An integrated control method, device and medium for unmanned systems

Abstract

The application relates to an integrated control method, device and medium for an unmanned system, the method realizes integrated control by arranging an integrated control module connected with an external control terminal, an executor and at least one sensor, and the method specifically comprises the following steps: obtaining a task instruction from the external control terminal and obtaining sensor information from the sensor; obtaining a control signal based on the task instruction and the sensor information through multi-source information fusion, obstacle type identification, judgment on whether obstacle avoidance is needed and optimal strategy determination under clustering; sending the control signal to the executor and obtaining execution result information from the executor; and obtaining execution evaluation information based on the task instruction and the execution result information. The application realizes end-to-end task execution and action analysis by using only one integrated module, the module supports seamless connection with the unmanned system body, avoids a long hierarchical call chain between traditional multiple modules, and reduces performance loss and information loss.

Description

Technical Field

[0001] This invention relates to the field of unmanned aerial vehicle control technology, and in particular to an integrated control method, device, and medium for unmanned systems. Background Technology

[0002] In recent years, with the continuous development of sensor, communication, and network technologies, unmanned systems have been widely used in both military and civilian fields due to their advantages such as no risk of personnel casualties, low cost, and ease of operation. The market generally offers dedicated integrated unmanned systems for specific tasks. However, because the components of these integrated systems, including communication, computing, and control, are not interchangeable, the versatility of their application in different scenarios is greatly limited. Modular design effectively solves this problem, offering advantages such as ease of maintenance, strong adaptability, and significant cost savings, achieving a precise match between the functionality of unmanned systems and diverse task requirements. However, unmanned systems often encapsulate specific functions such as communication, computing, and control separately. This leads to long hierarchical call chains, and performance degradation and information loss increase exponentially with the number of modules.

[0003] Chinese patent application CN201810143977.7 discloses an indoor visual navigation unmanned aerial vehicle (UAV) swarm flight control system and method. The system includes at least two UAVs, each UAV comprising: an optical flow sensor module for acquiring the UAV's planar position via image processing; an ultrasonic ranging module for acquiring the UAV's altitude; a short-range wireless communication module for communication between the UAVs; a memory for storing flight control programs corresponding to various flight tasks; and a controller connected to the optical flow sensor module, ultrasonic ranging module, short-range wireless communication module, and memory, for controlling the UAV's flight according to control commands, wherein the control commands are sent by the UAVs in the system.

[0004] The above methods enable independent flight and autonomous control; however, the above applications do not solve the problem of performance loss and information loss caused by the separate packaging of each part.

[0005] To enable unmanned systems (UMS) to perform tasks in complex real-world environments and with diverse mission requirements, the limitations of a single UMS in terms of perception range and onboard equipment prevent it from meeting specific mission requirements. Therefore, greater attention is being paid to the collaborative applications of UMS swarms. While UMS swarms possess distributed parallel computing and perception capabilities, enabling collaborative assistance and information sharing, their large overall scale often makes full-process logical control difficult. Consequently, there is currently a lack of an integrated method, device, or medium that can seamlessly connect with UMS swarms to address this issue. Summary of the Invention

[0006] The purpose of this invention is to overcome the shortcomings of the existing technology and provide an integrated control method, device, and medium for unmanned systems. By integrating communication, computing, control, and analysis into an integrated control module, the possibility of performance loss and information loss is reduced.

[0007] The objective of this invention can be achieved through the following technical solutions:

[0008] One aspect of the present invention provides an integrated control method for unmanned systems, applied to an integrated control module, the method specifically comprising the following steps:

[0009] Obtain task instructions and sensor information from the control terminal;

[0010] Based on the sensor information, fusion features are obtained through multi-source information fusion processing. Obstacle type identification is performed based on the fusion features to determine whether obstacle avoidance is required. The optimal strategy under the cluster is determined based on the task instructions, and control signals are obtained.

[0011] The control signal is sent to the actuator, and the execution result information is obtained from the actuator;

[0012] Based on the task instructions and the execution result information, execution evaluation information is obtained and sent to the control terminal.

