WO2022001119A1 - 一种多智能体系统结构及其控制方法 - Google Patents

一种多智能体系统结构及其控制方法 Download PDF

Info

Publication number
WO2022001119A1
WO2022001119A1 PCT/CN2021/076258 CN2021076258W WO2022001119A1 WO 2022001119 A1 WO2022001119 A1 WO 2022001119A1 CN 2021076258 W CN2021076258 W CN 2021076258W WO 2022001119 A1 WO2022001119 A1 WO 2022001119A1
Authority
WO
WIPO (PCT)
Prior art keywords
agent
communication network
main control
network
agent system
Prior art date
Application number
PCT/CN2021/076258
Other languages
English (en)
French (fr)
Inventor
郭胜
苏世杰
王月阳
王为民
唐文献
Original Assignee
镇江宇诚智能装备科技有限责任公司
江苏科技大学
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 镇江宇诚智能装备科技有限责任公司, 江苏科技大学 filed Critical 镇江宇诚智能装备科技有限责任公司
Publication of WO2022001119A1 publication Critical patent/WO2022001119A1/zh

Links

Images

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W84/00—Network topologies
    • H04W84/02—Hierarchically pre-organised networks, e.g. paging networks, cellular networks, WLAN [Wireless Local Area Network] or WLL [Wireless Local Loop]
    • H04W84/04—Large scale networks; Deep hierarchical networks
    • H04W84/06—Airborne or Satellite Networks
    • G—PHYSICS
    • G05—CONTROLLING; REGULATING
    • G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
    • G05B19/00—Programme-control systems
    • G05B19/02—Programme-control systems electric
    • G05B19/04—Programme control other than numerical control, i.e. in sequence controllers or logic controllers
    • G05B19/042—Programme control other than numerical control, i.e. in sequence controllers or logic controllers using digital processors
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W84/00—Network topologies
    • H04W84/18—Self-organising networks, e.g. ad-hoc networks or sensor networks
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
    • Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
    • Y02P90/02—Total factory control, e.g. smart factories, flexible manufacturing systems [FMS] or integrated manufacturing systems [IMS]

