.

[0005] The immune system is a highly effective defense mechanism developed by organisms over a long period of evolution, possessing powerful adaptive, recognition, and learning capabilities. Within the biological immune system, immune cells rapidly identify foreign pathogens (antigens) and, through replication and mutation, produce large numbers of targeted antibodies to combat them. The immune system also memorizes previously encountered antigens, enabling a rapid response the next time the same or similar antigens are encountered. By leveraging these advantages of the immune system and applying them to environmental perception in networked underwater robots, underwater robots can rapidly identify and adapt to complex and changing underwater environments.

[0006] The solution proposed by the present invention is to allocate identification subjects and build an environmental database according to the environmental perception task requirements of the underwater robot, add a dynamic update mechanism to update the environmental database, arrange sensors for the underwater robot, match the collected original data with the identification subjects as antigens, and copy and screen the identification subjects according to the matching results to obtain the current underwater environmental status indicators, perform weighted calculation on the underwater environmental status indicators to obtain the current underwater environmental complexity, and calculate the number of identification subject copies according to the current underwater environmental complexity, evaluate and optimize the copied identification subjects, and then perform environmental perception tasks based on the optimized copied identification subjects and the initial identification subjects of the copied identification subjects, share information through networking, set expected task results, calculate the difference between the actual results and the expected task results, and adjust according to the difference output, thereby improving the dynamic adaptability and collaborative efficiency of the underwater robot networking when performing underwater environmental perception. Summary of the Invention

[0007] In order to solve the above-mentioned technical problems, the present invention provides an underwater robot networking environment perception method based on immune mechanism.

[0008] The technical solution of the present invention is implemented as follows: a method for underwater robot network environment perception based on immune mechanism, comprising:

[0009] S1. Assign recognition subjects according to the environmental perception task requirements of the underwater robot; build an environmental database, and add a dynamic update mechanism to update the environmental database;

[0010] S2. Arrange sensors on the underwater robot and match the collected raw data with the identification subject;

[0011] S3. Based on the matching result in step S2, duplicate and screen the identified subject, obtain a current underwater environment condition index, perform weighted calculation on the underwater environment condition index to obtain the current underwater environment complexity, and calculate the number of duplicates of the identified subject based on the current underwater environment complexity;

[0012] S4, evaluating and optimizing the identification subject copied in step S3;

[0013] S5. The underwater robot performs the environmental perception task based on the optimized replica recognition subject in step S4 and the initial recognition subject of the replica recognition subject, and shares information through networking, sets the expected result of the task, calculates the difference between the actual result and the expected result of the task, and makes adjustments based on the difference output.

[0014] Furthermore, in step S1, according to the environmental perception task requirements of the underwater robot, an identification subject is allocated, and an environmental database is constructed, and a dynamic update mechanism is added to update the environmental database.

[0015] Furthermore, in step S1, the specific steps are:

[0016] According to the environmental perception task requirements of the underwater robot network, the task is decomposed into multiple subtask modules, and the subtask modules include an obstacle avoidance task module, a communication optimization task module, and a path planning task module. The underwater robot is assigned an identification subject, and each subtask module corresponds to one or more identification subjects. The identification subject is a feature code in binary form. An environmental database is constructed based on the historical environmental data read from the underwater robot storage device and the environmental information shared by the underwater robot in the real-time task. The identification subject is stored in the environmental database, and an antigen-antibody mapping table associated with the binary code is established. For each antigen, the antibody strategy corresponding to the antigen is determined and recorded in the mapping table in a one-to-one correspondence. During the execution of the environmental perception task, the database is updated in real time with newly collected antigens and verified antibody strategies;

[0017] Furthermore, the antigen and antibody strategy is:

[0018] The antigen is abnormal data in the underwater environment, and the antibody strategy is the processing measures taken by the recognition subject for the abnormal data in the underwater environment.

[0019] Furthermore, in step S2, sensors are arranged on the underwater robot, and the collected raw data is matched with the identification subject.