[0013] As a preferred technical solution, the multi-source information fusion processing process includes the following steps:

[0014] The sensor information is then normalized and standardized.

[0015] Data dimensionality reduction and data fusion are achieved through principal component analysis and neural networks;

[0016] For the features obtained after data fusion, a classifier is used to extract features and obtain fused features.

[0017] As a preferred technical solution, the obstacle type identification process includes the following steps:

[0018] The fused features obtained by fusing information from multiple sources are input into the trained obstacle recognition model, and the obstacle type recognition result is obtained based on the model's output.

[0019] As a preferred technical solution, the determination of whether obstacle avoidance is needed specifically includes the following steps:

[0020] The fused features obtained by fusing multi-source information are input into a preset dynamic model to determine whether there is a collision risk and whether obstacle avoidance is necessary.

[0021] As a preferred technical solution, the process of determining the optimal strategy under the cluster and obtaining control signals includes the following steps:

[0022] Based on the obstacle type identification results and the judgment results of whether obstacle avoidance is needed, the optimal policy is selected from the policy set including cooperation, competition and inaction according to the policy function obtained by pre-training, and the corresponding control signal is obtained.

[0023] As a preferred technical solution, the process of obtaining performance evaluation information includes the following steps:

[0024] Based on the task instructions and the execution result information, obtain the deviation evaluation index;

[0025] Determine whether the currently selected strategy is Pareto optimal for the cooperative task and obtain the determination result;

[0026] Based on the deviation evaluation index and the judgment result, execution evaluation information is obtained.

[0027] As a preferred technical solution, the sensor information includes infrared sensing information, ultrasonic information, acceleration information, altitude information, navigation information, and image information from an external gimbal.

[0028] As a preferred technical solution, the task instructions include individual task instructions and collaborative task instructions.

[0029] According to another aspect of the present invention, an integrated control device for unmanned systems is provided, comprising an integrated module and at least one sensor, a control terminal, and an actuator connected to the integrated module, the integrated module comprising:

[0030] The communication unit is used to establish a connection with an external control terminal, obtain task instruction information from the external control terminal, and feed back execution evaluation information to the external control terminal. The communication unit is also used to obtain sensor information of the sensor.

[0031] The computing unit is used to acquire control signals by fusing multi-source information, identifying obstacle types, determining whether obstacle avoidance is needed, and determining the optimal strategy under cluster conditions;

[0032] The control unit is used to send the control signal to the actuator and obtain execution result information from the actuator;

[0033] The analysis unit is used to obtain execution evaluation information based on the task instructions and the execution result information.

[0034] According to another aspect of the invention, a computer-readable storage medium is provided, comprising one or more programs executable by one or more processors of an electronic device, the one or more programs including instructions for performing an integrated control method for unmanned systems as described above.

[0035] Compared with the prior art, the present invention has the following advantages:

[0036] (1) Achieve integrated control of communication, calculation, control and analysis to reduce the impact of performance loss and information loss: For unmanned systems or unmanned system clusters, an integrated module is used to realize end-to-end task execution and action analysis. The integrated module can be seamlessly connected with the unmanned system body to realize communication with the control terminal, calculation of control signals, feedback and analysis of control results, avoid the long hierarchical call chain between traditional multi-modules, and reduce the problems of performance loss and information loss.

[0037] (2) Achieving autonomous maneuvering mission execution and mission planning and collaborative game control for unmanned system clusters: For the unmanned system itself, multi-source information fusion is used to autonomously complete path planning and obstacle avoidance. Simultaneously, by combining the unmanned system's onboard sensors, external ground control system, and external camera gimbal, an autonomous information interaction closed loop is achieved. When the external ground control system issues mission instructions to the unmanned system itself, an integrated module enables end-to-end mission execution and provides feedback on motion attitude information and mission execution status to the external ground control system. For the unmanned system cluster, multi-source information fusion is achieved. Combining reinforcement learning and game theory, reinforcement learning is used to output the optimal strategy, and game theory is used to determine whether the optimal strategy satisfies Pareto optimality for the unmanned system cluster mission. Simultaneously, by combining the unmanned system's onboard sensors, external ground control system, and external camera gimbal, an autonomous information interaction closed loop is achieved. When the external ground control system issues mission execution signals to the unmanned system cluster, an integrated module enables end-to-end execution of the optimal strategy and provides feedback on motion attitude information and mission execution status to the external ground control system. Attached Figure Description

