Multi-terminal cooperative sensing communication resource allocation modeling method and system

Abstract

The invention relates to the technical field of mobile communication, in particular to a multi-terminal cooperative sensing communication resource allocation modeling method and system. According to the multi-terminal cooperative sensing communication resource allocation modeling method, environment information is collected and analyzed in real time by utilizing the computing power of an edge server, and the environment information comprises the obstacle position and the AGV density; in a smart factory system comprising a plurality of IPPO base stations, a smart factory automated guided vehicle positioning resource allocation task model is constructed based on a sensing integrated network and comprises a sensing task model, a calculation task model, a communication task model and a resource allocation task model, and cooperative sensing resource allocation of complex tasks is realized. The positioning precision, the communication efficiency and the system performance are improved. According to the multi-terminal collaborative perception communication resource allocation modeling method and system, collaborative perception resource allocation of complex tasks is realized, and accurate perception of a large number of automatic guided vehicles (AGVs) and efficient transmission of various heterogeneous service data in a factory can be guaranteed under the condition of limited resources.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of mobile communication technology, in particular to a multi-terminal cooperative sensing communication resource allocation modeling method and system. BACKGROUND

[0002] Industrial manufacturing scenarios develop towards intelligent factories, and as a key component of intelligent factories, Auto Guided Vehicles (AGVs) undertake important tasks such as material handling and production line docking, and their applications are increasingly widespread. The emergence of AGVs not only improves production efficiency and reduces labor costs, but also puts forward higher requirements for the positioning and communication of AGVs.

[0003] To meet the requirements of intelligent factories for flexible scheduling of AGVs, a large number of wireless communication nodes are deployed in the factory to ensure seamless coverage of the wireless network in the workshop or even the factory site. With the development of sensing and communication integration technology, it is possible to use a large number of 5G / B5G base stations deployed in the environment to meet the communication and positioning requirements of AGVs.

[0004] On the one hand, the use of wireless resources can improve the sensing function, and the sensing function can share the same hardware and spectrum resources with the communication function to complete different types of sensing tasks; on the other hand, with the assistance of sensing capability, the efficiency of the process of establishing, managing and networking of communication links will be further improved, the quality of communication services will be improved, and the sensing and communication capabilities will be mutually beneficial.

[0005] However, in the sensing and communication integration system, communication and sensing tasks share limited wireless resources, and it is difficult to guarantee accurate sensing of a large number of AGVs and efficient transmission of various heterogeneous business data in the factory site under limited resources, therefore, the resource scheduling algorithm of sensing and communication integration needs to be specially designed to maximize the resource efficiency of sensing and communication.

[0006] Based on this, the present application provides a multi-terminal cooperative sensing communication resource allocation modeling method and system. SUMMARY

[0007] The present application provides a simple and efficient multi-terminal cooperative sensing communication resource allocation modeling method and system to overcome the defects of the prior art.

[0008] The present application is achieved by the following technical solutions:

[0009] A multi-terminal cooperative sensing communication resource allocation modeling method, comprising the following steps:

[0010] Step S1, real-time collection and analysis of environmental information

[0011] Real-time collection and analysis of environmental information, including obstacle location and AGV density, using the computing power of edge servers;

[0012] Step S2, constructing a task model based on the integrated sensing network

[0013] In a smart factory system comprising a plurality of IPPO base stations, a smart factory AGV positioning resource allocation task model is constructed based on the integrated sensing network, including a perception task model, a computing task model, a communication task model, and a resource allocation task model, to realize collaborative sensing resource allocation for complex tasks, thereby improving positioning accuracy, communication efficiency, and system performance.

[0014] In step S2, as a sensing target, the task model of the AGV includes a communication task model and a perception task model;

[0015] To achieve dynamic and accurate AGV perception, the overall processing delay of the perception task of each AGV must not exceed a custom threshold, i.e., the selection of multiple IPPO base stations for sensing a specific AGV, and the data offloading, transmission, and processing in the AGV perception task and computing task must be completed within the custom threshold range.

