A system and method for jointly allocating communication and computing resources for aircraft wing handling

Abstract

The present invention belongs to the technical field of aircraft wing handling methods, and specifically relates to a system and method for jointly allocating communication and computing resources for aircraft wing handling. The resource joint allocation system includes an edge server, a communication module, and a transport vehicle. The edge server is connected to multiple transport vehicles via the communication module. The resource allocation unit can effectively achieve an adaptive balance between computing load and control accuracy, and further dynamically selects a control algorithm for the control unit based on processor utilization and the remaining time of the control cycle. The optimization system of the present invention constructs a communication-control-computation-perception fusion architecture. Through reinforcement learning algorithms and control algorithms, the perception unit, control unit, resource allocation unit, and communication module are dynamically selected for joint control and resource allocation. This achieves the optimized allocation of various process flows and subsystem resources of the on-site aircraft wing transport vehicle, reduces performance bottlenecks under limited resource conditions, and effectively improves handling efficiency.

Description

Technical Field

[0001] The present invention belongs to the technical field of aircraft wing transportation methods, and in particular relates to a communication and computing resource joint allocation system and method for aircraft wing transportation. Background Art

[0002] In aircraft assembly and manufacturing, automated guided vehicle (AGV) systems are widely used for the precise handling of large components, such as aircraft wings, due to their flexibility and automation advantages. Due to the large size and extreme weight of aircraft wings, as well as the stringent assembly precision requirements, multiple AGVs are typically required to coordinate and distribute the load while maintaining synchronized movement.

[0003] However, existing technologies still have significant defects in the collaborative control of multiple AGVs, especially in communication coordination and joint allocation of computing resources. The collaborative handling of multiple AGVs requires real-time solution of multi-objective optimization problems such as kinematic constraints, load balancing and obstacle avoidance paths, which places extremely high demands on computing resources. Existing technologies usually adopt fixed priority scheduling or static task partitioning strategies, which are unable to dynamically adjust computing resources according to real-time working conditions. In addition, existing resource allocation models often regard communication, computing and mechanical control as independent modules, ignoring the coupling relationship between the three. For example, when multiple AGVs need to adjust the height of the lifting mechanism synchronously, if the computing resources are overly concentrated in the path planning module, it may cause a lag in the generation of motion control instructions, thereby causing wing attitude oscillation. The current multi-AGV system has bottlenecks in communication real-time, flexible scheduling of computing resources and cross-module collaboration, which seriously restricts the efficiency and safety of handling large components such as aircraft wings. Summary of the Invention

[0004] The technical problem to be solved by the present invention is to overcome the defect in the prior art that aircraft wing handling control cannot achieve effective communication and resource allocation, resulting in insufficient handling efficiency and safety, and to provide a communication and computing resource joint allocation system and method for aircraft wing handling.

[0005] The present invention discloses a joint distribution system of communication and computing resources for aircraft wing transportation, comprising an edge server, a communication module and a transport vehicle, wherein the edge server is communicatively connected with a plurality of the transport vehicles via the communication module;

[0006] The edge server includes a sensing unit, a control unit and a resource allocation unit, wherein the sensing unit, the control unit and the resource allocation unit include a plurality of virtualized containers;

[0007] The control unit presets a control algorithm for forming a control strategy based on the perception data, wherein the control algorithm includes a low complexity algorithm, a medium complexity algorithm, and a high complexity algorithm, and the computing power consumed during operation increases in sequence;

[0008] The resource allocation unit is configured to allocate the communication resources of the communication module and the computing resources of the control unit based on the resource utilization of the communication module and the edge server through a reinforcement learning algorithm; the resource allocation unit is configured to select a control algorithm for the control unit based on the processor utilization of the edge server and the remaining time of the control cycle, and to control the transmission frequency and data packet size of the corresponding data sent by the communication module;

[0009] The transport vehicle includes an actuator and a sensing mechanism; the sensing mechanism is used to obtain sensing data; the actuator is used to control the transport vehicle based on the control strategy;

[0010] The communication module performs network slicing on the transmitted data, and includes an uplink module and a downlink module. The uplink module is used to transmit the perception data of the perception mechanism of the transport vehicle to the perception unit of the edge server; the downlink module is used to transmit the control strategy of the control unit of the edge server to the execution mechanism of the transport vehicle.