[0038] Figure 1 This is a schematic diagram of the integrated communication-computing-control-analysis module in the embodiment;

[0039] Figure 2 A system relationship diagram of the integrated communication-computing-control-analysis module and method;

[0040] Figure 3 This is a schematic diagram illustrating the control flow of a multi-rotor unmanned aerial vehicle (UAV) as an example of an unmanned system.

[0041] The system includes: 1. Integrated communication-computing-control-analysis module; 2. External ground control system; 3. External camera gimbal; 4. Sensors; and 5. Unmanned system. Detailed Implementation

[0042] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0043] Example 1

[0044] To address the shortcomings of existing technologies, this embodiment provides an integrated control method for unmanned systems. A modular unit is designed for the unmanned system, integrating communication, computing, control, and diagnostic functions. This module supports seamless integration with the unmanned system itself, providing support for collaborative control of unmanned system clusters and full-process logic analysis and diagnostics.

[0045] like Figure 1 The communication-computing-control-analysis integrated module 1 in this embodiment includes a communication part, a computing part, a control part, and an analysis part. The communication-computing-control-analysis integrated module 1 is connected to the external ground control system 2, the external shooting gimbal 3, the sensor 4, and the unmanned system 5, respectively.

[0046] The external ground control system 2 is used to control or coordinately control the unmanned system itself by transmitting control commands to the unmanned system or unmanned system cluster via the wireless receiving device of the communication section of the integrated module. The external ground control system 2 is also used to receive analysis signals fed back by the unmanned system or unmanned system cluster via the wireless transmitting device of the communication section of the integrated module to correct control deviations or coordinate control deviations of the unmanned system itself. The control signals transmitted by the external ground control system 2 include path control, landing commands, attack commands, and return commands for the unmanned system itself, and cluster coordination commands, group attack commands, and group return commands for the unmanned system cluster. The control signals received by the external ground control system include the status, position, load, and mission execution status of the unmanned system or unmanned system cluster.

[0047] The external camera gimbal 3 is used to transmit high-resolution images of the unmanned system or unmanned system cluster's location and surrounding environment via the wireless receiving device of the unmanned system communication section. The transmitted signal from the external camera gimbal 3 includes environmental photos along the current path of the unmanned system or unmanned system cluster, photos of obstacles, photos of the locations of other cooperating unmanned systems, and photos of the locations of targets.

[0048] Sensor 4 is used to determine the internal situation of the unmanned system, specifically including an infrared sensor, an ultrasonic sensor, an accelerometer, an altitude sensor, and a navigation system. The infrared sensor uses infrared imaging to output an infrared image of the surrounding area of ​​the unmanned system. The ultrasonic sensor uses ultrasonic ranging to output information on the relative distances between the unmanned system and other unmanned systems, the relative distance between the unmanned system and a designated target location, and the relative altitude distance between the unmanned system and the ground. The accelerometer outputs acceleration and load information of the current unmanned system. The altitude sensor outputs the altitude information of the current unmanned system relative to a horizontal reference surface. The navigation system uses GPS to output the relative position of the current unmanned system in the current area.

[0049] The infrared sensor is specifically used to detect whether there are obstacles on the current path and whether there are other unmanned systems in the vicinity when capturing infrared image data of the current unmanned system. If obstacles or other unmanned systems are found, their positions are input to the integrated module for obstacle avoidance and collaborative task allocation. The ultrasonic sensor is specifically used to detect whether there are obstacles on the current path and whether there are other unmanned systems in the vicinity when probing the surroundings of the current unmanned system. If obstacles or other unmanned systems are found, their positions are input to the integrated module for obstacle avoidance and collaborative task allocation.