Definitions

  • the invention belongs to the technical field of intelligent autonomous unmanned systems, and particularly relates to a multi-agent system structure and a control method thereof.
  • Autonomous unmanned system technology is one of the key technologies that artificial intelligence focuses on, and its main features are intelligence, control systems, dynamic networking and human-machine relationships.
  • the network capabilities provided by 5G technology will meet the needs of three extreme business scenarios, providing users with higher rates and better business experience, including high-speed, high-bandwidth business capabilities, which can support ultra-high-definition 3D video, VR , AR and other services; the low power consumption and high connection density capabilities provided can support applications such as monitoring, sensors, and smart cities; the ultra-low latency, high-reliability communication capabilities provided can support autonomous driving, telemedicine, and smart factories , artificial intelligence and other applications. Therefore, 5G technology can promote the needs of intelligent and automated communication in all walks of life.
  • a multi-agent system is composed of a series of interacting agent units. Each agent unit expresses the structure, function and behavioral characteristics of the system through communication, cooperation, coordination, scheduling, management and control, etc., to complete a single intelligence. A large number of complex tasks that can not be done by the body unit. Multi-agent systems have autonomy, distribution, coordination, self-organization ability, learning ability and reasoning ability. Therefore, multi-agent systems are used to solve practical problems, which can replace a single agent unit or manual work, which is difficult or impossible to complete. work, and has strong robustness and reliability.
  • the design of the above-mentioned intelligent body unit and intelligent body unit depends on the specified application scenarios and functional requirements. Therefore, the formed intelligent body unit and system function are single, and the application scenarios have limitations; when there are task requirements for different application scenarios, it is necessary to The system is redesigned or transformed, that is, the existing system cannot meet the diverse and changing use requirements of application scenarios; in addition, the function of a single agent unit in the above-mentioned multi-agent system is a preset specific and single function. , when an agent unit in the system fails, it will affect the operation of the entire system, and the system has poor robustness and insufficient reliability.
  • the target localization method and device disclosed in a method for target localization using digital images, first obtain the target Gabor filtering shape template, use the Gabor filtering result for shape matching localization, and then use the similarity between skeleton features determine the area where the target is located; patent a method and system for target positioning and identification of underwater robots (ZL201710209500.X) discloses a method for target positioning using sonar information; patented wireless sensor network static target positioning method and system (ZL201310145553.1 ) discloses a method for locating a target using a wireless network.
  • the multi-agent system needs to be able to perceive environmental information in various ways and perform multi-source heterogeneous signal fusion processing. Then, control the operation of the multi-agent system. If the target of the multi-agent system moves randomly, it is difficult for the existing multi-agent system to predict the possible position of the target at the next moment, so that the intelligent unit cannot obtain the information of the target at the next moment, or even lose the target.
  • Purpose of the invention Aiming at the problems existing in the prior art, to provide a multi-agent system and a control method thereof, which can quickly and reliably apply the multi-agent system in different scenarios and perform different tasks.
  • the initial operation strategy of the system is automatically generated, and the operation strategy is changed in real time during the operation process to improve the autonomy and coordination of the multi-agent system.
  • the present invention provides a multi-agent system, including a self-organizing wireless communication network, the self-organizing wireless communication network is connected to the Internet, and the Internet is respectively connected to a gateway device and a first communication network.
  • the base station is connected, the first communication network is also connected with a multi-channel interaction system, and the multi-channel interaction system is connected with a user; the gateway device is also connected with a data server.
  • the intelligent body unit includes a main control module and a plurality of peripheral modules connected thereto, and the main control module is connected with the peripheral modules through an interface;
  • the main control module includes a microcontroller, the common pins on the microcontroller are connected with the main control interface, the output line of the main control interface is connected with the ring-shaped main control terminal, the microcontroller is also connected with a power supply, and the main control module is connected to the main control module. Distribute multiple identical main control interfaces and main control terminals.
  • peripheral module includes a perception module, a communication module, a positioning module and an execution module;
  • the perception module perceives through images, voices, actions, and life signals
  • the communication module communicates through 5G, 4G, GPRS, CDMA, satellite communication, underwater acoustic communication, Zig-Bee, Bluetooth, and Wi-Fi;
  • the positioning module performs positioning through GPS, Wi-Fi, Bluetooth, and ultrasonic positioning methods.
  • the main control interface is provided with an AC oscillating circuit, a first digital switch and a second digital switch,
  • the AC oscillating circuit is used to generate an AC signal S i of a specific frequency and voltage, and the output line of the AC oscillating circuit is connected to the S line; the main control interface is connected to the first digital switch through the N 1 line, and the main control N 1 through the interface line and the second digital switch.
  • the main control module is connected to the peripheral module at any angle through the ring main control terminal, and a certain main control terminal of the main control module can be connected to multiple peripheral modules in parallel at the same time; a plurality of peripheral modules are distributed on the peripheral module.
  • Terminal, the peripheral terminal is connected to the peripheral interface, and the output line of the peripheral interface is connected to the peripheral actuator
  • the peripheral interface includes a voltage regulator circuit, a band-pass filter, a rectifier, and a latch;
  • the voltage regulator circuit and the output terminal of the power supply line is connected to the peripheral effector, the peripheral interface between the peripheral terminals and peripheral actuator wire N 1 - N 2 is provided with a third digital switch;
  • the AC signal S i input by the AC signal line S transmits a sinusoidal signal to the rectifier through a band-pass filter. After passing through the rectifier, the sinusoidal signal is converted into a DC signal to control the output state of the latch connected to it. The output of the latch passes through the rectifier.
  • the NOT gate circuit is connected with the third digital switch, and controls the opening or closing of the third digital switch;
  • the self-organizing wireless communication network is composed of an intelligent body unit and a wireless communication network transmission device.
  • the self-organizing wireless communication network is divided into an underwater self-organizing wireless communication network, an air self-organizing wireless communication network and a terrestrial self-organizing wireless communication network; correspondingly, the intelligent body units are also divided into underwater intelligent body units, aerial agent unit and terrestrial agent unit.
  • the underwater intelligent body unit is an underwater unmanned vehicle; the aerial intelligent body unit is an unmanned aerial vehicle; the land intelligent body unit is an unmanned vehicle.
  • the underwater self-organizing wireless communication network consists of an underwater unmanned vehicle, a water transfer station, a communication satellite and a satellite ground receiving station, and the water transfer station is connected with several underwater unmanned vehicles, so The underwater unmanned vehicle is connected with the satellite ground receiving station through the communication satellite.
  • the air self-organizing wireless communication network includes an unmanned aerial vehicle and a second communication network base station, and the second communication network base station is connected with several unmanned aerial vehicles.
  • the terrestrial self-organizing wireless communication network includes an unmanned vehicle and a third communication network base station, and the third communication network base station is connected to several unmanned vehicles.
  • a control method for a multi-agent system as described above characterized in that it comprises the following steps:
  • Step 1 Set the application scenario of the multi-agent system
  • the agent system select the sea, land and air joint search and rescue application scenario through the multi-channel interactive system, and generate the operating conditions and parameters of the multi-agent system according to the selected application scenario and the multi-agent knowledge base system, including the search and rescue coverage area, system operation Duration, search and rescue time period and detection target; the user inputs parameters through the multi-channel interactive system and uploads them to the data server;
  • Step 2 Build a multi-agent system
  • the multi-channel interaction system judges the composition type and quantity of the agent units in the multi-agent system according to the selected application scenario and the multi-agent knowledge base system, and determines the positioning method and communication method according to the number of the agent units and the movement path;
  • the multi-agent system communication network system consists of three parts: the interconnection network system, the satellite relay communication network system and the self-organizing wireless network system; the interconnection network system interconnects the wireless self-organizing network and the data storage server; the satellite relay communication network system The relay network connecting the underwater wireless ad hoc network and the interconnection network; the wireless ad hoc network, responsible for the network services of the multi-agent system, is divided into the terrestrial wireless ad hoc network, the air wireless ad hoc network and the underwater wireless ad hoc network ;
  • the underwater wireless self-organizing network uploads data and receives instructions from the multi-channel interactive system through the communication satellite equipment on the underwater unmanned vehicle; the communication satellite receives the data uploaded by the underwater wireless self-organizing network and the instructions from the multi-channel interactive system, And they are respectively downloaded to the satellite ground receiving station and the underwater wireless ad hoc network.
  • the satellite ground receiving station accesses the World Wide Web through the ground network to receive and send data and instructions; the land and air wireless ad hoc networks pass through the existing communication network.
  • the ground base station is connected to the World Wide Web to complete the data and instruction interaction with the data server and the multi-channel interactive system; the multi-agent system on land, sea and air completes the main controller selection, data transmission and instruction reception through the wireless self-organizing network;
  • Step 4 Deploy and run a multi-agent system and perceive environmental information
  • One of the intelligent body units is selected as the main control intelligent body unit;
  • the multi-channel interactive system sends control instructions to the main control intelligent body unit through the communication network, and the main control intelligent body unit sends instructions to control other intelligent body units to move to the deployment position ;
  • Intelligent body unit collection including environmental information information of image, voice, motion, and life signal, and transmit the collected information to the control module of the intelligent body unit;
  • the intelligent body unit transmits information to the main control intelligent body unit through the communication network, and the transmission path of the information is determined by the communication topology structure;
  • Step 5 Perform multi-source heterogeneous information fusion processing on the transmitted information
  • the multi-channel interactive system sends motion control instructions through the communication network according to the real-time positioning information of the target, and the intelligent unit moves after receiving the instructions, ensuring that the target is always in the best monitoring position of the intelligent unit; at the same time, the multi-channel interactive system displays the target position information;
  • the user views the location information of the target in real time on the multi-channel interactive system, controls the operation of the intelligent body unit through the operation of the multi-channel interactive system, and sends out actions, gestures, and voice commands, which are sensed by the multi-channel interactive system or the intelligent body unit.
  • the agent unit executes.
  • step (2) the specific steps of constructing the multi-agent system in the step (2) are as follows:
  • the multi-channel interaction system judges the composition type of the intelligent unit in the multi-agent system according to the selected application scenario and the multi-agent knowledge base system.
  • the intelligent unit in the sea, land and air joint search and rescue application scenario includes unmanned vehicles, unmanned aerial vehicles and water. unmanned aerial vehicle;
  • the multi-channel interactive system determines the detection method according to the search and rescue time period. If the working time period is daytime, the image or life sensing module is selected for detection. If the search and rescue time period is night, the infrared or life sensing module is selected for detection; the detection method is determined After that, the control system determines the number of intelligent body units according to the search and rescue coverage area and the detection range of a single intelligent body unit, and generates the movement path of the intelligent body unit at the same time;
  • the communication method between each agent unit is determined according to the distance D between each agent unit during operation:
  • D distance between each agent unit during operation:
  • 5G can be used
  • 4G GPRS
  • communication satellite short-wave communication
  • short-wave communication etc.
  • D ⁇ 500m Zig-Bee
  • Bluetooth Bluetooth
  • Wi-Fi wireless broadband
  • the multi-channel interactive system determines which positioning method to use according to the application scenario and positioning accuracy requirements, or uses a combination of multiple positioning methods: in the open field, the positioning accuracy is m-level, then GPS positioning is used; indoors, the positioning accuracy is cm-level , then use WIFI or Bluetooth; under water, the positioning accuracy is cm level: use ultrasonic positioning;
  • the multi-channel interactive system determines the number of mobile power stations that the multi-agent system needs to configure according to the endurance and running time of the agent unit.
  • the method for networking the multi-agent system in the step (3) includes the following steps:
  • the multi-agent system forms a network without obvious master-slave relationship according to the networking protocol, that is, each network node acts as the master node and initiates network transmission requests, and realizes the master-slave distribution of nodes through software according to task requirements;
  • Each multi-agent system relies on the flooding protocol to broadcast data packets, mainly including the address and machine code of the agent unit.
  • Each agent unit identifies the address and machine code of other agent units according to the received data packets. , and establish a routing table for it, and randomly assign an agent unit to the main controller;
  • the main control agent unit communicates with the upper computer through the gateway node and the communication network;
  • the host computer has a new algorithm for assigning the master controller of the multi-agent system, then re-designate the master controller agent unit, re-assign master-slave control nodes, repeat step (3.3), if not, the network is in the stage of waiting for the response command ;