[0011] Furthermore, the perception unit includes a data fusion container, a mechanical analysis container and an anomaly detection container; the data fusion container is used to integrate the perception data of the lidar, inertial measurement unit and odometer to form the positioning data of the transport vehicle; the mechanical analysis container is used to evaluate the force of the wing workpiece based on the force data; the anomaly detection container is used to monitor the jump of the perception data; the control unit includes a scanning and mapping container, a formation control container and a planning and scheduling container, which are used to generate a control strategy based on the output of the perception unit.

[0012] Furthermore, the sensing mechanism includes a laser radar, an inertial measurement unit, an odometer and a force sensor, and the force sensor is used to monitor the force data of the transport vehicle and the wing workpiece; the actuator includes a chassis actuator and an upper body actuator, which are used to control the transport vehicle based on the control strategy.

[0013] Furthermore, the resource allocation unit is used to allocate initial computing power to the scanning and mapping container, formation control container and planning and scheduling container of the control unit, the scanning and mapping container allocates 100~500 TOPS computing power, the formation control container allocates 10~100 TOPS computing power, and the planning and scheduling container allocates 500~1000 TOPS computing power; allocates an initial network slicing strategy to the downlink module of the communication module, allocates eMBB slices to the control strategy output by the scanning and mapping container, allocates uRLLC slices to the formation control container, and allocates mMTC slices to the planning and scheduling container.

[0014] Furthermore, the reinforcement learning algorithm of the resource allocation unit is used to: define a state space, which includes process type, communication system status, computing system status, and task requirement matrix; define an action space, which includes the communication resource allocation strategy of the communication module and the computing resource allocation strategy of the edge server; define a reward function, the reward value of the reward function includes the satisfaction of task requirements, resource utilization and delay penalty; train the resource allocation strategy based on the PPO reinforcement learning algorithm, select the optimal action based on the current state, and control the execution of the communication resource allocation strategy and the computing resource allocation strategy; update the state space, and repeatedly generate the communication resource allocation strategy and the computing resource allocation strategy.

[0015] Furthermore, the task requirement matrix is ​​[Q, C, T, P, S], where Q is the QoS requirement, C is the computing power requirement, T is the control period, P is the control accuracy, and S is the perception fusion dimension.

[0016] Furthermore, the control algorithms contained in the scanning and mapping container, formation control container, and planning and scheduling container of the control unit include low-complexity, medium-complexity, and high-complexity algorithms, respectively; in the corresponding containers and their control cycles, when the utilization rate of the allocated computing resources is greater than 80% or the remaining control cycle time is less than 20%, switch to the low-complexity algorithm; when the utilization rate of the allocated computing resources is greater than 50% and less than or equal to 80% and the remaining control cycle time is greater than 50%, switch to the medium-complexity algorithm; when the utilization rate of the allocated computing resources is less than 50% and the remaining control cycle time is greater than 80%, switch to the high-complexity algorithm.

[0017] Furthermore, the low-complexity algorithm is a PID control algorithm, and the computing power consumption is less than 10TOPS; the medium-complexity algorithm is an LQR control algorithm, and the computing power consumption is 10~50TOPS; the high-complexity algorithm is an MPC control algorithm, and the computing power consumption is greater than 50TOPS.

[0018] A method for jointly allocating communication and computing resources for aircraft wing transportation is applicable to the above-mentioned joint allocation system for communication and computing resources for aircraft wing transportation. The method comprises: a resource allocation unit allocates communication resources of a communication module and computing resources of a control unit based on the resource utilization of the communication module and an edge server through a reinforcement learning algorithm; the resource allocation unit selects a control algorithm for the control unit based on the processor utilization of the edge server and the remaining time of the control cycle, and controls the sending frequency and data packet size of corresponding data sent by the communication module; a sensing mechanism of a transport vehicle acquires sensing data; an actuator controls the transport vehicle based on a control strategy; the communication module performs network slicing on the transmitted data, and an uplink module transmits the sensing data of the sensing mechanism of the transport vehicle to the sensing unit of the edge server; and a downlink module transmits the control strategy of the control unit of the edge server to the actuator of the transport vehicle.