[0050] The communication section of the integrated module includes a wireless receiver and a wireless transmitter. The wireless receiver establishes communication connections with the external ground control system 2 and the external camera pan-tilt unit 3. The wireless transmitter establishes port connections with the sensors used to assess the internal status of the unmanned system and the analysis section of the integrated module, as well as with the external ground control system. The wireless transmitter transmits control signals to and from the sensor input / output ports for assessing the internal status of the unmanned system and the input / output ports of the analysis section of the integrated module.

[0051] The computing portion of the integrated module is hardware-supported by a portable computer, which establishes a port connection with the communication portion of the integrated module. The portable computer receives image and command information from the communication portion of the integrated module as input, performs multi-source information fusion, and uses intelligent algorithms for comprehensive analysis. It then performs task planning and allocation based on the input signals, converting the input signals into control signals for output. The specific computational tasks include multi-source information fusion, path determination and planning, obstacle type identification, determining whether obstacle avoidance is necessary, and collaborative game analysis to output the optimal strategy.

[0052] The control section of the integrated module includes hardware support provided by the main control system and gyroscopes, and establishes a port connection with the computing section. The main control system receives control information from the computing section of the integrated module as input, executes corresponding acceleration, deceleration, and other actions, and records the corresponding execution commands as output execution information. The gyroscope receives control information from the computing section of the integrated module as input, executes corresponding turning actions, and records the corresponding execution commands as output execution information. Specific output information from the control section includes: whether it is necessary to adjust the main control system to perform acceleration, deceleration, landing, and return operations to speed up, slow down, pause, or stop the current task; whether it is necessary to adjust the gyroscope to perform turning operations to avoid obstacles or avoid deviating from the path; whether it is necessary to execute sub-tasks in the collaborative task; and whether to take one action from the set of cooperation, competition, and inaction strategies for the unmanned system swarm at the current moment.

[0053] The analysis section of the integrated module is supported by a microcomputer. It establishes port connections with the control and communication sections of the integrated module. The microcomputer takes execution information received from the control section and instruction information received from the communication section as input, analyzes the current execution status, and outputs the analysis results to the communication system of the integrated module. The specific analysis content of the analysis section includes: combining the input information from the communication section and the output information from the control section of the integrated model, analyzing whether there are deviations between the task commands issued by the external ground control system and the action commands executed by the control section, thus achieving macro-level control of the unmanned system's own control; and analyzing whether the actions executed by the control section satisfy Pareto optimality in the collaborative tasks of the unmanned system swarm, thus achieving macro-level control of the unmanned system's own control.

[0054] Specifically, the steps of multi-source information fusion are as follows: input information from the communication part of the integrated module is preprocessed through normalization and standardization; data dimensionality reduction and data fusion are achieved through principal component analysis and neural networks; feature extraction is achieved through a classifier; and the processed data is stored. The purpose of multi-source information fusion is to obtain more accurate data and reduce data ambiguity.

[0055] Specifically, the steps for identifying obstacle types are as follows: inputting information from the communication section of the integrated module, including sensors for judging the internal situation and a large number of obstacle images collected by the external gimbal; performing preliminary processing on the input information through multi-source information fusion; converting the processed images into an obstacle image training set; inputting the training set into a convolutional neural network for image data training to generate a training model; inputting infrared images of the current unmanned system's surroundings captured by infrared sensors into the trained model to achieve obstacle recognition, and transmitting the identified obstacle information to the external ground control system via a wireless transmitter. The calculation formula for the convolutional neural network is shown in formula (1):

[0056] Y=σ(conv2(W,X)+b) (1)

[0057] Where W and b are the weight and bias parameters of the convolutional neural network, respectively, X refers to the input variable, which is generally the pixel value of the image, conv2() refers to the two-dimensional convolution function, and σ refers to the activation function in deep learning.