  • the upper computer transmits the task instruction through the interconnected communication network; the task instruction comes from the user instruction extracted by the upper computer, or the task instruction generated according to the knowledge base in the data server;
  • the communication network sends the task instructions issued by the upper computer to each intelligent unit layer by layer;
  • Each intelligent unit transmits the corresponding command response or data back to the data server or the host computer through the land, sea and air wireless ad hoc network.
  • the multi-agent system control method can quickly and reliably apply the multi-agent system in different scenarios and perform different tasks.
  • the system will automatically generate the initial operation strategy of the system according to the selected application scenarios and input parameters, and The system operation strategy will change in real time during the operation process, improving the autonomy and coordination of the multi-agent system;
  • a multi-domain, multimedia, and multi-level knowledge base system is established to provide support for the dynamic change of the operation strategy of the multi-agent system and ensure the accurate operation of the multi-agent system in different application scenarios;
  • the knowledge base system and the multi-channel The interactive system interface is set up with a secondary retrieval mechanism to avoid excessive noise in information retrieval and improve retrieval efficiency and accuracy;
  • the intelligent body unit adopts a modular design and is composed of multiple functional modules. Each functional module is connected by a quick plug-in method, which can be quickly spliced and combined into intelligent body units with different functions, and intelligent body units with the same function. Different forms can be spliced, so that the intelligent unit can perform tasks in three states of sea, land and air, and meet the needs of multi-scenario applications.
  • FIG. 1 is a schematic structural diagram of a multi-agent system structure of the present invention.
  • Fig. 2 is the flow chart of the control method of the multi-agent system structure of the present invention.
  • Fig. 3 is the structural schematic diagram of the agent in Fig. 1;
  • Figure 4 is a schematic diagram of the interface connection of the agent module.
  • a multi-agent system structure includes an ad hoc wireless communication network, the ad hoc wireless communication network is connected to the Internet 10, and the Internet 10 is in turn connected to a gateway device 11 and a first communication network base station 13 respectively
  • the first communication network is also connected with the multi-channel interaction system 15 , and the multi-channel interaction system 15 is connected with the user 14 ; the gateway device 11 is also connected with the data server 12 .
  • the self-organizing wireless communication network is composed of the agent unit 1 and the wireless communication network transmission device.
  • the self-organizing wireless communication network is divided into an underwater self-organizing wireless communication network, an air self-organizing wireless communication network and a terrestrial self-organizing wireless communication network; correspondingly, the intelligent body unit 1 is also divided into an underwater intelligent body unit 1. , Air Agent Unit 1 and Land Agent Unit 1.
  • the underwater intelligent body unit 1 is an underwater unmanned vehicle 2 ; the aerial intelligent body unit 1 is an unmanned aerial vehicle 7 ; the land intelligent body unit 1 is an unmanned vehicle 9 .
  • the underwater self-organizing wireless communication network is composed of an underwater unmanned vehicle 2, a water transfer station 3, a communication satellite 4 and a satellite ground receiving station 5, and the water transfer station 3 is connected with several underwater unmanned aerial vehicles.
  • the vehicle 2 is connected, and the underwater unmanned vehicle 2 is connected with the satellite ground receiving station 5 through the communication satellite 4 .
  • the air ad hoc wireless communication network includes an unmanned aerial vehicle 7 and a second communication network base station 6 , and the second communication network base station 6 is connected to several unmanned aerial vehicles 7 .
  • the terrestrial ad hoc wireless communication network includes an unmanned vehicle 9 and a third communication network base station 8 , and the third communication network base station 8 is connected to several unmanned vehicles 9 .
  • the modularly designed agent includes a main control module 1.1 and multiple peripheral modules 1.9, and the peripheral module 1.9 includes a perception module 1.91, a communication module 1.92, a positioning module 1.93 and an execution module 1.94;
  • the agent can include multiple and multi-type perception modules 1.91, communication modules 1.92, positioning modules 1.93 and execution modules 1.94;
  • the perception module 1.91 includes the types of modules that perceive signals such as images, voices, actions, and life signals;
  • the communication module 1.92 includes Module types of 5G, 4G, GPRS, CDMA, satellite communication, underwater acoustic communication, Zig-Bee, Bluetooth (Bluetooth), wireless broadband (Wi-Fi) and other communication methods;
  • positioning module 1.93 includes GPS, Wi-Fi, Bluetooth, Module types of positioning methods such as ultrasonic;
  • Execution module 1.94 includes types of wheels, propellers, manipulators, lights, etc.;
  • Each functional module is provided with interface 1.10 in multiple directions, which is convenient for quick combination, expansion, and replacement of functional modules, and at the same time, the interface 1.10 can realize signal and energy transmission between components;
  • the connection principle of interface 1.10 is shown in Figure 4.
  • the main control module 1.1 includes a microcontroller 1.2.
  • the common pins on the microcontroller 1.2 are connected to the main control interface 1.3.
  • the output line of the main control interface 1.3 is connected to the ring main control terminal 1.4.
  • Connection, the microcontroller 1.2 is connected to the power supply 1.5, and the main control module 1.1 is distributed with multiple identical main control interfaces 1.3 and main control terminals 1.4;
  • the output line of the main control interface 1.3 includes a power line V, a ground line G, an AC signal line S and a number of control lines. In this embodiment, there are two control lines, which are N 1 and N 2 respectively .
  • the main control interface 1.3 is provided with AC
  • the oscillating circuit 1.6 is used to generate an alternating current signal S i of a specific frequency and voltage.
  • the output line of the alternating current oscillating circuit 1.6 is connected to the S line.
  • the main control interface 1.3 is also provided with a first digital switch 1.7 and a second digital switch 1.8. are respectively connected to lines N 1 - N 2 line, a digital micro-controller controls the first switching state and a second digital switch 1.7 to 1.8, so that N 1 - N 2 selectivity of a particular connection pin microcontroller with 1.2 ;
  • the main control module 1.1 is connected to the peripheral module 1.9 at any angle through the ring main control terminal 1.4, and a certain main control terminal 1.4 of the main control module 1.1 can be connected to multiple peripheral modules 1.9 in parallel at the same time; the peripheral modules 1.9 are distributed on the A plurality of peripheral terminals 1.110, the peripheral terminal 1.110 is connected to the peripheral interface 1.111, and the output line of the peripheral interface 1.111 is connected to the peripheral actuator 1.113;
  • Voltage regulating circuit comprises a peripheral interface 1.111 1.114, 1.114 regulator circuit and the output terminal of the actuator peripheral power cable 1.113, 1.111 peripheral interface peripherals and peripheral terminals actuators 1.110 1.113 N 2 and N 1 of the wire There is a third digital switch 1.112 between;
  • the AC signal S i input by the AC signal line S transmits a sinusoidal signal to the rectifier 116 through the band-pass filter 1.115, and the sinusoidal signal is converted into a DC signal after passing through the rectifier 1.116 to control the output state of the latch 1.117 connected to it.
  • the output terminal of 1.117 is connected to the third digital switch 1.112 through the NOT gate circuit, and controls the opening or closing of the third digital switch 1.112;
  • Output of the rectifier is connected between 1.118 and 1.116 N 1 line via the digital switch, the output terminal of the latch 1.117 1.118 controlling the digital switch is turned off or closed;
  • the band-pass filters 1.115 on different peripheral modules 1.9 can only pass the alternating current signal S i of a specific frequency and voltage, and do not repeat each other.
  • a control method of the above-mentioned multi-agent system structure comprising the following steps:
  • Step 1 Set the application scenario of the multi-agent system
  • the running agent system select the sea, land and air joint search and rescue application scenario through the multi-channel interaction system 15, and generate the operating conditions and parameters of the multi-agent system 1 system according to the selected application scenario and the multi-agent knowledge base system, including the search and rescue coverage area. , the system running time, the search and rescue time period and the detection target; the user 14 inputs the parameters through the multi-channel interactive system 15 and uploads them to the data server 12;
  • Step 2 Build a multi-agent system 1 system
  • the multi-channel interaction system 15 determines the composition type and quantity of the agent units 1 in the multi-agent system 1 system according to the selected application scenario and the knowledge base system of the multi-agent system 1, and determines the positioning method according to the number of the agent units 1 and the movement path. and means of communication;
  • the communication network system consists of three parts: the interconnection network system, the satellite relay communication network system and the self-organizing wireless network system; the interconnection network system interconnects the wireless self-organizing network and the data storage server; the satellite relay communication system The network system connects the underwater wireless ad hoc network and the relay network of the interconnecting network; the wireless ad hoc network is responsible for the network services of the multi-agent system 1, and is divided into terrestrial wireless ad hoc networks, air wireless ad hoc networks, and underwater wireless ad hoc networks. self-organizing network;
  • the underwater wireless self-organizing network uploads data and receives commands from the multi-channel interactive system 15 through the communication satellite 4 equipment on the underwater unmanned vehicle 2; the communication satellite 4 receives the data uploaded by the underwater wireless self-organizing network and multi-channel interaction
  • the instructions of the system 15 are transmitted to the satellite ground receiving station 5 and the underwater wireless ad hoc network respectively, and the satellite ground receiving station 5 is connected to the World Wide Web through the ground network to receive and send data and instructions;
  • the network accesses the World Wide Web through the ground base station of the existing communication network to complete the data and instruction interaction with the data server 12 and the multi-channel interaction system 15; the multi-agent system 1 of the land, sea and air completes the selection of the main controller, data transmission and instruction reception;
  • Step 4 Deploy and run the multi-agent system 1 system and perceive environmental information
  • One of the intelligent body units 1 is selected as the main control intelligent body unit 1; the multi-channel interactive system 15 sends control instructions to the main control intelligent body unit 1 through the communication network, and the main control intelligent body unit 1 sends instructions to control other intelligent The body unit 1 moves to the deployment position;
  • the intelligent body unit 1 collects environmental information including images, voices, actions, and life signals, and transmits the collected information to the control module of the intelligent body unit 1;
  • the intelligent body unit 1 transmits the information to the main control intelligent body unit 1 through the communication network, and the transmission path of the information is determined by the communication topology structure;
  • Step 5 Perform multi-source heterogeneous information fusion processing on the transmitted information
  • Step 6 Multi-agent system 1 system task assignment execution
  • the multi-channel interaction system 15 sends motion control instructions through the communication network according to the real-time positioning information of the target, and the agent unit 1 moves after receiving the instructions, ensuring that the target is always in the best monitoring position of the agent unit 1; at the same time, the multi-channel interaction system 15 displays target location information;
  • Step 7. User 14 controls the operation
  • the user 14 checks the location information of the target in real time on the multi-channel interactive system 15, controls the operation of the intelligent body unit 1 through the operation of the multi-channel interactive system 15, and issues actions, gestures, and voice commands, and the multi-channel interactive system 15 or the intelligent body unit 1 After sensing, a running command is issued to control the execution of the intelligent body unit 1.
  • step (2) the specific steps of constructing the multi-agent system 1 system in the step (2) are as follows:
  • the multi-channel interaction system 15 determines the composition type of the agent unit 1 in the multi-agent system according to the selected application scenario in combination with the multi-agent knowledge base system.
  • the multi-channel interactive system 15 determines the detection method according to the search and rescue time period. If the working time period is daytime, the image or life sensing module is selected for detection. If the search and rescue time period is night, the infrared or life sensing module is selected for detection; the detection method After the determination, the control system determines the number of intelligent body units 1 according to the search and rescue coverage area and the detection range of a single intelligent body unit 1, and generates the movement path of the intelligent body unit 1 at the same time;
  • the communication mode between each agent unit 1 is determined according to the distance D between each agent unit 1 during operation: in land and air scenarios, when D>500m, you can Adopt 5G, 4G, GPRS, communication satellite 4, short-wave communication, etc.; when D ⁇ 500m, Zig-Bee, Bluetooth (Bluetooth), wireless broadband (Wi-Fi), etc. can be used; underwater scene uses underwater acoustic communication;
  • the multi-channel interactive system 15 determines which positioning method to use according to the application scenario and positioning accuracy requirements, or uses a combination of multiple positioning methods: in an open field, the positioning accuracy is m-level, GPS positioning is used; indoors, the positioning accuracy is cm level, use WIFI or Bluetooth; under water, the positioning accuracy is cm level: use ultrasonic positioning;
  • the multi-channel interaction system 15 determines the number of mobile power stations that need to be configured in the multi-agent system according to the endurance and running time of the agent unit 1 .
  • the method for networking a multi-agent system in the step (3) includes the following steps:
  • the multi-agent system forms a network without obvious master-slave relationship according to the networking protocol, that is, each network node acts as the master node and initiates network transmission requests, and realizes the master-slave distribution of nodes through software according to task requirements;
  • Each multi-agent system relies on the flooding protocol to broadcast data packets, mainly including the address and machine code of the agent unit.
  • Each agent unit identifies the address and machine code of other agent units according to the received data packets. , and establish a routing table for it, and randomly assign an agent unit to the main controller;
  • the main control agent unit communicates with the upper computer through the gateway node and the communication network;
  • the host computer has a new algorithm for assigning the master controller of the multi-agent system, then re-designate the master controller agent unit, re-assign master-slave control nodes, repeat step (3.3), if not, the network is in the stage of waiting for the response command ;
  • the upper computer transmits the task instruction through the interconnected communication network; the task instruction comes from the user instruction extracted by the upper computer, or the task instruction generated according to the knowledge base in the data server;
  • the communication network sends the task instructions issued by the host computer to each intelligent unit layer by layer;
  • Each intelligent unit transmits the corresponding command response or data back to the data server or the upper computer through the land, sea and air wireless ad hoc network.