[0020] Furthermore, in step S2, the specific steps are:

[0021] Sensors are arranged on different parts of the underwater robot, and the sensors include sonar sensors, lidar sensors, signal strength sensors, interference detection sensors, global positioning system receivers and inertial measurement units. The abnormal data collected by each sensor is used as raw data, and the raw data is preprocessed. The preprocessing is to use wavelet transform to denoise the raw data, and normalize the denoised raw data. Feature extraction is performed on the preprocessed data, and the feature extraction is divided into a certain number of intervals. Each interval is represented by a binary number. The subtask module intervals are combined according to the subtask modules in step S1 to obtain the binary antigen code corresponding to each subtask module. The binary antigen code is consistent with the length of the recognition subject feature code. The Hamming distance algorithm is used to calculate the distance between the binary antigen code and the identification subject feature code corresponding to each subtask module. The Hamming distance is the number of different positions of the binary bits corresponding to the binary antigen code and the identification subject feature code, and a matching threshold is preset. When the Hamming distance is less than and equal to the preset threshold, it indicates that the antigen and the identification subject are successfully matched. If there are multiple identification subjects that successfully match the antigen, the identification subjects that have successfully matched are sorted according to the historical success rate, and the identification subjects with a high historical success rate are preferentially selected for matching. When the Hamming distance is greater than the threshold, it indicates that the antigen and the identification subject have failed to match. The antigen that currently fails to match is recorded and uploaded to the environment database, and the identification subject is reassigned to the subtask module corresponding to the antigen that failed to match.

[0022] Furthermore, the preset matching threshold is:

[0023] For the obstacle avoidance task module, the matching threshold is preset to 3;

[0024] For the communication optimization task module, the matching threshold is preset to 2;

[0025] For the path planning task module, the matching threshold is preset to 4.

[0026] Furthermore, in step S3, based on the matching results in step S2, the identification subject is copied and screened, the current underwater environmental condition index is obtained, the current underwater environmental condition index is weightedly calculated to obtain the current underwater environmental complexity, and the number of identification subject copies is calculated based on the current underwater environmental complexity.

[0027] Furthermore, in step S3, the specific steps are:

[0028] Based on the matching result in step S2, the affinity between the recognition subject and the antigen is calculated, and the affinity threshold is set to 0.6. The recognition subjects with an affinity higher than the affinity threshold are screened and replicated. The current underwater environmental condition index is obtained through the sensor arranged in step S2. The underwater environmental condition index includes the current obstacle density, the degree of communication interference, and the complexity of the terrain. The underwater environmental condition index is weighted and calculated to obtain the current underwater environmental complexity. The number of recognition subject replications is calculated based on the current underwater environmental complexity.

[0029] Furthermore, the affinity is calculated as:

[0030]

[0031] Among them, A f is the affinity between the recognition subject and the antigen, H is the Hamming distance between the recognition subject feature code and the antigen code, and n is the number of coding bits;

[0032] Furthermore, the underwater environmental condition index is weighted and calculated as follows:

[0033] C=ω1×Z+ω2×T+ω3×D

[0034] Where C is the complexity of the current underwater environment, Z is the obstacle density, T is the communication interference level, D is the terrain complexity, ω1, ω2, and ω3 are the weights corresponding to the obstacle density, communication interference level, and terrain complexity, respectively;

[0035] Furthermore, the number of copies of the identification subject is calculated as:

[0036] N=α×C+β×A f +n k

[0037] Among them, N is the number of copies of the identification subject, C is the current underwater environmental condition indicator, and A f To identify the affinity between the subject and the antigen, n k is the basic replication number, α is the complexity coefficient of the current underwater environment, and β is the affinity influence coefficient.

[0038] Furthermore, in step S4, the identification subject copied in step S3 is evaluated and optimized.