[0019] A computer-readable storage medium stores a computer program thereon, which, when executed by a processor, implements the above-mentioned method for jointly allocating communication and computing resources for aircraft wing handling.

[0020] Beneficial effects: The present invention discloses a joint allocation system for communication and computing resources for aircraft wing transportation, comprising an edge server, a communication module, and a transport vehicle, wherein the edge server is communicatively connected to a plurality of transport vehicles via the communication module; a resource allocation unit is used to allocate the communication resources of the communication module and the computing resources of the control unit based on the resource utilization of the communication module and the edge server through a reinforcement learning algorithm; the resource allocation unit is used to select a control algorithm for the control unit based on the processor utilization and the remaining time of the control cycle of the edge server, and to control the transmission frequency and data packet size of the corresponding data sent by the communication module. The resource allocation unit can consider the differences in the communication and computing resource requirements of each process in the aircraft wing transportation through the reinforcement learning algorithm, effectively and dynamically allocate the communication and computing resources of each process in the aircraft wing transportation, achieve an adaptive balance between computing load and control accuracy, and further dynamically select a control algorithm for the control unit based on the processor utilization and the remaining time of the control cycle, thereby achieving more efficient resource allocation. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0022] Figure 1 Schematic diagram of the system structure of the present invention;

[0023] Figure 2 Schematic diagram of the resource allocation method of the present invention. DETAILED DESCRIPTION

[0024] To make the above-mentioned objects, features, and advantages of the present application more clearly understood, the specific embodiments of the present application are described in detail below with reference to the accompanying drawings. The following description sets forth many specific details to facilitate a full understanding of the present application. However, the present application can be implemented in many other ways than those described herein, and those skilled in the art can make similar improvements without violating the scope of the present application. Therefore, the present application is not limited to the specific embodiments disclosed below.

[0025] In the description of this application, it should be understood that, unless otherwise expressly specified or limited, the terms "installed," "connected," "connected," "fixed," and the like should be understood in a broad sense. For example, they may refer to fixed connections, detachable connections, or integration; mechanical connections, electrical connections; direct connections, or indirect connections through an intermediate medium; internal connections between two elements, or interactions between two elements, unless otherwise expressly limited. A person of ordinary skill in the art can understand the specific meanings of the above terms in this application based on the specific circumstances.

[0026] Example 1:

[0027] Reference Figure 1 As shown, the present invention discloses a joint distribution system of communication and computing resources for aircraft wing transportation, including an edge computing server (ECS), a communication module and a transportation vehicle, wherein the edge server is communicatively connected to a plurality of the transportation vehicles through the communication module;

[0028] The edge server includes a perception unit, a control unit and a resource allocation unit, and the perception unit, control unit and resource allocation unit include multiple virtualized containers; in this embodiment, the virtualized container is a Docker container, which is orchestrated and managed by the container orchestration platform Kubernetes, and computing resources are allocated by the container orchestration platform Kubernetes.

[0029] The control unit presets a control algorithm for forming a control strategy based on the perception data, wherein the control algorithm includes a low complexity algorithm, a medium complexity algorithm, and a high complexity algorithm, and the computing power consumed during operation increases in sequence;

[0030] The resource allocation unit is configured to allocate the communication resources of the communication module and the computing resources of the control unit based on the resource utilization of the communication module and the edge server through a reinforcement learning algorithm; the resource allocation unit is configured to select a control algorithm for the control unit based on the processor utilization of the edge server and the remaining time of the control cycle, and to control the transmission frequency and data packet size of the corresponding data sent by the communication module;

[0031] The transport vehicle includes an actuator and a sensing mechanism; the sensing mechanism is used to obtain sensing data; the actuator is used to control the transport vehicle based on the control strategy;

[0032] The communication module performs network slicing on the transmitted data, and includes an uplink module and a downlink module. The uplink module is used to transmit the perception data of the perception mechanism of the transport vehicle to the perception unit of the edge server; the downlink module is used to transmit the control strategy of the control unit of the edge server to the execution mechanism of the transport vehicle.