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Signal Processing (AREA)
  • Astronomy & Astrophysics (AREA)
  • Automation & Control Theory (AREA)
  • Mobile Radio Communication Systems (AREA)

Abstract

本发明公开了一种多智能体系统结构，包括自组织无线通信网络，所述自组织无线通信网络与互联网连接，所述互联网分别又与网关设备和第一通信网络基站连接，所述第一通信网络还与多通道交互系统连接，所述多通道交互系统与用户连接；所述网关设备还与数据服务器连接。本发明可以迅速、可靠的将多智能体系统应用在不同场景中，执行不同的任务，系统会根据选择的应用场景和输入的参数，自动生成系统初始运行策略，并且在运行过程中实时改变运行策略，具有较高的自主性和协调性。

Description

一种多智能体系统结构及其控制方法 技术领域
本发明属于智能自主无人系统技术领域，特别涉及一种多智能体系统结构及其控制方法。
背景技术
自主无人系统技术是人工智能重点关注的关键技术之一，其主要特点是智能化、控制系统、动态组网和人机关系。为此，5G技术提供的网络能力将满足三大极限业务场景需求，为用户提供更高速率和更好的业务体验，包括提供的高速率、高带宽业务能力，可支持超高清3D视频、VR、AR等业务；提供的低功耗、高连接密度能力，可支持监控、传感器、智慧城市等应用；提供的超低时延、高可靠性通信能力，可支持自动驾驶、远程医疗、智慧工厂、人工智能等应用。因此5G技术可促进各行各业的智能化和自动化通信需求。
多智能体系统是由一系列相互作用的智能体单元构成，各个智能体单元之间通过通信、合作、协调、调度、管理和控制等方式来表达系统的结构、功能及行为特性，完成单个智能体单元不能完成的大量而又复杂的工作。多智能体系统具有自主性、分布性、协调性，并具有自组织能力、学习能力和推理能力，因而采用多智能体系统解决实际问题，可以代替单个智能体单元或人工，很难或无法完成的工作，并具有很强的鲁棒性和可靠性。
多智能体系统自20世纪70年代被提出以来，就在各个领域迅速地得到了应用，例如军事协同作战、航空编队、城市管理、智能交通、联合搜救、海底探测等。专利“一种无人飞行器城市空中管理系统”(CN201711285844.5)、“一种应用于智能交通的无人机系统”(CN201810734095.8)、“一种无人机群搜索救援方法及系统”(CN201810892369.6)、“用于海底电缆巡检的水下机器人控制系统及方法”(CN201711259977.5)等，实现了多智能体系统在具体指定的不同场景下的应用，并有效、迅速的完成指定的相关作业任务，自动化、智能化程度较高；“一种基于图形化表示的多智能体系统生成方法”(ZL201410133930.4)采用图形化方式编辑构建自定义的多智能体系统，提高了智能体单元的开发效率，该方法可以根据用户需求，快速生成多种多智能体系统；
但上述智能体单元及智能体单元的设计取决于指定的应用场景和功能需求，因此形成的智能体单元及系统功能单一，应用场景存在局限性；当有不同应用场景任务需求时，需要对原系统进行重新设计或改造，即现有系统无法满足应用场景多样、变化的使用需求；此外，上述多智能体系统中的单个智能体单元，在系 统中的功能是预先设定的具体、单一功能，当系统中某个智能体单元发生故障时，会影响整个系统的运行，系统鲁棒性差、可靠性不足。
在多样、变化的应用场景中，多智能体系统对环境信息的感知精准度对于控制多智能体系统运行至关重要，但目前多智能体系统感知场景信息的方式单一，如专利一种数字图像中的目标定位方法以及装置(ZL201410359215.2)公开了一种利用数字图像进行目标定位的方法，首先获取目标Gabor滤波形状模板，利用Gabor滤波结果进行形状匹配定位，再利用骨架特征之间的相似度判定目标所在区域；专利一种水下机器人目标定位识别方法和系统(ZL201710209500.X)公开了一种利用声呐信息进行目标定位的方法；专利无线传感器网络静止目标定位方法及系统(ZL201310145553.1)公开了一种利用无线网络进行目标定位的方法。
当多智能体系统的应用环境变化影响了信息获取的精度，则会导致感知的环境信息精准度降低，因此需要多智能体系统能够采用多种方式感知环境信息，进行多源异构信号融合处理后，控制多智能体系统运行。若多智能体系统的目标是随机运动的，现有多智能体系统很难预测下一时刻目标可能出现的位置，导致智能体单元不能较好的获取下一时刻目标的信息，甚至丢失目标。
此外，现有的多智能体系统在实际应用过程中，人机交互性差，无法实现多通道全息的信息交换，用户操控系统的便捷度和精准度较低。
因此，需要开发适用于多领域场景，具备动态运行策略，自主性、协调性、可靠性更高，且人机交互友好的多智能体系统。
发明内容
发明目的：针对现有技术存在的问题，提供一种多智能体系统及其控制方法，可以迅速、可靠的将多智能体系统应用在不同场景中，执行不同的任务，系统会根据选择的应用场景和输入的参数，自动生成系统初始运行策略，并且在运行过程中实时改变运行策略，提升多智能体系统自主性和协调性。
技术方案：为解决上述技术问题，本发明提供一种多智能体系统，包括自组织无线通信网络，所述自组织无线通信网络与互联网连接，所述互联网分别又与网关设备和第一通信网络基站连接，所述第一通信网络还与多通道交互系统连接，所述多通道交互系统与用户连接；所述网关设备还与数据服务器连接。
进一步的，所述智能体单元包括主控模块和与之连接的多个外设模块，所述主控模块通过接口与外设模块连接；
所述各个模块之间通过接口连接进行信号和能量传递；
所述主控模块包括微控制器，微控制器上的常用引脚与主控接口连接，主控 接口的输出线与环状主控端子连接，微控制器还连接有电源，主控模块上分布多个相同的主控接口和主控端子。
进一步的，所述外设模块包括感知模块、通信模块、定位模块和执行模块；
所述感知模块通过图像、语音、动作、生命信号进行感知；
所述通信模块通过5G、4G、GPRS、CDMA、卫星通信、水声通信、Zig-Bee、蓝牙、Wi-Fi方式进行通信；
所述定位模块通过GPS、Wi-Fi、蓝牙、超声波的定位方式进行定位。
进一步的，所述主控接口上设有交流振荡电路、第一数字开关和第二数字开关，
所述交流振荡电路用于生成特定频率与电压的交流电信号S i，交流振荡电路的输出线与S线连接；所述主控接口通过N 1线与第一数字开关连接，所述主控接口通过N 1线与第二数字开关。
进一步的，主控模块通过环状主控端子与外设模块以任意角度相连，且主控模块的某个主控端子可同时并行连接多个外设模块；外设模块上分布多个外设端子，外设端子与外设接口连接，外设接口输出线连接外设执行器
进一步的，所述外设接口内包括调压电路、带通滤波器、整流器、锁存器；
所述调压电路的输出端与外设执行器的电源线连接，外设接口在外设端子和外设执行器的N 1和N 2线之间设有第三数字开关；
交流信号线S输入的交流电信号S i经过带通滤波器向整流器传递正弦信号，正弦信号经过整流器后被转为直流信号控制与其相连的锁存器的输出状态，锁存器的输出端通过非门电路与第三数字开关连接，控制第三数字开关的断开或闭合；
整流器的输出端与N 1线之间通过第四数字开关连接，锁存器的输出端控制第四数字开关的断开或闭合。
进一步的，所述自组织无线通信网络由智能体单元和无线通信网络传输装置构成。
进一步的，所述自组织无线通信网络分为水下自组织无线通信网络、空中自组织无线通信网络和陆地自组织无线通信网络；对应的，智能体单元也分为水下智能体单元、空中智能体单元和陆地智能体单元。
进一步的，所述水下智能体单元为水下无人航行器；所述空中智能体单元为 无人机；所述陆地智能体单元为无人车。
进一步的，所述水下自组织无线通信网络由水下无人航行器、水上中转站、通信卫星和卫星地面接收站构成，所述水上中转站与若干个水下无人航行器连接，所述水下无人航行器通过通信卫星与卫星地面接收站连接。
进一步的，所述空中自组织无线通信网络包括无人机和第二通信网络基站，所述第二通信网络基站与若干个无人机连接。
进一步的，所述陆地自组织无线通信网络包括无人车和第三通信网络基站，所述第三通信网络基站与若干个无人车连接。
一种如上所述的多智能体系统的控制方法，其特征在于，包括如下步骤：
步骤1、设置多智能体系统应用场景