[0039] Furthermore, in step S4, the specific steps are:

[0040] For the replica recognition subject generated in step S3, the replica recognition subject is matched multiple times with the antigen corresponding to the initial recognition subject of the replica recognition subject, and the success rate of successful matching is calculated. If the matching success rate is greater than and equal to 90% or above, it means that the replica recognition subject can be used. If the matching success rate is less than 90%, the matching is continued by lowering the matching threshold described in step S2. If the requirement is not met again, the successfully matched replica recognition subject is screened out, and the antibody strategy of the successfully matched replica recognition subject is verified multiple times in the current underwater environment to test the environmental adaptability. When the antibody strategy verification success rate is greater than and equal to 80% or above, it means that the environmental adaptability of the replica recognition subject is good. If the antibody strategy verification success rate is less than 80%, the replica recognition subject with a success rate less than 80% is discarded, and the current underwater environment complexity coefficient and the affinity influence coefficient are adjusted according to the current environmental complexity to generate a discarded number of replica recognition subjects again, and the generated replica recognition subjects are matched and screened again until the environmental adaptability of the replica recognition subject reaches the expected success rate.

[0041] Furthermore, in step S5, the underwater robot performs the environmental perception task based on the optimized replica recognition subject in step S4 and the initial recognition subject of the replica recognition subject, and shares information through networking, sets the expected result of the task, calculates the difference between the actual result and the expected result of the task, and makes adjustments based on the difference output.

[0042] Furthermore, in step S5, the specific steps are:

[0043] The underwater robot applies the optimized replica recognition subject in step S4 to the corresponding subtask module, and performs environmental perception task processing in combination with the initial recognition subject of the replica recognition subject. During the task processing, the underwater robot continuously monitors environmental changes through the sensors arranged in step S2. If a significant change in the underwater environmental conditions is detected, the antibody strategy of the replica recognition subject and the initial recognition subject of the replica recognition subject is adjusted, and the situation is fed back to other subtask modules for coordinated adjustment. In the task execution process, the underwater robot uses LoRa technology to build a wireless communication network, and introduces the MQTT protocol for data transmission and message distribution during task processing. Each underwater robot shares information according to the LoRa-MQTT protocol to collaboratively process the environmental perception task, and sets the expected result of the task. The relative error method is used to calculate the difference between the expected result of the environmental perception task and the actual environmental perception task result, and the antibody strategy of the replica recognition subject currently executing the environmental perception task and the initial recognition subject of the replica recognition subject are adjusted according to the difference output. The antibody strategy is adjusted as follows:

[0044] When the difference is less than or equal to 10%, the antibody strategies of the replica recognition subject performing the environmental perception task and the initial recognition subject of the replica recognition subject are marked as stable strategies and stored in the environmental database;

[0045] When the difference is greater than 10%, the process from step S1 to step S4 is executed again.

[0046] Beneficial effects:

[0047] The present invention allocates recognition subjects according to the environmental perception task requirements of the underwater robot, constructs an environmental database, adds a dynamic update mechanism to update the environmental database, arranges sensors for the underwater robot, matches the collected original data with the recognition subjects as antigens, and copies and screens the recognition subjects according to the matching results to obtain current underwater environmental status indicators. The underwater environmental status indicators are weightedly calculated to obtain the current underwater environmental complexity, and the number of copies of the recognition subjects is calculated according to the current underwater environmental complexity. The copied recognition subjects are evaluated and optimized, and then the environmental perception task is performed according to the optimized copied recognition subjects and the initial recognition subjects of the copied recognition subjects. Information is shared through networking, the expected result of the task is set, the difference between the actual result and the expected result of the task is calculated, and adjustment is made according to the difference output, thereby improving the dynamic adaptability and collaborative efficiency of the underwater robot networking when performing underwater environmental perception. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] Figure 1 This is a structural block diagram of an underwater robot networking environment perception method based on an immune mechanism in an embodiment of the present invention;

[0049] Figure 2 This is a flowchart of the steps of a method for underwater robot networking environment perception based on an immune mechanism in an embodiment of the present invention;

[0050] Figure 3 This is a flow chart of the identification subject replication steps of an underwater robot networking environment perception method based on the immune mechanism in an embodiment of the present invention. DETAILED DESCRIPTION

[0051] In order to make the objects and advantages of the present invention more clearly understood, the present invention is further described below in conjunction with embodiments; it should be understood that the specific embodiments described herein are merely used to explain the present invention and are not intended to limit the present invention.