[0033] Specifically, the perception unit includes a data fusion container, a mechanical analysis container, and an anomaly detection container; the data fusion container is used to integrate the perception data of the lidar, inertial measurement unit, and odometer to form the positioning data of the transport vehicle; the mechanical analysis container is used to evaluate the force of the wing workpiece based on the force data; the anomaly detection container is used to monitor the jump of the perception data; the control unit includes a scanning and mapping container, a formation control container, and a planning and scheduling container, which are used to generate a control strategy based on the output of the perception unit.

[0034] The sensing mechanism includes a laser radar, an inertial measurement unit, an odometer and a force sensor, and the force sensor is used to monitor the force data of the transport vehicle and the wing workpiece; the actuator includes a chassis actuator and an upper body actuator, which are used to control the transport vehicle based on the control strategy.

[0035] In this embodiment, the resource allocation unit can consider the differences in communication and computing resource requirements of each process in the aircraft wing transportation through a reinforcement learning algorithm, effectively and dynamically allocate communication and computing resources for each process in the aircraft wing transportation, achieve an adaptive balance between computing load and control accuracy, and further dynamically select a control algorithm for the control unit based on processor utilization and the remaining time of the control cycle to achieve more efficient resource allocation.

[0036] Example 2:

[0037] Reference Figure 2As shown, this embodiment provides a method for jointly allocating communication and computing resources for aircraft wing transportation, which is applicable to the joint allocation system for communication and computing resources for aircraft wing transportation described in Example 1, and the method includes: a resource allocation unit allocates communication resources of the communication module and computing resources of the control unit based on the resource utilization of the communication module and the edge server through a reinforcement learning algorithm; the resource allocation unit selects a control algorithm for the control unit based on the processor utilization of the edge server and the remaining time of the control cycle, and controls the sending frequency and data packet size of the corresponding data sent by the communication module; the perception mechanism of the transport vehicle obtains perception data; the execution mechanism controls the transport vehicle based on the control strategy; the communication module performs network slicing on the transmitted data, and the uplink module transmits the perception data of the perception mechanism of the transport vehicle to the perception unit of the edge server; the downlink module transmits the control strategy of the control unit of the edge server to the execution mechanism of the transport vehicle.

[0038] In this embodiment, when the joint allocation method of communication and computing resources transported by the aircraft wing is turned on, the resource allocation unit allocates initial computing power to the scanning and mapping container, formation control container and planning and scheduling container of the control unit, the scanning and mapping container allocates 100~500 TOPS computing power, the formation control container allocates 10~100 TOPS computing power, and the planning and scheduling container allocates 500~1000 TOPS computing power; the initial network slicing strategy is allocated to the downlink module of the communication module, the eMBB slice is allocated for the control policy output by the scanning and mapping container, the uRLLC slice is allocated for the formation control container, and the mMTC slice is allocated for the planning and scheduling container.

[0039] Specifically, a field-side demand matrix is ​​established for each process. The field-side demand matrix is ​​[QoS requirements, computing power density, control cycle, control accuracy, and perception data fusion dimension]. Specifically, the demand dimensions of the demand matrix are shown in Table 1:

[0040] Table 1

[0041] Demand Dimension Parameter Symbol Parameter Description QoS requirements Q Comprehensive communication quality requirements (including bandwidth / latency / reliability) Hashrate density C Computing power required per unit time (TOPS) Control cycle T Minimum execution interval of control instructions Control accuracy P Positioning / force accuracy of actuators Perception fusion dimension S Fusion depth of multimodal perception data

[0042] The specific expression of QoS requirements in the requirement dimension of the requirement matrix is:

[0043] Table 2

[0044] grade Parameter Description Value range 1 Low bandwidth, high number of connections, and high latency tolerance Bandwidth ≤ 10Mbps, latency ≤ 1s, number of connections ≥ 10,000 2 Medium bandwidth, medium number of connections, medium latency tolerance Bandwidth ≤ 50Mbps, latency ≤ 500ms, number of connections ≥ 1000 3 High bandwidth, medium latency, and medium reliability Bandwidth ≥ 100Mbps, latency ≤ 50ms, 99.99% reliability 4 Ultra-high bandwidth, low latency, and high reliability Bandwidth ≥ 500Mbps, latency ≤ 20ms, 99.999% reliability 5 Ultra-low latency, ultra-high reliability, medium bandwidth Bandwidth ≥ 50Mbps, latency ≤ 5ms, 99.9999% reliability

[0045] Specifically, QoS requirements are based on network slicing characteristics (mMTC: 1-2, eMBB: 3-4, uRLLC: 5), corresponding to different priorities of bandwidth, latency, and reliability.