启动运行智能体系统，通过多通道交互系统选择海陆空联合搜救应用场景，根据选取的应用场景结合多智能体知识库系统生成多智能体系统的运行条件和参数，具体包括搜救覆盖面积，系统运行时长，搜救时间段和检测目标；用户通过多通道交互系统输入参数，并上传到数据服务器中；
步骤2、构建多智能体系统
多通道交互系统根据选择的应用场景结合多智能体知识库系统判断多智能体系统中智能体单元的组成类型和数量，并根据智能体单元数量和运动路径确定定位方式和通信方式；
步骤3、多智能体系统组网
多智能体系统通信网络系统由互连网络系统、卫星中继通信网络系统和自组织无线网络系统三部分构成；互连网络系统互连无线自组织网络和数据存储服务器；卫星中继通信网络系统联系水下无线自组织网络和互连网络的中继网络；无线自组织网络，负责多智能体系统的网络服务，分为陆地无线自组织网络、空中无线自组织网络和水下无线自组织网络；
水下无线自组织网络通过水下无人航行器上的通信卫星设备上传数据和接收来自多通道交互系统的指令；通信卫星接收水下无线自组织网络上传的数据和多通道交互系统的指令，并分别下传给卫星地面接收站和水下无线自组织网络，卫星地面接收站通过地面网络接入万维网进行数据和指令的接收和发送；陆上和空中无线自组织网络则通过现有通信网络的地面基站接入万维网完成与数据服 务器和多通道交互系统的数据和指令交互；陆海空中的多智能体系统通过无线自组织网络完成主控制器选择、数据传输和指令接收；
步骤4、部署运行多智能体系统并感知环境信息
智能体单元中选择一个智能体单元作为主控智能体单元；多通道交互系统通过通信网络将控制指令发送到主控智能体单元，主控智能体单元发送指令控制其它智能体单元运动到部署位置；
智能体单元采集，包括图像、语音、动作、生命信号的环境信息信息，并将采集到的信息传送给智能体单元的控制模块；
智能体单元通过通信网络将信息传输到主控智能体单元，由通信拓扑结构确定信息的传输路径；
步骤5、将传输的信息进行多源异构信息融合处理
步骤6、多智能体系统任务分配执行
多通道交互系统根据目标实时定位信息，通过通信网络发送运动控制指令，智能体单元接收到指令后运动，保证目标始终在智能体单元的最佳监测位置；同时多通道交互系统显示目标位置信息；
步骤7、用户控制操作
用户在多通道交互系统上实时查看目标的位置信息，通过多通道交互系统操作控制智能体单元运行，并发出动作、手势、语音指令，由多通道交互系统或智能体单元感知后发出运行指令控制智能体单元执行。
进一步的，所述步骤(2)中构建多智能体系统的具体步骤如下：
多通道交互系统根据选择的应用场景结合多智能体知识库系统判断多智能体系统中智能体单元的组成类型，海陆空联合搜救应用场景下的智能体单元包括无人车、无人机和水下无人航行器；
多通道交互系统根据搜救时间段来确定检测方式，若工作时间段为白天，则选择图像或生命感知模块进行检测，若搜救时间段为夜晚，则选择红外或生命感知模块进行检测；检测方式确定后，控制系统根据搜救覆盖面积和单个智能体单元的检测范围，确定智能体单元的数量，同时生成智能体单元的运动路径；
多智能体系统数量和运动路径确定后，根据运行过程中各个智能体单元之间的距离D来确定各个智能体单元之间的通信方式：陆地和空中场景下，当 D>500m，可采用5G、4G、GPRS、通信卫星、短波通信等；当D<500m，可采用Zig-Bee、蓝牙(Bluetooth)、无线宽带(Wi-Fi)等；水下场景则采用水声通信；
多通道交互系统根据应用场景和定位精度需求，确定采用何种定位方式，或多种定位方式组合使用：在开阔场地，定位精度为m级，则采用GPS定位；在室内，定位精度为cm级，则采用WIFI或蓝牙；在水下，定位精度为cm级：则采用超声波定位；
多通道交互系统根据智能体单元的续航能力和运行时长，判断多智能体系统需要配置移动电站的数量。
进一步的，所述步骤(3)中多智能体系统组网的方法包括以下步骤：
(3.1)多智能体系统组网
多智能体系统依据组网协议组成无明显主从关系的网络，即每个网络节点都做为主节点和发起网络传输请求，根据任务需求通过软件实现节点的主从分配；
(3.2)建立路由表
每个多智能体系统依赖泛洪协议向外广播数据包，主要包括智能体单元的地址、机器码，每个智能体单元根据接收到的数据包，识别出其它智能体单元的地址和机器码，并为其建立路由表，随机分配一个智能体单元为主控制器；
(3.3)主控制节点接入主干网络
主控智能体单元通过网关节点和通信网络与上位机通信；
(3.4)上位机接收信息
若上位机有新的分配多智能体系统主控制器的算法，则重新指定主控制器智能体单元，重新分配主从控制节点，重复步骤(3.3)，若无，则网络处于等待响应指令阶段；
(3.5)上位机发出任务指令
上位机通过互联通信网络传送任务指令；任务指令来源于上位机提取的用户指令，或根据数据服务器中知识库生成的任务指令；
(3.6)通信网络逐层响应
通信网络将上位机发出的任务指令逐层发送到各个智能体单元；
(3.7)智能体单元响应
各个智能体单元通过陆海空无线自组织网络将相应的指令响应或者数据传 送回数据服务器或者上位机。
与现有技术相比，本发明的优点在于：
(1)多智能体系统控制方法可以迅速、可靠的将多智能体系统应用在不同场景中，执行不同的任务，系统会根据选择的应用场景和输入的参数，自动生成系统初始运行策略，并且系统运行策略会在运行过程中实时改变，提升多智能体系统自主性和协调性；
(2)建立了多领域、多媒体、多层次的知识库系统，为多智能体系统实现运行策略动态改变提供支撑，保障多智能体系统在不同应用场景下的精准运行；知识库系统与多通道交互系统接口设置二级检索机制，避免信息检索噪音过大，提高检索效率和精准度；
(3)智能体单元采用模块化设计，由多个功能模块组成，每个功能模块通过快速插装的方式连接，可以快速拼接组合成具备不同功能的智能体单元，并且同一功能的智能体单元可拼接不同的形态，使得智能体单元实现海陆空三种状态下执行任务的功能，满足多场景应用需求。
附图说明
图1为本发明多智能体系统结构的结构示意图；
图2为本发明多智能体系统结构的控制方法流程图；
图3为图1中智能体的结构示意图；
图4为智能体模块接口连接原理图。
图中：1、智能体单元；2、水下无人航行器；3、水上中转站；4、通信卫星，5、卫星地面接收站，6、第二通信网络基站，7、无人机，8、第三通信网络基站，9、无人车，10、互联网，11、网关设备，12、数据服务器，13、第一通信网络基站，14、用户，15、多通道交互系统；
1.1、主控模块；1.2、微控制器；1.3、主控接口；1.4、主控端子；1.5、电源；1.6、交流振荡电路；1.7、第一数字开关；1.8、第二数字开关；1.9、外设模块；1.91、感知模块；1.92、通信模块；1.93、定位模块；1.94、执行模块；1.10、接口；1.110、外设端子；1.111、外设接口；1.112、第三数字开关；1.113、外设执行器；1.114、调压电路；1.115、带通滤波器；1.116、整流器；1.117、锁存器；1.118、第四数字开关。
具体实施方式
下面结合附图和具体实施方式，进一步阐明本发明。本发明描述的实施例仅仅是本发明的一部分实施例，而不是全部的实施例。基于本发明中的实施例，本领域普通技术人员在没有做出创造性劳动的前提下所得到的其他实施例，都属于本发明所保护的范围。
如图1所示，一种多智能体系统结构，包括自组织无线通信网络，所述自组织无线通信网络与互联网10连接，所述互联网10分别又与网关设备11和第一通信网络基站13连接，所述第一通信网络还与多通道交互系统15连接，所述多通道交互系统15与用户14连接；所述网关设备11还与数据服务器12连接。
进一步的，所述自组织无线通信网络由智能体单元1和无线通信网络传输装置构成。
进一步的，所述自组织无线通信网络分为水下自组织无线通信网络、空中自组织无线通信网络和陆地自组织无线通信网络；对应的，智能体单元1也分为水下智能体单元1、空中智能体单元1和陆地智能体单元1。
进一步的，所述水下智能体单元1为水下无人航行器2；所述空中智能体单元1为无人机7；所述陆地智能体单元1为无人车9。
进一步的，所述水下自组织无线通信网络由水下无人航行器2、水上中转站3、通信卫星4和卫星地面接收站5构成，所述水上中转站3与若干个水下无人航行器2连接，所述水下无人航行器2通过通信卫星4与卫星地面接收站5连接。
进一步的，所述空中自组织无线通信网络包括无人机7和第二通信网络基站6，所述第二通信网络基站6与若干个无人机7连接。