[0052] The preferred implementation methods of the present invention are described below with reference to the accompanying drawings. Those skilled in the art should understand that these implementation methods are only used to explain the technical principles of the present invention and are not intended to limit the scope of protection of the present invention.

[0053] See also Figure 1-Figure 3 As shown, the embodiment of the present invention provides an underwater robot network environment perception method based on an immune mechanism, including:

[0054] S1. Assign recognition subjects according to the environmental perception task requirements of the underwater robot; build an environmental database, and add a dynamic update mechanism to update the environmental database;

[0055] S2. Arrange sensors on the underwater robot and match the collected raw data with the identification subject;

[0056] S3. Based on the matching result in step S2, duplicate and screen the identified subject, obtain a current underwater environment condition index, perform weighted calculation on the underwater environment condition index to obtain the current underwater environment complexity, and calculate the number of duplicates of the identified subject based on the current underwater environment complexity;

[0057] S4, evaluating and optimizing the identification subject copied in step S3;

[0058] S5. The underwater robot performs the environmental perception task based on the optimized replica recognition subject in step S4 and the initial recognition subject of the replica recognition subject, and shares information through networking, sets the expected result of the task, calculates the difference between the actual result and the expected result of the task, and makes adjustments based on the difference output.

[0059] like Figure 2 As shown, in step S1, according to the environmental perception task requirements of the underwater robot, an identification subject is allocated; and an environmental database is constructed, and a dynamic update mechanism is added to update the environmental database.

[0060] Specifically, in this embodiment, according to the environmental perception task requirements of the underwater robot network, the environmental perception task is decomposed into three subtask modules: obstacle avoidance task module, communication optimization task module, and path planning task module. Such task decomposition enables the underwater robot to quickly and accurately call corresponding processing measures when responding to different environmental perception task requirements, thereby improving the pertinence and efficiency of task execution. The obstacle avoidance task module is assigned an identification subject with the ability to identify and avoid obstacles such as reefs and shipwrecks, with a feature code of 101011; the communication optimization task module is assigned an identification subject that can reduce the impact of electromagnetic interference and ensure stable communication, with a feature code of 011100; the path planning task module is assigned an identification subject that can plan an efficient path that avoids obstacles and adapts to complex terrain, with a feature code of 110010;

[0061] Historical environmental data is read from the underwater robot storage device to build an environmental database. This data records the location, size, shape of obstacles encountered in the past detection in the environmental perception task area, the frequency, intensity and location of communication interference, the undulation of the terrain, the distribution of canyons and seamounts, and other information. An antigen-antibody mapping table associated with binary coding is established. For each antigen, the antibody strategy adopted by the recognition subject corresponding to the antigen is determined and recorded in the mapping table with a one-to-one correspondence. In actual environmental perception, once an antigen (abnormal data in the underwater environment) is detected, the underwater robot can quickly find the corresponding antibody strategy (the processing measures of the recognition subject) in the mapping table according to the binary code, thereby achieving a rapid response to environmental changes. During the execution of the task, if new antigen data is found, that is, when abnormal data is detected in the underwater environment during the current task execution, the abnormal data is recorded in the database, and the effective antibody strategy for the abnormal data (that is, the operation or strategy for solving the abnormal data) is recorded and uploaded to the environmental database.

[0062] like Figure 2 As shown, in step S2, sensors are arranged on the underwater robot, and the collected raw data is matched with the identification subject.

[0063] Specifically, in this embodiment, a variety of sensors are installed at different parts of the underwater robot. A high-precision sonar sensor and a lidar sensor are installed at the front end of the underwater robot to accurately detect the distance, shape, and size of obstacles ahead. A signal strength sensor and an interference detection sensor are installed on the side of the fuselage to monitor the communication signal strength and interference in real time. A global positioning system receiver and an inertial measurement unit are installed on the top to obtain the precise position and posture information of the robot. Abnormal data collected by each sensor is used as raw data, and the raw data is preprocessed. The preprocessing includes denoising the raw data using a wavelet transform and normalizing the denoised raw data. The normalization includes using a minimum-maximum normalization method to convert data with different dimensions and value ranges collected by different sensors into a unified and comparable numerical range.