[0016] To sense targets, IPPO base stations need to emit radar beams in several directions of interest; radar sensing performance is guided by beam pattern error and mean square cross-correlation pattern to design beamforming;

[0017] The perception task model is constructed by beam pattern error and mean square cross-correlation pattern;

[0018] 1) Beam pattern error

[0019] Beam pattern error is used to evaluate the quality of radar beams, the purpose is to match the expected beam pattern, and optimize the transmission power in multiple directions of interest; the beam pattern error of the kth IPPO base station is denoted as η k , which is calculated as the mean square error between the expected beam pattern and the designed beam pattern, denoted as:

[0020]

[0021] where θ m represents the azimuth angle of the target AGV m in the polar coordinates of IPPO base station k, ρ k (θ m ) represents the given expected beam pattern, a k (θ m ) represents the steering vector of IPPO base station k at AGV m, is a k (θ m ) is the conjugate transpose matrix of x k is the set of AGVs served by IPPO base station k, |M k | represents the number of elements in set M k

[0022] S x,k represents the transmit signal covariance matrix, denoted as:

[0023]

[0024] where x k represents the base station transmit signal vector of IPPO base station k, constructed according to the array structure, is the conjugate transpose matrix of x k .

[0025] 2) Mean-square cross-correlation pattern:

[0026] To design high-quality probe beams, it is necessary to reduce the cross-correlation between the considered directions. The mean-square cross-correlation pattern φ k of the kth IPPO base station is denoted as:

[0027]

[0028] In addition, considering the mutual interference between multiple IPPO base stations, in the multi-base station scenario, the interference plus noise power is included in the sensing task model as an additional indicator of radar sensing; the interference plus noise power is denoted as Θ k .

[0029]

[0030] where K represents the set of all IPPO base stations, M j represents the set of AGVs for which IPPO base station j is responsible for sensing; represents the sensing beam weight vector of IPPO base station j to AGV m, represents the communication beam weight matrix; represents the sensing link channel matrix from IPPO base station j to IPPO base station k in the radar echo channel, i.e., Radar-to-Radar interference; H kj represents the communication link channel matrix from IPPO base station j to IPPO base station k in the signal interference path, i.e., Comm-to-radar interference; represents the noise power (Gaussian white noise) at the receiving end of IPPO base station k.

[0031] ​These metrics jointly measure the performance of radar sensing in a multi-cell coordinated IPPO system, which are crucial for designing effective beamforming strategies and optimizing the overall system performance.

[0032] The sensing task decomposition of the integrated communication and sensing network is a key step to realize the fusion of communication and sensing functions.

[0033] In the step S2, the complex sensing task is decomposed into subtasks executable by each node in the network through the calculation task model, while ensuring efficient allocation and execution of the task;

[0034] The sensing task is composed of Y subtasks with dependencies, denoted by set Y; based on this, a weighted directed acyclic graph DAG is used to simulate multi-task service, denoted as G={V, E}, wherein the nodes in V={1, 2,..., Y} represent subtasks, and the edges in E={(i, j)|i, j∈V} represent the dependency relationship between subtasks; in addition, the dependency indicates that a subtask can only be executed after its predecessor task is completed and the result is received;

[0035] The node represents a subtask, and has a computation attribute, denoted as w={w i |i∈V}, which represents the number of CPU cycles required by the subtask; if the subtask y is executed on the resource node RNn, it is denoted as x y,n =1, otherwise 0.

[0036] For the resource node RNn in the network, the total resource amount is f n (cycle / s), and the fixed computation ability allocated to the subtask is defined as f n,y (cycle / s), y∈Y, and the computation duration is:

[0037]

[0038] wherein w i is the computation requirement of the subtask i.

[0039] In addition to the computation resource requirement, the edge between the nodes represents the predecessor-successor relationship of the task, and has a communication attribute, which indicates that the transmission bandwidth between the nodes needs to be ensured to meet the transmission requirement r i,j ,(i, k)∈E between the subtasks to maintain smooth cooperation.

[0040] For the directed acyclic graph DAG task, since the subtasks have the predecessor-successor relationship, only after the successor tasks are all completed and the data is transmitted to the nodes where the subtasks are located, the subtask can start to be executed.