[0046] The specific expression of computing power demand in the demand dimension of the demand matrix is:

[0047] Table 3

[0048] interval Parameter Description Value range Low computing power Simple computing tasks, low-power devices 1~10 TOPS Medium computing power Real-time control or lightweight algorithms 10~100 TOPS High computing power Complex real-time calculation, multi-sensor fusion 100~500 TOPS Ultra-high computing power Large-scale optimization, deep learning inference 500~1000 TOPS

[0049] Specifically, the computing power needs to reflect different demands for processing power per unit time.

[0050] The specific expression of the control cycle demand in the demand dimension of the demand matrix is:

[0051] Table 4

[0052] interval Parameter Description Value range Ultra-high frequency Extremely short cycle, ultra-fast response 10μs~100μs high frequency Short cycle, fast control 100μs~10ms intermediate frequency Medium-duration, coordination tasks 10ms~100ms low frequency Long cycle, planning or scheduling 100ms~1s

[0053] Specifically, the control cycle requirements reflect different real-time requirements of tasks.

[0054] The specific expression of the control accuracy requirement in the requirement dimension of the requirement matrix is:

[0055] Table 5

[0056] interval Parameter Description Value range Ultra-high precision Extremely high positioning or force requirements 0.001%~0.01% High precision High-precision tasks, industrial-grade requirements 0.01%~0.1% medium accuracy Routine tasks, centimeter-level accuracy 0.1%~1% Low precision Coarse positioning, meter-level accuracy 1%~5%

[0057] Specifically, the control accuracy requirement reflects the degree of precision required of the actuator.

[0058] The specific expression of the perception fusion dimension requirement in the requirement dimension of the requirement matrix is:

[0059] Table 6

[0060] grade Parameter Description Value range 1 Raw data processing, no fusion 1 2 Feature-level fusion to extract key information 2 3 Decision-level fusion, multi-modal comprehensive decision-making 3

[0061] Specifically, the perception fusion dimension requirement reflects the fusion depth of perception data, from raw data to high-level decision making.

[0062] Establish initial on-site requirements for each process and allocate resources accordingly:

[0063] Table 7

[0064] Process Demand Matrix Resource Allocation in Converged Systems Joint scanning and mapping [3 High computing power medium frequency medium precision 3] eMBB slice, 100-500 TOPS Formation control and handling [5 Medium Hashrate, High Frequency, Ultra-High Precision 2] uRLLC slices, 10-100 TOPS Planning and Scheduling [2 Ultra-high computing power, low frequency and low precision 1] mMTC slicing, 500-1000 TOPS

[0065] In this embodiment, the reinforcement learning algorithm of the resource allocation unit includes:

[0066] Step S1: Define a state space, which includes process type, communication system state, computing system state, and task requirement matrix. Specifically, in this embodiment, the communication system state includes bandwidth utilization, latency, and packet loss rate, and the computing system state includes CPU utilization, GPU utilization, and memory occupancy. The task requirement matrix is ​​[Q, C, T, P, S], where Q represents QoS requirements, C represents computing power, T represents control period, P represents control accuracy, and S represents the perception fusion dimension.

[0067] Step S2: Define an action space, where the action space includes the communication resource allocation strategy of the communication module and the computing resource allocation strategy of the edge server; wherein the communication resource allocation strategy includes bandwidth allocation and network slice switching, and the computing resource allocation strategy includes CPU frequency, GPU computing power, and memory allocation.

[0068] Step S3: defining a reward function, wherein the reward value of the reward function includes the satisfaction of task requirements, resource utilization and delay penalty;

[0069] The reward function is expressed as:

[0070] ;

[0071] in, Indicates the satisfaction of task requirements, ranging from ; Indicates resource utilization, ranging from ; Indicates delay penalty; The weight is adjustable and is set according to process requirements.