[0064] The pre-processed data is subjected to feature extraction, and the feature extraction is divided into a certain number of intervals, each of which is represented by a binary number. Taking the obstacle avoidance task module as an example, the key features of the distance, shape, and moving speed of the object are extracted from the pre-processed data, and the distance is divided into 0-5 meters (binary representation is 00), 5-10 meters (01), 10-20 meters (10), and more than 20 meters (11). The shape is divided into regular (00), approximately regular (01), and irregular (10). The moving speed is divided into stationary (00), low speed (01), and high speed (10). The preset distance of the detected object is 11 meters, the shape is regular, and the speed is high. The binary antigen code of the obstacle avoidance task module is combined to be 010100. The Hamming distance algorithm is used to calculate the binary antigen code (100010) of the obstacle avoidance task module and the obstacle avoidance task module recognition subject feature code (1 01011), compare the number of different binary bit positions corresponding to the two encodings, and obtain a Hamming distance of 2, which is less than the preset obstacle avoidance task module matching threshold, indicating that the current antigen is successfully matched with the obstacle avoidance task module recognition subject. If there are multiple successfully matched recognition subjects, they are sorted according to the historical success rate, and the recognition subject with the high success rate is preferentially selected for matching. The historical success rate is obtained by counting the number of times the recognition subject successfully performs the environmental perception task in multiple similar environments in the past and the total number of executions. If there is currently recognition subject A that has successfully avoided obstacles 8 times in the past 10 similar obstacle avoidance scenarios, then the historical success rate of the recognition subject A is 80%, and there is currently recognition subject B that has successfully avoided obstacles 7 times in the past 10 similar obstacle avoidance scenarios, then the historical success rate of the recognition subject B is 70%. Recognition subject A is selected for matching according to the historical success rate sorting.

[0065] like Figure 2-Figure 3 As shown, in step S3, according to the matching result in step S2, the identification subject is copied and screened, the current underwater environmental condition index is obtained, the underwater environmental condition index is weightedly calculated to obtain the current underwater environmental complexity, and the number of identification subject copies is calculated based on the current underwater environmental complexity.

[0066] Specifically, in this embodiment, based on the matching result in step S2, the affinity between the obstacle avoidance task module and the antigen is calculated as:

[0067]

[0068] It shows that the obstacle avoidance task module recognition subject can be copied. The current underwater environment condition indicators are obtained through the sensors arranged in step S2. The current underwater environment condition indicators are preset as obstacle density Z = 0.4, communication interference level T = 0.7, and terrain complexity D = 0.5. The corresponding weights are ω1 = 0.4, ω2 = 0.3, and ω3 = 0.3, respectively. The current underwater environment complexity is:

[0069] C=0.4×0.4+0.3×0.7+0.3×0.5=0.52

[0070] Assuming the basic replication number is 2, the underwater environment complexity coefficient is 2, and the affinity influence coefficient is 1.5, the number of identification subject replications is calculated based on the current underwater environment complexity, which is calculated as:

[0071] N=2×0.52+1.5×0.67+2=4.045

[0072] Round up to an integer, that is, the number of main body copies recognized by the current obstacle avoidance task module is 5.

[0073] like Figure 2 As shown, in step S4, the recognition subject copied in step S3 is evaluated and optimized.