[0058] Specifically, the steps for determining whether to avoid an obstacle are as follows: input information from the communication section of the integrated module, including distance information to the obstacle captured by the ultrasonic sensor, acceleration information of the unmanned system body identified by the accelerometer, position information of the unmanned system body captured by the navigation system, and relative position information between the obstacle and the unmanned system body captured by the external camera gimbal; perform preliminary processing on the input information using a multi-source information fusion method; import the processed information into the dynamic model of the unmanned system itself, output whether there is a risk of collision, and then determine whether to avoid the obstacle. The dynamic equation of the unmanned system can be described by formula (2):

[0059]

[0060] Among them, F U The external force acting on the unmanned system is M, where m is the mass of the unmanned system. U The acceleration matrix of the unmanned system, v U and w U These refer to the velocity and angular velocity of the unmanned system, respectively.

[0061] Specifically, the steps for outputting the optimal strategy in the cluster cooperative game are as follows: input information from the communication part of the integrated module, including distance information from other unmanned system clusters captured by ultrasonic sensors, cooperative task information of the unmanned system cluster transmitted by the external ground control system, current position information of the unmanned system cluster captured by the navigation system, load information captured by the accelerometer, and relative position information of the unmanned system cluster captured by the external camera pan-tilt unit; convert the input information into a state transition function p(s′|s,a) and a reward function R(s,a), train it using reinforcement learning, and obtain the strategy function π(a|s) of the unmanned system cluster; according to the strategy function of the unmanned system cluster, make one of the actions in the strategy set, which includes three actions: cooperation, competition, and inaction. The solution formula for the strategy function is shown in (3):

[0062]

[0063] Where R(s,a) refers to the reward function, p(s′|s,a) refers to the state transition function, γ is the discount factor, and V π (s′) refers to the value function, Q π (s,a) refers to the state value function, and π(a|s) refers to the policy function.

[0064] Specifically, the macro-control steps of the analysis section of the integrated module are as follows: When analyzing whether there is a deviation between the task commands issued by the external ground control system and the action commands executed by the control section, the input information of the communication section and the output information of the control section are both converted into sequence information as the input information of the analysis section. It is then judged whether the sequence information of the two is the same and an analysis report is output. The analysis report is then transmitted to the communication section for feedback to the external ground control system. When analyzing whether the actions executed by the control section satisfy Pareto optimality in the collaborative task of the unmanned system cluster, the input information of the communication section and the output information of the control section are both converted into sequence information as the input information of the analysis section. Combining the body action of the unmanned system and the overall action of the unmanned system cluster, it is calculated whether the current unmanned system cluster satisfies Pareto optimality in executing the collaborative task and an analysis report is output. The analysis report is then transmitted to the communication section for feedback to the external ground control system.

[0065] like Figure 2-3 The multi-rotor UAV is a highly stable and expandable six-rotor industrial UAV with an all-carbon fiber fuselage. This embodiment will use the multi-rotor UAV as an unmanned system and implement it based on the technical solution of this invention.

[0066] The integrated communication-computing-control-analysis module 1 distinguishes between two mission execution scenarios for the multi-rotor UAV based on the different execution commands from the external ground control system, providing support for full-process logical analysis and diagnosis. The specific details are as follows:

[0067] Scenario 1: The current multi-rotor drone needs to perform the tasks of the multi-rotor drone itself.

[0068] Scenario 2: The current multi-rotor drone swarm needs to perform collaborative tasks.

[0069] like Figure 3 The above describes a flowchart (a) of a multi-rotor drone performing its own tasks, and a flowchart (b) of a multi-rotor drone swarm performing collaborative tasks. When a multi-rotor drone faces situation 1, the specific steps for performing the relevant tasks using the integrated module and method include:

[0070] Step S101: Build an integrated equipment platform including an external ground control system and an external camera gimbal, and deploy sensors and an integrated communication-computing-control-analysis module for unmanned systems.

[0071] Step S102: The external ground control system issues mission instructions for the multi-rotor UAV body to the unmanned system;

[0072] Step S103: Input the mission commands from the external ground control system and the data captured by the external camera pan-tilt unit and sensors into the integrated module via a wireless receiving device, and let the integrated module analyze the current situation.