[0072] Step S4: training the resource allocation strategy based on the PPO reinforcement learning algorithm, selecting the optimal action based on the current state, and controlling the execution of the communication resource allocation strategy and the computing resource allocation strategy;

[0073] Step S5: Update the state space and repeatedly generate the communication resource allocation strategy and the computing resource allocation strategy.

[0074] In this embodiment, the control algorithms contained in the scanning and mapping container, formation control container, and planning and scheduling container of the control unit include a low-complexity algorithm, a medium-complexity algorithm, and a high-complexity algorithm, respectively; in the corresponding containers and their control cycles, when the allocated computing resource utilization is greater than 80% or the remaining control cycle time is less than 20%, switch to the low-complexity algorithm; when the allocated computing resource utilization is greater than 50% and less than or equal to 80% and the remaining control cycle time is greater than 50%, switch to the medium-complexity algorithm; when the allocated computing resource utilization is less than 50% and the remaining control cycle time is greater than 80%, switch to the high-complexity algorithm.

[0075] Specifically, the low-complexity algorithm is a PID control algorithm with a computing power consumption of less than 10TOPS; the medium-complexity algorithm is an LQR control algorithm with a computing power consumption of 10~50TOPS; and the high-complexity algorithm is an MPC control algorithm with a computing power consumption of more than 50TOPS.

[0076] In this embodiment, the method for the resource allocation unit to control the sending frequency and data packet size of the corresponding data sent by the communication module includes controlling the communication module to process the corresponding data, and when using a low-complexity algorithm, reducing the data packet sending frequency and reducing the size of the data in the data packet; when using a higher-complexity algorithm, increasing the data packet sending frequency accordingly and increasing the size of the data in the data packet.

[0077] Example 3:

[0078] This embodiment provides a computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, the method for jointly allocating communication and computing resources for aircraft wing handling described in the second embodiment is implemented.

[0079] The technical features of the above-mentioned embodiments can be combined arbitrarily. In order to make the description concise, not all possible combinations of the technical features in the above-mentioned embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0080] The above-described embodiments merely represent several implementation methods of the present application. While the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the patent application. It should be noted that a person of ordinary skill in the art may make various modifications and improvements without departing from the spirit of the present application, and these modifications and improvements fall within the scope of protection of the present application. Therefore, the scope of protection of the present patent application shall be determined by the appended claims.

Claims

1. A system for jointly allocating communication and computing resources for aircraft wing transport, characterized in that: It includes an edge server, a communication module and a transport vehicle, wherein the edge server is communicatively connected with a plurality of the transport vehicles through the communication module; The edge server includes a sensing unit, a control unit and a resource allocation unit, wherein the sensing unit, the control unit and the resource allocation unit include a plurality of virtualized containers; The control unit presets a control algorithm for forming a control strategy based on the perception data, wherein the control algorithm includes a low complexity algorithm, a medium complexity algorithm, and a high complexity algorithm, and the computing power consumed during operation increases in sequence; The resource allocation unit is configured to allocate the communication resources of the communication module and the computing resources of the control unit based on the resource utilization of the communication module and the edge server through a reinforcement learning algorithm; the resource allocation unit is configured to select a control algorithm for the control unit based on the processor utilization of the edge server and the remaining time of the control cycle, and to control the transmission frequency and data packet size of the corresponding data sent by the communication module; The reinforcement learning algorithm of the resource allocation unit is used to: define a state space, the state space including the process type, the communication system state, the computing system state, and the task requirement matrix; define an action space, the action space including the communication resource allocation strategy of the communication module and the computing resource allocation strategy of the edge server; A reward function is defined, wherein the reward value of the reward function includes the satisfaction of task requirements, resource utilization, and delay penalty; a resource allocation strategy is trained based on the PPO reinforcement learning algorithm, an optimal action is selected based on the current state, and a communication resource allocation strategy and a computing resource allocation strategy are controlled and executed; Update the state space and repeatedly generate communication resource allocation strategies and computing resource allocation strategies; The transport vehicle includes an actuator and a sensing mechanism; the sensing mechanism is used to obtain sensing data; the actuator is used to control the transport vehicle based on the control strategy; The communication module performs network slicing on the transmitted data, and includes an uplink module and a downlink module. The uplink module is used to transmit the perception data of the perception mechanism of the transport vehicle to the perception unit of the edge server; the downlink module is used to transmit the control strategy of the control unit of the edge server to the execution mechanism of the transport vehicle.