[0074] Specifically, in this embodiment, for the replica recognition subject generated in step S3, the replica recognition subject is matched multiple times with the antigen corresponding to the initial recognition subject of the replica recognition subject. 100 matches are preset, and 92 matches are successful, indicating that the matching success rate is greater than 90%, indicating that the replica recognition subject can be used. If the matching success rate is less than 90%, the matching is continued by lowering the matching threshold in step S2 until the matching success rate meets the requirement. The lower matching threshold allows a large difference between the replica recognition subject and the antigen to still be considered a match. If the requirement is not met again, the successfully matched replica recognition subject is screened out for verification;

[0075] For the initially available replication recognition subjects, the antibody strategy generated for the identified antigens in the current underwater environment is verified multiple times. The verification is preset to 10 times. If 8 times are successful and the verification success rate is greater than 80%, it is considered that the replication recognition subject has good environmental adaptability. If the verification success rate is less than 80%, the replication recognition subject with a success rate less than 80% is discarded, and the current underwater environment complexity coefficient and affinity influence coefficient are adjusted according to the complexity of the current underwater environment to generate a discarded number of replication immune monomers. When the current underwater environment complexity is high, the current underwater environment complexity coefficient and affinity influence coefficient are increased. When the current underwater environment complexity is low, the current underwater environment complexity coefficient and affinity influence coefficient are reduced. The generated replication recognition subjects are then re-matched and screened until the environmental adaptability of the replication recognition subject reaches the expected success rate, ensuring that all generated replication immune monomers can adapt to the current underwater environment and perform environmental perception tasks.

[0076] like Figure 2 As shown, in step S5, the underwater robot performs the environmental perception task based on the optimized replica recognition subject in step S4 and the initial recognition subject of the replica recognition subject, and shares information through networking, sets the expected result of the task, calculates the difference between the actual result and the expected result of the task, and makes adjustments based on the difference output.

[0077] Specifically, in this embodiment, the optimized copy recognition subject is applied to the corresponding subtask module, and the environmental perception task is performed in combination with the initial recognition subject, giving full play to the advantages of the recognition subject and improving the accuracy and efficiency of environmental perception. During the execution of the environmental perception task, if the sensor detects that the current underwater environmental conditions have changed significantly, the antibody strategies of the copy recognition subject and the initial recognition subject are adjusted, and the antibody strategy corresponding to the current underwater environmental conditions is found according to the environmental database, and the situation is fed back to other subtask modules for joint adjustment. The underwater robot uses LoRa technology to build a wireless communication network and introduces the MQTT protocol for data transmission and message distribution when processing the environmental perception task. Each underwater robot shares its own detection data, the environmental conditions encountered, and the treatment measures taken through the wireless communication network, and cooperates to detect the environmental perception task and realize collaborative operation. Multi-robot collaboration can share perception tasks, avoid duplication of work, and improve overall perception efficiency. In the environmental perception task, different robots can reasonably plan monitoring routes based on shared information to cover a larger area.

[0078] The expected result of the environmental perception task is set, and the relative error method is used to calculate the difference between the actual result and the expected result. Adjustments are made based on the difference. This is to continuously optimize the underwater robot's environmental perception task process. The relative error method is:

[0079]

[0080] Where C is the difference, A is the expected result of the environmental perception task, and B is the actual result;

[0081] When the difference is less than or equal to 10%, the effective antibody strategy is stored in the environmental database to provide a reference for subsequent tasks. When the difference is greater than 10%, the S1 to S4 processes are re-executed for comprehensive adjustments. This feedback adjustment mechanism enables the underwater robot to continuously improve according to the actual perception situation, thereby enhancing its adaptability and perception capabilities in complex underwater environments.

[0082] Thus far, the technical solutions of the present invention have been described in conjunction with the preferred embodiments shown in the accompanying drawings. However, it will be readily understood by those skilled in the art that the scope of protection of the present invention is obviously not limited to these specific embodiments. Without departing from the principles of the present invention, those skilled in the art may make equivalent changes or substitutions to the relevant technical features, and the technical solutions after such changes or substitutions will fall within the scope of protection of the present invention.

[0083] The foregoing description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Those skilled in the art will readily appreciate that the present invention is susceptible to various modifications and variations. Any modifications, equivalent substitutions, or improvements made within the spirit and principles of the present invention shall be included within the scope of protection of the present invention.