[0041] The start time and end time of the subtask i are defined as t start,i and t end,i, the allocation decision vector is defined as α=(α1,α2,…,α K ), where α K =n, which means deploying subtask i to resource node RNn; the total execution time overhead of the task is T delay is the end time of the last completed subtask, expressed as:

[0042]

[0043] In step S2, the communication task model is responsible for implementing the data unloading process. The transmission signal x of the kth IPPO base station k Expressed as:

[0044]

[0045] The channel coefficient from the kth IPPO base station to the macro base station is expressed as H k ∈C L×L ; When allocating resources, allocate the available spectrum B to different AGVs and tasks, using b ik represents the spectrum resources allocated to the communication task between the automatic guided vehicle AGVm and the IPPO base station k, and the data offloading rate of the kth IPPO base station Expressed as:

[0046]

[0047] Among them, I M represents the M-order identity matrix;

[0048] In addition, the computing resources of M edge servers are allocated to different IPPO base stations to process the perception tasks of the automatic guided vehicle AGV, using c ikm Indicates that the IPPO base station n k The AGV v i The sensing task is offloaded to the edge server s m The allocation of computing resources for processing;

[0049] The IPPO base station allocates resources for each AGV. The total resource allocation for all AGVs’ perception and communication cannot exceed the bandwidth resource B of a single IPPO base station. total and total power P total , which is expressed as follows:

[0050]

[0051] If each IPPO base station provides the sensing data offloading function, the total sensing power cannot exceed the base station sensing power upper limit Cc, which is expressed as

[0052] The total throughput of the AGV represents the sum of the data transmission rates between all AGVs and the edge server, thereby improving the overall communication efficiency of the system; the average delay of the AGV The average value of the data transmission delay between the AGV and the edge server is represented by D

[0053] Time delay constraint

[0054] Specifically, the end-to-end delay is represented as: D proc (t) = D trans (t) + D comp (t) ≤ D max

[0055] where D trans (t) represents the transmission delay, D comp (t) represents the calculation delay, and D max represents the maximum delay;

[0056]

[0057] Since the AGV has a motion speed, an excessively long processing time will cause the positioning to deviate, affecting the positioning accuracy, therefore the performance of the IPPO base station for AGV positioning includes the end-to-end processing process of the perception data, and the positioning error The Cramer-Rao lower bound (CRLB) is used to constrain:

[0058]

[0059] where c is the speed of light, and SNR(t) is the signal-to-noise ratio;

[0060]

[0061] In the step S2, the target of the resource allocation task model is to maximize the system communication resource use efficiency under the premise of accurate AGV positioning; the resource allocation task model is constructed as follows:

[0062]

[0063] where G s (t) is the positioning accuracy of the AGV at time t, R c (t) is the spectrum efficiency at time t, and η comp (t) is the resource efficiency at time t;

[0064] The constraint conditions are as follows:

[0065] C1: Ensure that the start time of subtask i should be greater than or equal to zero, expressed as: t start,i ≥ 0;

[0066] C2: Ensure that the execution of the subtask is atomic, expressed as: t end,i = t start,i + t i,k ;

[0067] C3: Ensure that the subtasks are assigned to different resource nodes Rn, because each subtask has different characteristics, expressed as:

[0068] C4: A subtask can only be executed after all its predecessor tasks are completed and the results are received, expressed as:

[0069]

[0070] C5: The total sum of resource allocation for all automated guided vehicles AGV perception and communication cannot exceed the bandwidth resources of a single IPPO base station, expressed as:

[0071]

[0072] C6: The total sum of resource allocation for all automated guided vehicles AGV perception and communication cannot exceed the total power of a single IPPO base station, expressed as:

[0073]

[0074] C7: Perception needs to satisfy that the perception deviation is less than a user-defined upper bound, expressed as:

[0075] A multi-terminal collaborative perception communication resource allocation modeling system for implementing the above method, comprising a data acquisition module, a perception task model construction module, a computing task model construction module, a communication task model construction module and a resource allocation task model construction module.

[0076] A multi-terminal collaborative perception communication resource allocation modeling computing device, comprising:

[0077] One or more processors, one or more memories, and one or more programs, wherein the one or more programs are stored in the one or more memories and configured to be executed by the one or more processors, the one or more programs comprising instructions for executing any of the above methods.

[0078] A computer-readable storage medium storing one or more programs, the one or more programs comprising instructions which, when executed by a multi-terminal co-sensing communication resource allocation modeling computing device, cause the multi-terminal co-sensing communication resource allocation modeling computing device to perform any of the above-described methods.

[0079] The beneficial effects of the present application are: the multi-terminal co-sensing communication resource allocation modeling method and system realize the co-sensing resource allocation of complex tasks, and can guarantee the accurate perception of a large number of automatic guided vehicles AGVs and the efficient transmission of various heterogeneous business data in a factory area under limited resource conditions. BRIEF DESCRIPTION OF DRAWINGS

[0080] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the drawings needed to be used in the embodiments or prior art description will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor on the basis of these drawings.