2. A system for jointly allocating communication and computing resources for aircraft wing transport according to claim 1, characterized in that: The perception unit includes a data fusion container, a mechanical analysis container, and an anomaly detection container; the data fusion container is used to integrate the perception data of the lidar, inertial measurement unit, and odometer to form the positioning data of the transport vehicle; the mechanical analysis container is used to evaluate the force on the wing workpiece based on the force data; the anomaly detection container is used to monitor the jump of the perception data; the control unit includes a scanning and mapping container, a formation control container, and a planning and scheduling container, which are used to generate a control strategy based on the output of the perception unit.

3. A system for jointly allocating communication and computing resources for aircraft wing transport according to claim 2, characterized in that: The sensing mechanism includes a laser radar, an inertial measurement unit, an odometer and a force sensor, and the force sensor is used to monitor the force data of the transport vehicle and the wing workpiece; the actuator includes a chassis actuator and an upper body actuator, which are used to control the transport vehicle based on the control strategy.

4. The communication and computing resource joint allocation system for aircraft wing transport according to claim 1, characterized in that: The resource allocation unit is used to allocate initial computing power to the scanning and mapping container, formation control container and planning and scheduling container of the control unit, where the scanning and mapping container allocates 100~500 TOPS computing power, the formation control container allocates 10~100 TOPS computing power, and the planning and scheduling container allocates 500~1000 TOPS computing power; allocates an initial network slicing strategy to the downlink module of the communication module, allocates eMBB slices to the control policy output by the scanning and mapping container, allocates uRLLC slices to the formation control container, and allocates mMTC slices to the planning and scheduling container.

5. The communication and computing resource joint allocation system for aircraft wing transport according to claim 1, characterized in that: The task requirement matrix is ​​[Q, C, T, P, S], where Q is the QoS requirement, C is the computing power requirement, T is the control period, P is the control accuracy, and S is the perception fusion dimension.

6. The communication and computing resource joint allocation system for aircraft wing transport according to claim 1, characterized in that: The control algorithms contained in the scanning and mapping container, formation control container, and planning and scheduling container of the control unit include a low-complexity algorithm, a medium-complexity algorithm, and a high-complexity algorithm, respectively; in the corresponding container and its control cycle, when the allocated computing resource utilization is greater than 80% or the remaining control cycle time is less than 20%, switch to the low-complexity algorithm; when the allocated computing resource utilization is greater than 50% and less than or equal to 80% and the remaining control cycle time is greater than 50%, switch to the medium-complexity algorithm; when the allocated computing resource utilization is less than 50% and the remaining control cycle time is greater than 80%, switch to the high-complexity algorithm.

7. A system for jointly allocating communication and computing resources for aircraft wing transport according to claim 6, characterized in that: The low-complexity algorithm is a PID control algorithm with a computing power consumption of less than 10TOPS; the medium-complexity algorithm is an LQR control algorithm with a computing power consumption of 10~50TOPS; the high-complexity algorithm is an MPC control algorithm with a computing power consumption of more than 50TOPS.

8. A method for jointly allocating communication and computing resources for aircraft wing handling, characterized in that: The communication and computing resource joint allocation system for aircraft wing transportation as described in claim 1 comprises: a resource allocation unit allocates communication resources of the communication module and computing resources of the control unit based on the resource utilization of the communication module and the edge server through a reinforcement learning algorithm; the resource allocation unit selects a control algorithm for the control unit based on the processor utilization of the edge server and the remaining time of the control cycle, and controls the sending frequency and data packet size of the corresponding data sent by the communication module; the perception mechanism of the transport vehicle obtains perception data; the execution mechanism controls the transport vehicle based on the control strategy; the communication module performs network slicing on the transmitted data, and the uplink module transmits the perception data of the perception mechanism of the transport vehicle to the perception unit of the edge server; the downlink module transmits the control strategy of the control unit of the edge server to the execution mechanism of the transport vehicle.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method for jointly allocating communication and computing resources for aircraft wing handling according to claim 8 is implemented.

Images