Claims

1. A method for underwater robot network environment perception based on immune mechanism, characterized in that: The following steps are involved: S1. Assign recognition subjects according to the environmental perception task requirements of the underwater robot; build an environmental database, and add a dynamic update mechanism to update the environmental database; S2. Arrange sensors on the underwater robot and match the collected raw data with the identification subject; S3. Based on the matching result in step S2, duplicate and screen the identified subject, obtain a current underwater environment condition index, perform weighted calculation on the underwater environment condition index to obtain the current underwater environment complexity, and calculate the number of duplicates of the identified subject based on the current underwater environment complexity; S4, evaluating and optimizing the identification subject copied in step S3; S5. The underwater robot performs the environmental perception task based on the optimized replica recognition subject in step S4 and the initial recognition subject of the replica recognition subject, and shares information through networking, sets the expected result of the task, calculates the difference between the actual result and the expected result of the task, and makes adjustments based on the difference output.

2. The method for underwater robot network environment perception based on immune mechanism according to claim 1 is characterized by: According to the requirements of the underwater robot's environmental perception task, the identification subject is allocated; and an environmental database is constructed, and a dynamic update mechanism is added to update the environmental database: According to the environmental perception task requirements of the underwater robot network, the task is decomposed into multiple subtask modules, which include an obstacle avoidance task module, a communication optimization task module and a path planning task module. Identification subjects are assigned to the underwater robot. Each subtask module corresponds to one or more identification subjects. The identification subject is a feature code in binary form. An environmental database is constructed based on the historical environmental data read from the underwater robot storage device and the environmental information shared by the underwater robot in the real-time task. The identification subject is stored in the environmental database, and an antigen-antibody mapping table associated with binary coding is established. For each antigen, the antibody strategy corresponding to the antigen is determined and recorded in the mapping table with a one-to-one correspondence. During the execution of the environmental perception task, the database updates the newly collected antigens and verified antibody strategies in real time.

3. The method for underwater robot network environment perception based on immune mechanism according to claim 2 is characterized by: The antigen is abnormal data in the underwater environment, and the antibody strategy is the processing measures taken by the recognition subject for the abnormal data in the underwater environment.

4. The method for underwater robot network environment perception based on immune mechanism according to claim 1, characterized in that: Arrange the sensors on the underwater robot and match the collected raw data with the identification subject as follows: Arrange sensors on different parts of the underwater robot respectively, take the abnormal data collected by each sensor as raw data, pre-process the raw data, perform feature extraction on the pre-processed data, divide the feature extraction into a certain number of intervals, each interval is represented by a binary number, and combine the subtask module intervals according to the subtask modules in step S1 to obtain the binary antigen code corresponding to each subtask module. The length of the binary antigen code is consistent with the identification subject feature code. Use the Hamming distance algorithm to calculate the distance between the binary antigen code and the identification subject feature code corresponding to each subtask module. The Hamming distance is the number of different positions of the binary bits corresponding to the binary antigen code and the recognition subject feature code, and a matching threshold is preset. When the Hamming distance is less than and equal to the preset threshold, it means that the antigen and the recognition subject are successfully matched. If there are multiple recognition subjects that successfully match the antigen, the successfully matched recognition subjects are sorted according to the historical success rate, and the recognition subjects with a high historical success rate are preferentially selected for matching. When the Hamming distance is greater than the threshold, it means that the antigen and the recognition subject have failed to match. The antigen that currently fails to match is recorded and uploaded to the environmental database, and the recognition subject is reassigned to the subtask module corresponding to the failed antigen match.

5. The method for underwater robot network environment perception based on immune mechanism according to claim 4, characterized in that: The preset matching threshold is: For the obstacle avoidance task module, the matching threshold is preset to 3; For the communication optimization task module, the matching threshold is preset to 2; For the path planning task module, the matching threshold is preset to 4.