进一步的，所述陆地自组织无线通信网络包括无人车9和第三通信网络基站8，所述第三通信网络基站8与若干个无人车9连接。
如图3所示，模块化设计的智能体包括主控模块1.1和多个外设模块1.9，外设模1.9包括感知模块1.91、通信模块1.92、定位模块1.93和执行模块1.94；
智能体可包含多个、多类型的感知模块1.91、通信模块1.92、定位模块1.93和执行模块1.94；其中感知模块1.91包括感知图像、语音、动作、生命信号等信号的模块类型；通信模块1.92包括5G、4G、GPRS、CDMA、卫星通信、水声通信、Zig-Bee、蓝牙(Bluetooth)、无线宽带(Wi-Fi)等通信方式的模块类型；定位模块1.93包括GPS、Wi-Fi、蓝牙、超声波等定位方式的模块类型；执行模 块1.94包括轮子、螺旋桨、机械手、照明灯等类型；
每个功能模块在多个方向上设有接口1.10，便于快速组合、扩展、更换功能模块，同时接口1.10可以实现部件之间信号和能量传递；
接口1.10连接原理如图4所示，主控模块1.1包括微控制器1.2，微控制器1.2上的常用引脚与主控接口1.3连接，主控接口1.3的输出线与环状主控端子1.4连接，微控制器1.2连接有电源1.5，主控模块1.1上分布多个相同的主控接口1.3和主控端子1.4；
主控接口1.3的输出线包括电源线V、地线G、交流信号线S和若干控制线，本实施例为两根控制线，分别为N 1和N 2，主控接口1.3上设有交流振荡电路1.6，用于生成特定频率与电压的交流电信号S i，交流振荡电路1.6的输出线与S线连接，主控接口1.3上还设有第一数字开关1.7和第二数字开关1.8，分别与线与N 1和N 2线连接，微控制器控制第一数字开关1.7和第二数字开关1.8的状态，使N 1和N 2选择性的与微控制器1.2上的特定引脚连接；
主控模块1.1通过环状主控端子1.4与外设模块1.9以任意角度相连，且主控模块1.1的某个主控端子1.4可同时并行连接多个外设模块1.9；外设模块1.9上分布多个外设端子1.110，外设端子1.110与外设接口1.111连接，外设接口1.111输出线连接外设执行器1.113；
外设接口1.111包括调压电路1.114，调压电路1.114的输出端与外设执行器1.113的电源线连接，外设接口1.111在外设端子1.110和外设执行器1.113的N 1和N 2线之间设有第三数字开关1.112；
交流信号线S输入的交流电信号S i经过带通滤波器1.115向整流器116传递正弦信号，正弦信号经过整流器1.116后被转为直流信号控制与其相连的锁存器1.117的输出状态，锁存器1.117的输出端通过非门电路与第三数字开关1.112连接，控制第三数字开关1.112的断开或闭合；
整流器1.116的输出端与N 1线之间通过数字开关1.118连接，锁存器1.117的输出端控制数字开关1.118的断开或闭合；
不同外设模块1.9上的带通滤波器1.115只能通过特定频率和电压的交流电信号S i，且互不重复。
一种如上所述的多智能体系统结构的控制方法，包括如下步骤：
步骤1、设置多智能体系统应用场景
启动运行智能体系统，通过多通道交互系统15选择海陆空联合搜救应用场景，根据选取的应用场景结合多智能体知识库系统生成多智能体系统1系统的运行条件和参数，具体包括搜救覆盖面积，系统运行时长，搜救时间段和检测目标；用户14通过多通道交互系统15输入参数，并上传到数据服务器12中；
步骤2、构建多智能体系统1系统
多通道交互系统15根据选择的应用场景结合多智能体系统1知识库系统判断多智能体系统1系统中智能体单元1的组成类型和数量，并根据智能体单元1数量和运动路径确定定位方式和通信方式；
步骤3、多智能体系统1系统组网
多智能体系统1系统通信网络系统由互连网络系统、卫星中继通信网络系统和自组织无线网络系统三部分构成；互连网络系统互连无线自组织网络和数据存储服务器；卫星中继通信网络系统联系水下无线自组织网络和互连网络的中继网络；无线自组织网络，负责多智能体系统1的网络服务，分为陆地无线自组织网络、空中无线自组织网络和水下无线自组织网络；
水下无线自组织网络通过水下无人航行器2上的通信卫星4设备上传数据和接收来自多通道交互系统15的指令；通信卫星4接收水下无线自组织网络上传的数据和多通道交互系统15的指令，并分别下传给卫星地面接收站5和水下无线自组织网络，卫星地面接收站5通过地面网络接入万维网进行数据和指令的接收和发送；陆上和空中无线自组织网络则通过现有通信网络的地面基站接入万维网完成与数据服务器12和多通道交互系统15的数据和指令交互；陆海空中的多智能体系统1通过无线自组织网络完成主控制器选择、数据传输和指令接收；
步骤4、部署运行多智能体系统1系统并感知环境信息
智能体单元1中选择一个智能体单元1作为主控智能体单元1；多通道交互系统15通过通信网络将控制指令发送到主控智能体单元1，主控智能体单元1发送指令控制其它智能体单元1运动到部署位置；
智能体单元1采集，包括图像、语音、动作、生命信号的环境信息信息，并将采集到的信息传送给智能体单元1的控制模块；
智能体单元1通过通信网络将信息传输到主控智能体单元1，由通信拓扑结 构确定信息的传输路径；
步骤5、将传输的信息进行多源异构信息融合处理
步骤6、多智能体系统1系统任务分配执行
多通道交互系统15根据目标实时定位信息，通过通信网络发送运动控制指令，智能体单元1接收到指令后运动，保证目标始终在智能体单元1的最佳监测位置；同时多通道交互系统15显示目标位置信息；
步骤7、用户14控制操作
用户14在多通道交互系统15上实时查看目标的位置信息，通过多通道交互系统15操作控制智能体单元1运行，并发出动作、手势、语音指令，由多通道交互系统15或智能体单元1感知后发出运行指令控制智能体单元1执行。
进一步的，所述步骤(2)中构建多智能体系统1系统的具体步骤如下：
多通道交互系统15根据选择的应用场景结合多智能体知识库系统判断多智能体系统中智能体单元1的组成类型，海陆空联合搜救应用场景下的智能体单元1包括无人车9、无人机7和水下无人航行器2；
多通道交互系统15根据搜救时间段来确定检测方式，若工作时间段为白天，则选择图像或生命感知模块进行检测，若搜救时间段为夜晚，则选择红外或生命感知模块进行检测；检测方式确定后，控制系统根据搜救覆盖面积和单个智能体单元1的检测范围，确定智能体单元1的数量，同时生成智能体单元1的运动路径；
智能体单元1数量和运动路径确定后，根据运行过程中各个智能体单元1之间的距离D来确定各个智能体单元1之间的通信方式：陆地和空中场景下，当D>500m，可采用5G、4G、GPRS、通信卫星4、短波通信等；当D<500m，可采用Zig-Bee、蓝牙(Bluetooth)、无线宽带(Wi-Fi)等；水下场景则采用水声通信；
多通道交互系统15根据应用场景和定位精度需求，确定采用何种定位方式，或多种定位方式组合使用：在开阔场地，定位精度为m级，则采用GPS定位；在室内，定位精度为cm级，则采用WIFI或蓝牙；在水下，定位精度为cm级：则采用超声波定位；
多通道交互系统15根据智能体单元1的续航能力和运行时长，判断多智能体系统需要配置移动电站的数量。
所述步骤(3)中多智能体系统组网的方法包括以下步骤：
(3.1)多智能体系统组网
多智能体系统依据组网协议组成无明显主从关系的网络，即每个网络节点都做为主节点和发起网络传输请求，根据任务需求通过软件实现节点的主从分配；
(3.2)建立路由表
每个多智能体系统依赖泛洪协议向外广播数据包，主要包括智能体单元的地址、机器码，每个智能体单元根据接收到的数据包，识别出其它智能体单元的地址和机器码，并为其建立路由表，随机分配一个智能体单元为主控制器；
(3.3)主控制节点接入主干网络
主控智能体单元通过网关节点和通信网络与上位机通信；
(3.4)上位机接收信息
若上位机有新的分配多智能体系统主控制器的算法，则重新指定主控制器智能体单元，重新分配主从控制节点，重复步骤(3.3)，若无，则网络处于等待响应指令阶段；
(3.5)上位机发出任务指令
上位机通过互联通信网络传送任务指令；任务指令来源于上位机提取的用户指令，或根据数据服务器中知识库生成的任务指令；
(3.6)通信网络逐层响应
通信网络将上位机发出的任务指令逐层发送到各个智能体单元；
(3.7)智能体单元响应
各个智能体单元通过陆海空无线自组织网络将相应的指令响应或者数据传送回数据服务器或者上位机。