[0073] Step S104: All input information is fused from multiple sources through the integrated module. The processed data is then fed into the pre-trained convolutional neural network to complete obstacle recognition and send the obstacle information to the external ground control system. The processed data is then fed into the dynamic model to calculate and identify whether there is a collision risk, and then to determine whether it is necessary to avoid the obstacle. The information on whether to perform obstacle avoidance action is then sent to the external ground control system.

[0074] Step S105: The external ground control system receives the analysis report from the integrated module and performs a subjective evaluation analysis. If the current action of the multi-rotor UAV meets the task command requirements issued by the external ground control system, the multi-rotor UAV is allowed to execute the current action by default. If the current action of the multi-rotor UAV does not meet the task command requirements issued by the external ground control system, the external ground control system will send a command to stop the current action of the multi-rotor UAV and require the integrated module to collect all data again and repeat steps S3 to S5 until the action of the multi-rotor UAV meets the task command requirements issued by the external ground control system, and then stop the process.

[0075] When a multi-rotor drone faces situation 2, the specific steps for performing the relevant tasks using the integrated module and method include:

[0076] Step S201: Build an integrated equipment platform including an external ground control system and an external camera gimbal, and deploy sensors and an integrated communication-computing-control-analysis module for unmanned systems;

[0077] Step S202: The external ground control system issues a collaborative task command for the multi-rotor UAV swarm to the unmanned system;

[0078] Step S203: Input the mission commands from the external ground control system and the data captured by the external camera pan-tilt unit and sensors into the integrated module via a wireless receiving device, and let the integrated module analyze the current situation.

[0079] Step S204: All input information is fused through the integrated module to form multi-source information. The processed data is then fed into the already trained reinforcement learning model to obtain the policy function π(a|s) of the current multi-rotor UAV cluster. Based on the policy function, the actions of the current multi-rotor UAV cluster are determined. It is determined whether the current actions meet the Pareto optimality of the multi-rotor UAV cluster cooperative task, and the analysis information is sent to the external ground control system.

[0080] Step S205: The external ground control system receives the analysis report from the integrated module and performs a subjective evaluation analysis. If the current actions of the multi-rotor UAV cluster meet the requirements of the collaborative task instructions issued by the external ground control system, the multi-rotor UAV cluster is allowed to execute the current actions by default. If the current actions of the multi-rotor UAV cluster do not meet the requirements of the collaborative task instructions issued by the external ground control system, the external ground control system will send an instruction to stop the current actions of the multi-rotor UAV cluster and require the integrated module to recollect all data and repeat steps S3 to S5 until the actions of the multi-rotor UAV cluster meet the requirements of the task instructions issued by the external ground control system, and then stop the process.

[0081] Compared with the prior art, the present invention has the following beneficial effects:

[0082] (1) The communication-computing-control-analysis integrated module and method for unmanned systems of the present invention builds an integrated equipment platform including an external ground control system and an external camera gimbal for unmanned systems or unmanned system clusters. The unmanned system body uses only one integrated module to realize end-to-end task execution and motion analysis. The module supports seamless connection with the unmanned system body, avoiding the long hierarchical call chain between traditional multi-modules, and reducing performance loss and information loss problems.

[0083] (2) The integrated communication-computation-control-analysis module and method for unmanned systems of the present invention realizes autonomous maneuvering mission execution of unmanned systems. For the unmanned system itself, through multi-source information fusion, it autonomously completes path planning and obstacle avoidance processing. At the same time, by combining the sensors carried by the unmanned system itself, the external ground control system and the external camera gimbal, an autonomous information interaction closed loop is realized. When the external ground control system gives the unmanned system a mission execution signal, the integrated module realizes end-to-end mission execution and feeds back motion attitude information and mission execution status to the external ground control system.

[0084] (3) The integrated communication-computation-control-analysis module and method for unmanned systems of the present invention realizes task planning and collaborative game control of unmanned system clusters. For unmanned system clusters, multi-source information fusion is realized. Combining reinforcement learning and game theory methods, the optimal strategy is output using reinforcement learning, and game theory is used to determine whether the optimal strategy satisfies the Pareto optimality of the unmanned system cluster task. At the same time, by combining the sensors carried by the unmanned system itself, the external ground control system, and the external camera gimbal, an autonomous information interaction closed loop is realized. When the external ground control system gives the unmanned system cluster task execution signal, the integrated module realizes the end-to-end execution of the optimal strategy and feeds back motion attitude information and task execution status to the external ground control system.