6. The method for underwater robot network environment perception based on immune mechanism according to claim 1, characterized in that: According to the matching result in step S2, the identification subject is copied and screened as follows: Calculate the affinity between the recognition subject and the antigen, set an affinity threshold, screen out the recognition subject with an affinity higher than the affinity threshold, and replicate it, wherein the affinity is calculated as: Among them, A f is the affinity between the recognition subject and the antigen, H is the Hamming distance between the recognition subject feature code and the antigen code, and n is the number of coding bits.

7. The method for underwater robot network environment perception based on immune mechanism according to claim 1, characterized in that: The underwater environmental condition index is weighted and calculated as follows: C=ω1×Z+ω2×T+ω3×D Among them, C is the complexity of the current underwater environment, Z is the obstacle density, T is the communication interference level, D is the terrain complexity, and ω1, ω2, and ω3 are the weights corresponding to the obstacle density, communication interference level, and terrain complexity, respectively.

8. The method for underwater robot network environment perception based on immune mechanism according to claim 1, characterized in that: The number of copies of the identification subject is calculated as: N=α×C+β×A f +n k Among them, N is the number of copies of the identification subject, C is the current underwater environmental condition indicator, and A f To identify the affinity between the subject and the antigen, n k is the basic replication number, α is the complexity coefficient of the current underwater environment, and β is the affinity influence coefficient.

9. The method for underwater robot network environment perception based on immune mechanism according to claim 1, characterized in that: The identification subject copied in step S3 is evaluated and optimized as follows: For the replica recognition subject generated in step S3, the replica recognition subject is matched multiple times with the antigen corresponding to the initial recognition subject of the replica recognition subject, and the success rate of successful matching is calculated. If the matching success rate is greater than and equal to 90% or above, it means that the replica recognition subject can be used. If the matching success rate is less than 90%, the matching is continued by lowering the matching threshold described in step S2. If the requirement is not met again, the successfully matched replica recognition subject is screened out, and the antibody strategy of the successfully matched replica recognition subject is verified multiple times in the current underwater environment to test the environmental adaptability. When the antibody strategy verification success rate is greater than and equal to 80% or above, it means that the environmental adaptability of the replica recognition subject is good. If the antibody strategy verification success rate is less than 80%, the replica recognition subject with a success rate less than 80% is discarded, and the current underwater environment complexity coefficient and the affinity influence coefficient are adjusted according to the current environmental complexity to generate a discarded number of replica recognition subjects again, and the generated replica recognition subjects are matched and screened again until the environmental adaptability of the replica recognition subject reaches the expected success rate.

10. The method for underwater robot network environment perception based on immune mechanism according to claim 1, characterized in that: The underwater robot performs the environmental perception task based on the optimized replica recognition subject in step S4 and the initial recognition subject of the replica recognition subject, and sets the expected result of the task through networking and sharing information, calculates the difference between the actual result and the expected result of the task, and adjusts the output according to the difference as follows: The underwater robot applies the optimized replica recognition subject in step S4 to the corresponding subtask module, and performs environmental perception task processing in combination with the initial recognition subject of the replica recognition subject. During the task processing, the underwater robot continuously monitors environmental changes through the sensors arranged in step S2. If a significant change in the underwater environmental conditions is detected, the antibody strategy of the replica recognition subject and the initial recognition subject of the replica recognition subject is adjusted, and the situation is fed back to other subtask modules for coordinated adjustment. In the task execution process, the underwater robot uses LoRa technology to build a wireless communication network, and introduces the MQTT protocol for data transmission and message distribution during task processing. Each underwater robot shares information according to the LoRa-MQTT protocol to collaboratively process the environmental perception task, and sets the expected result of the task. The relative error method is used to calculate the difference between the expected result of the environmental perception task and the actual environmental perception task result, and the antibody strategy of the replica recognition subject currently executing the environmental perception task and the initial recognition subject of the replica recognition subject are adjusted according to the difference output. The antibody strategy is adjusted as follows: When the difference is less than or equal to 10%, the antibody strategies of the replica recognition subject performing the environmental perception task and the initial recognition subject of the replica recognition subject are marked as stable strategies and stored in the environmental database; When the difference is greater than 10%, the process from step S1 to step S4 is executed again.

Images