Claims (10)

  1. 一种多智能体系统结构，其特征在于：包括自组织无线通信网络，所述智能体单元和无线通信网络传输装置与互联网连接，所述互联网分别又与网关设备和第一通信网络基站连接，所述第一通信网络还与多通道交互系统连接，所述多通道交互系统与用户连接；所述网关设备还与数据服务器连接。
  2. 根据权利要求1所述的一种多智能体系统结构，其特征在于：所述智能体单元包括主控模块和与之连接的多个外设模块，所述主控模块通过接口与外设模块连接；
    所述各个模块之间通过接口连接进行信号和能量传递；
    所述主控模块包括微控制器，微控制器上的常用引脚与主控接口连接，主控接口的输出线与环状主控端子连接，微控制器还连接有电源，主控模块上分布多个相同的主控接口和主控端子。
  3. 根据权利要求2所述的一种多智能体系统结构，其特征在于：所述外设模块包括感知模块、通信模块、定位模块和执行模块；
    所述感知模块通过图像、语音、动作、生命信号进行感知；
    所述通信模块通过5G、4G、GPRS、CDMA、卫星通信、水声通信、Zig-Bee、蓝牙、Wi-Fi方式进行通信；
    所述定位模块通过GPS、Wi-Fi、蓝牙、超声波的定位方式进行定位。
  4. 根据权利要求2所述的一种多智能体系统结构，其特征在于：所述主控接口上设有交流振荡电路、第一数字开关和第二数字开关，
    所述交流振荡电路用于生成特定频率与电压的交流电信号S i，交流振荡电路的输出线与S线连接；所述主控接口通过N 1线与第一数字开关连接，所述主控接口通过N 1线与第二数字开关。
  5. 根据权利要求2所述的一种多智能体系统结构，其特征在于：主控模块通过环状主控端子与外设模块以任意角度相连，且主控模块的某个主控端子可同时并行连接多个外设模块；外设模块上分布多个外设端子，外设端子与外设接口连接，外设接口输出线连接外设执行器。
  6. 根据权利要求2所述的一种多智能体系统结构，其特征在于：所述外设接口内包括调压电路、带通滤波器、整流器、锁存器；
    所述调压电路的输出端与外设执行器的电源线连接，外设接口在外设端子和 外设执行器的N 1和N 2线之间设有第三数字开关；
    交流信号线S输入的交流电信号S i经过带通滤波器向整流器传递正弦信号，正弦信号经过整流器后被转为直流信号控制与其相连的锁存器的输出状态，锁存器的输出端通过非门电路与第三数字开关连接，控制第三数字开关的断开或闭合；
    整流器的输出端与N 1线之间通过第四数字开关连接，锁存器的输出端控制第四数字开关的断开或闭合。
  7. 根据权利要求1所述的一种多智能体系统结构，其特征在于：所述自组织无线通信网络分为水下自组织无线通信网络、空中自组织无线通信网络和陆地自组织无线通信网络；对应的，智能体单元也分为水下无人航行器、无人机和无人车。
  8. 根据权利要求3所述的一种多智能体系统结构，其特征在于：所述水下自组织无线通信网络由水下无人航行器、水上中转站、通信卫星和卫星地面接收站构成，所述水上中转站与若干个水下无人航行器连接，所述水下无人航行器通过通信卫星与卫星地面接收站连接；
    所述空中自组织无线通信网络包括无人机和第二通信网络基站，所述第二通信网络基站与若干个无人机连接；
    所述陆地自组织无线通信网络包括无人车和第三通信网络基站，所述第三通信网络基站与若干个无人车连接。
  9. 一种如权利要求1-8之一所述的多智能体系统结构的控制方法，其特征在于，包括如下步骤：
    (1)设置多智能体系统应用场景：
    启动运行多智能体系统，通过多通道交互系统选择海陆空联合搜救应用场景，根据选取的应用场景结合多智能体知识库系统生成多智能体系统的运行条件和参数，具体包括搜救覆盖面积，系统运行时长，搜救时间段和检测目标；用户通过多通道交互系统输入参数，并上传到数据服务器中；
    (2)构建多智能体系统：
    多通道交互系统根据选择的应用场景结合多智能体知识库系统判断多智能体系统中智能体单元的组成类型和数量，并根据智能体单元数量和运动路径确定定位方式和通信方式；
    (3)构建多智能体系统组网；
    (4)部署运行多智能体系统并感知环境信息：
    智能体单元中选择一个智能体单元作为主控智能体单元；多通道交互系统通过通信网络将控制指令发送到主控智能体单元，主控智能体单元发送指令控制其它智能体单元运动到部署位置；
    智能体单元采集，包括图像、语音、动作、生命信号的环境信息信息，并将采集到的信息传送给智能体单元的控制模块；
    智能体单元通过通信网络将信息传输到主控智能体单元，由通信拓扑结构确定信息的传输路径；
    (5)将传输的信息进行多源异构信息融合处理；
    (6)多智能体系统任务分配执行：
    多通道交互系统根据目标实时定位信息，通过通信网络发送运动控制指令，智能体单元接收到指令后运动，保证目标始终在智能体单元的最佳监测位置；同时多通道交互系统显示目标位置信息；
    (7)用户控制操作：
    用户在多通道交互系统上实时查看目标的位置信息，通过多通道交互系统操作控制智能体单元运行，并发出动作、手势、语音指令，由多通道交互系统或智能体单元感知后发出运行指令控制智能体单元执行。
  10. 根据权利要求9所述的一种多智能体系统结构的控制方法，其特征在于，所述步骤(3)中构建多智能体系统组网的具体步骤如下：
    多智能体系统通信网络系统由互连网络系统、卫星中继通信网络系统和自组织无线网络系统三部分构成；互连网络系统互连无线自组织网络和数据存储服务器；卫星中继通信网络系统联系水下无线自组织网络和互连网络的中继网络；无线自组织网络，负责多智能体系统的网络服务，分为陆地无线自组织网络、空中无线自组织网络和水下无线自组织网络；
    水下无线自组织网络通过水下无人航行器上的通信卫星设备上传数据和接收来自多通道交互系统的指令；通信卫星接收水下无线自组织网络上传的数据和多通道交互系统的指令，并分别下传给卫星地面接收站和水下无线自组织网络，卫星地面接收站通过地面网络接入万维网进行数据和指令的接收和发送；陆上和 空中无线自组织网络则通过现有通信网络的地面基站接入万维网完成与数据服务器和多通道交互系统的数据和指令交互；陆海空中的多智能体系统通过无线自组织网络完成主控制器选择、数据传输和指令接收。
PCT/CN2021/076258 2020-06-30 2021-02-09 一种多智能体系统结构及其控制方法 WO2022001119A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202010611458.6 2020-06-30
CN202010611458.6A CN111836409B (zh) 2020-06-30 2020-06-30 一种多智能体系统结构及其控制方法

Publications (1)

Publication Number Publication Date
WO2022001119A1 true WO2022001119A1 (zh) 2022-01-06

Family

ID=72900637

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2021/076258 WO2022001119A1 (zh) 2020-06-30 2021-02-09 一种多智能体系统结构及其控制方法

Country Status (2)

Country Link
CN (1) CN111836409B (zh)
WO (1) WO2022001119A1 (zh)