[0085] Example 2

[0086] This embodiment provides an electronic device, including: one or more processors and a memory, wherein the memory stores one or more programs, the one or more programs including instructions for executing the integrated control method for unmanned systems as described in Embodiment 1.

[0087] Example 3

[0088] This embodiment provides a computer-readable storage medium including one or more programs executable by one or more processors of an electronic device, the one or more programs including instructions for performing an integrated control method for unmanned systems as described in Embodiment 1.

[0089] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any person skilled in the art can easily conceive of various equivalent modifications or substitutions within the technical scope disclosed in the present invention, and these modifications or substitutions should all be covered within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A control method for an integrated control device for unmanned systems, characterized by, The method is applied to an integrated control module and specifically comprises the following steps: obtaining a task instruction and sensor information from a control terminal; based on the sensor information, obtaining fusion features through multi-source information fusion processing, identifying an obstacle type based on the fusion features, determining whether obstacle avoidance is needed, determining an optimal strategy under a swarm based on the task instruction, and obtaining a control signal; sending the control signal to an actuator and obtaining execution result information from the actuator; based on the task instruction and the execution result information, obtaining execution evaluation information and sending it to the control terminal, the process of the multi-source information fusion processing comprises the following steps: normalizing and standardizing the sensor information; realizing data dimension reduction and data fusion through principal component analysis and neural networks; for the features after data fusion, using a classifier to extract features and obtain fusion features, the process of determining an optimal strategy under a swarm and obtaining a control signal comprises the following steps: based on the obstacle type identification result and the determination result of whether obstacle avoidance is needed, selecting an optimal strategy from a strategy set including cooperation, competition and inaction according to a strategy function obtained through pre-training, and obtaining a corresponding control signal, the process of obtaining execution evaluation information comprises the following steps: obtaining a deviation evaluation index based on the task instruction and the execution result information; determining whether the currently selected strategy meets the Pareto optimality of a cooperative task and obtaining a determination result; based on the deviation evaluation index and the determination result, obtaining execution evaluation information, the integrated control device integrated module, at least one sensor, control terminal and actuator connected to the integrated module, the integrated module comprises: a communication unit for obtaining task instruction information and feeding back execution evaluation information, the communication unit is also used for obtaining sensor information of the sensor; a computing unit for obtaining a control signal through multi-source information fusion, obstacle type identification, determination of whether obstacle avoidance is needed and determination of an optimal strategy under a swarm; a control unit for sending the control signal to the actuator and obtaining execution result information from the actuator; an analysis unit for obtaining execution evaluation information based on the task instruction and the execution result information.

2. The control method of the integrated control device for unmanned systems according to claim 1, wherein the process of obstacle type identification comprises the following steps: inputting the fusion features obtained through multi-source information fusion into a trained obstacle identification model and obtaining an obstacle type identification result based on the output of the model.

3. The control method of the integrated control device for unmanned systems according to claim 1, wherein the process of determining whether obstacle avoidance is needed specifically comprises the following steps: inputting the fusion features obtained through multi-source information fusion into a preset dynamics model, determining whether there is a collision risk, and determining whether obstacle avoidance is needed.

4. The control method of the integrated control device for unmanned systems according to claim 1, wherein The sensor information includes infrared sensor information, ultrasonic information, acceleration information, height information, navigation information and image information from an external shooting gimbal.

5. The control method of the integrated control device for unmanned systems according to claim 1, wherein The task instruction includes a single task instruction and a cooperative task instruction.

6. A computer-readable storage medium, characterized in that, one or more programs for execution by one or more processors of an electronic device, the one or more programs including instructions for performing the control method of the integrated control device for unmanned systems according to any one of claims 1-5.

Images