Cited By (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114581748A (zh) * 2022-05-06 2022-06-03 南京大学 基于机器学习的多智能体感知融合系统及其实现方法
CN116260882A (zh) * 2023-05-15 2023-06-13 中国人民解放军国防科技大学 一种低通信流量的多智能体调度异步一致性方法和装置
CN116353861A (zh) * 2022-09-23 2023-06-30 武汉理工大学 察打补一体且可分体、回收、补给的跨介质无人平台及系统
CN117289668A (zh) * 2023-11-24 2023-12-26 深圳市陶氏精密技术有限公司 分布式减速机网络协同控制方法、装置、设备及存储介质
CN117320218A (zh) * 2023-11-28 2023-12-29 杭州亿时照明工程设计有限公司 智慧照明系统的舒适度一致性控制方法

Families Citing this family (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111836409B (zh) * 2020-06-30 2023-06-09 镇江宇诚智能装备科技有限责任公司 一种多智能体系统结构及其控制方法
CN111835838A (zh) * 2020-06-30 2020-10-27 江苏科技大学 一种多智能体系统及其控制方法
CN116014740B (zh) * 2023-03-22 2024-03-01 国网浙江义乌市供电有限公司 一种配电网多元资源能源聚合控制器最少化部署方法

Citations (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104283935A (zh) * 2013-07-09 2015-01-14 上海海事大学 一种海洋互联网系统及其方法
US20190223237A1 (en) * 2018-01-18 2019-07-18 Electronics And Telecommunications Research Institute Unmanned vehicle controlling system and method of operating same
CN110166265A (zh) * 2018-02-11 2019-08-23 陕西爱尚物联科技有限公司 一种网络管控的方法及其模块
CN110944032A (zh) * 2019-10-14 2020-03-31 国网山东省电力公司应急管理中心 一种基于泛在电力物联网的自组网综合感知智能识别预警方法
CN111198575A (zh) * 2020-02-27 2020-05-26 西北工业大学 一种无人机飞行控制器
CN111781871A (zh) * 2020-06-30 2020-10-16 镇江宇诚智能装备科技有限责任公司 一种智能体结构及其多外设模块拼接与识别方法
CN111836409A (zh) * 2020-06-30 2020-10-27 镇江宇诚智能装备科技有限责任公司 一种多智能体系统结构及其控制方法
CN111835838A (zh) * 2020-06-30 2020-10-27 江苏科技大学 一种多智能体系统及其控制方法

Family Cites Families (13)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US8753164B2 (en) * 2007-10-11 2014-06-17 Lego A/S Toy construction system
US20100074116A1 (en) * 2008-09-25 2010-03-25 Wayne-Dalton Corp. System and Method of Controlling a Wireless Radio-Frequency Network Using a Gateway Device
US20100269143A1 (en) * 2009-04-21 2010-10-21 Irving Rabowsky System and Method for Satellite Enhanced Command, Control, and Surveillance Services Between Network Management Centers and Unmanned Land and Aerial Devices
CN102739786B (zh) * 2012-06-22 2013-04-24 渤海大学 基于泛在网络的建筑火灾智能救助系统及方法
EP2937997B1 (en) * 2013-06-25 2018-11-28 Fuji Electric Co., Ltd. Signal transmission circuit
CN205752715U (zh) * 2016-03-31 2016-11-30 深圳贝尔创意科教有限公司 连接结构及应用该连接结构的电子装置
CN107547587A (zh) * 2016-06-24 2018-01-05 南京中兴软件有限责任公司 一种定位方法及装置
KR20180056326A (ko) * 2016-11-18 2018-05-28 주식회사 승우 비행체를 이용한 통신서비스 시스템
US10986602B2 (en) * 2018-02-09 2021-04-20 Intel Corporation Technologies to authorize user equipment use of local area data network features and control the size of local area data network information in access and mobility management function
CN108650010A (zh) * 2018-03-26 2018-10-12 西南电子技术研究所（中国电子科技集团公司第十研究所） 智能测控通信网络系统
CN111292523B (zh) * 2018-12-06 2023-04-07 中国信息通信科技集团有限公司 网络智能体系统
CN110111210B (zh) * 2019-05-16 2021-04-23 刘锋 一种混合智能社交网络系统、社交方法及设备
CN110390431A (zh) * 2019-07-19 2019-10-29 大连海事大学 一种基于无人设备群体智能算法的搜救网及其调度方法

Patent Citations (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104283935A (zh) * 2013-07-09 2015-01-14 上海海事大学 一种海洋互联网系统及其方法
US20190223237A1 (en) * 2018-01-18 2019-07-18 Electronics And Telecommunications Research Institute Unmanned vehicle controlling system and method of operating same
CN110166265A (zh) * 2018-02-11 2019-08-23 陕西爱尚物联科技有限公司 一种网络管控的方法及其模块
CN110944032A (zh) * 2019-10-14 2020-03-31 国网山东省电力公司应急管理中心 一种基于泛在电力物联网的自组网综合感知智能识别预警方法
CN111198575A (zh) * 2020-02-27 2020-05-26 西北工业大学 一种无人机飞行控制器
CN111781871A (zh) * 2020-06-30 2020-10-16 镇江宇诚智能装备科技有限责任公司 一种智能体结构及其多外设模块拼接与识别方法
CN111836409A (zh) * 2020-06-30 2020-10-27 镇江宇诚智能装备科技有限责任公司 一种多智能体系统结构及其控制方法
CN111835838A (zh) * 2020-06-30 2020-10-27 江苏科技大学 一种多智能体系统及其控制方法

Cited By (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114581748A (zh) * 2022-05-06 2022-06-03 南京大学 基于机器学习的多智能体感知融合系统及其实现方法
CN114581748B (zh) * 2022-05-06 2022-09-23 南京大学 基于机器学习的多智能体感知融合系统及其实现方法
CN116353861A (zh) * 2022-09-23 2023-06-30 武汉理工大学 察打补一体且可分体、回收、补给的跨介质无人平台及系统
CN116260882A (zh) * 2023-05-15 2023-06-13 中国人民解放军国防科技大学 一种低通信流量的多智能体调度异步一致性方法和装置
CN117289668A (zh) * 2023-11-24 2023-12-26 深圳市陶氏精密技术有限公司 分布式减速机网络协同控制方法、装置、设备及存储介质
CN117289668B (zh) * 2023-11-24 2024-02-02 深圳市陶氏精密技术有限公司 分布式减速机网络协同控制方法、装置、设备及存储介质
CN117320218A (zh) * 2023-11-28 2023-12-29 杭州亿时照明工程设计有限公司 智慧照明系统的舒适度一致性控制方法
CN117320218B (zh) * 2023-11-28 2024-02-20 杭州亿时照明工程设计有限公司 智慧照明系统的舒适度一致性控制方法

Also Published As

Publication number Publication date
CN111836409B (zh) 2023-06-09
CN111836409A (zh) 2020-10-27

Similar Documents

Publication Publication Date Title
WO2022001119A1 (zh) 一种多智能体系统结构及其控制方法
WO2022001120A1 (zh) 一种多智能体系统及其控制方法
Alsamhi et al. Survey on artificial intelligence based techniques for emerging robotic communication
Erdelj et al. UAVs that fly forever: Uninterrupted structural inspection through automatic UAV replacement
WO2022021739A1 (zh) 一种语义智能变电站机器人仿人巡视作业方法及系统
CN102707675B (zh) 群机器人控制器、群机器人控制方法及控制器终端
CN103389699B (zh) 基于分布式智能监测控制节点的机器人监控及自主移动系统的运行方法
CN111766879A (zh) 一种基于自主协同导航的智能车编队系统
Curiac Towards wireless sensor, actuator and robot networks: Conceptual framework, challenges and perspectives
WO2019041874A1 (zh) 飞行器控制方法及装置
CN113282083B (zh) 一种基于机器人操作系统的无人车编队实验平台
Smith et al. Multi-robot deployment and coordination with embedded graph grammars
CN110045750A (zh) 一种基于四旋翼无人机的室内场景构建系统及其实现方法
Chen et al. A fast coordination approach for large-scale drone swarm
CN112731914A (zh) 一种5g智慧工厂的云化agv应用系统
Pashkin et al. Pilot communication protocols for group of mobile robots in USAR scenarios
CN114793239A (zh) 一种内河智能船舶域控制器功能实现的系统及方法
Pan et al. The application of the multi-agent coverage and self-healing control based on a swarm intelligence SONM and potential function approach
CN110614639A (zh) 一种基于ros的变电站巡检机器人系统及其方法
CN205899376U (zh) 一种自主式球形两栖多机器人通信系统
Al-Ali et al. Description of EU Projects Demonstrate Self-Adaptive CPSs (V1. 0)
Rasheed et al. ROBO-FLOCK: Development of a low-cost Leader-follower Swarm of Mobile Robots
Bauer et al. Persistent UAV Formation Flight by Dynamic Agent Replacement and Leader Selection
Gunawardhana et al. Indoor Autonomous Multi-Robot Communication System
CN112168072A (zh) 一种清洁机器人的软件结构

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 21832675

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 21832675

Country of ref document: EP

Kind code of ref document: A1