[0081] ATTACHMENT Figure 1 The figure is a schematic diagram of the intelligent factory automatic guided vehicle positioning resource allocation task model. DETAILED DESCRIPTION

[0082] In order to make the technical personnel in the art better understand the technical solutions in the present application, the technical solutions in the embodiments of the present application will be described clearly and completely below in conjunction with the embodiments of the present application. Obviously, the described embodiments are only some of the embodiments of the present application, not all. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor should be within the scope of protection of the present application.

[0083] There are multiple IPPO base stations (IBS) in the intelligent factory system, each IPPO base station has a certain processing capability and can connect to the edge cloud for calculation offloading or backhaul data; at the same time, each IPPO base station is equipped with several antennas.

[0084] The target user of the perception and communication in the factory area is the automatic guided vehicle AGV, each IPPO base station collects perception data, converts the signal to digital bits by sending a probe signal and receiving a return signal. The data volume represents the data volume required for perceiving each perception target, which is a constant determined by the actual configuration of radar perception; if the perception scheduling z k,m =1, it means that the mth target is perceived by the kth IPPO base station.

[0085] In the IPPO system, data offloading is the process of sending radar perception data collected by IPPO base stations to edge servers, which is crucial for training machine learning models and implementing edge intelligence. The offloading rate of IPPO base stations sending their collected perception data to macro base stations is not less than a custom rate threshold to ensure timely transmission of data.

[0086] After the macro base station receives the perception data from the IPPO base station, the edge server uses the received data to train the machine learning model.

[0087] The multi-terminal cooperative perception communication resource allocation modeling method includes the following steps:

[0088] Step S1, real-time collection and analysis of environmental information

[0089] Real-time collection and analysis of environmental information using the computing power of the edge server, including obstacle location and AGV density;

[0090] Step S2, building a task model based on a common-sense integrated network

[0091] In a smart factory system containing several IPPO base stations, a smart factory AGV positioning resource allocation task model is built based on a common-sense integrated network, including a perception task model, a computing task model, a communication task model, and a resource allocation task model. This model realizes cooperative perception resource allocation for complex tasks to improve positioning accuracy, communication efficiency, and system performance.

[0092] In step S2, as a perception target, the task model of the AGV includes a communication task model and a perception task model;

[0093] To achieve dynamic and accurate AGV perception, the overall processing delay of each AGV's perception task cannot exceed a custom threshold. The selection of multiple IPPO base stations for perceiving a specific AGV, the offloading, transmission, and processing of data in the AGV perception task and computing task must be completed within the custom threshold range.

[0094] To perceive targets, IPPO base stations need to emit radar beams in several directions of interest; radar perception performance is guided by beam pattern error and mean square cross-correlation pattern to design beamforming;

[0095] The perception task model is constructed through beam pattern error and mean square cross-correlation pattern.

[0096] 1) Beam pattern error

[0097] The beam pattern error is used to evaluate the quality of the radar beam, with the goal of matching the desired beam pattern and optimizing the transmit power of interest in multiple directions. The beam pattern error of the kth IPPO base station is expressed as η k , calculated as the mean square error between the desired beam pattern and the designed beam pattern, expressed as:

[0098]

[0099] Among them, θ m represents the azimuth of the target automated guided vehicle AGV m in the polar coordinates of the IPPO base station k, ρ k (θ m ) represents a given desired beam pattern, a k (θ m ) represents the guidance vector of IPPO base station k at the automatic guided vehicle AGV m, is a k (θ m )’s conjugate transposed matrix; M k It is a collection of automatic guided vehicles (AGVs) served by IPPO base station k. k | represents the set M k The number of elements in ;

[0100] S x,k represents the covariance matrix of the transmitted signal, which is expressed as:

[0101]

[0102] Among them, x k represents the base station transmission signal vector of IPPO base station k, constructed according to the array structure, is x k The conjugate transposed matrix of .

[0103] 2) Mean square cross correlation mode:

[0104] In order to design a high-quality sounding beam, it is necessary to reduce the cross-correlation between the considered directions. The mean square cross-correlation pattern φ of the kth IPPO base station k Expressed as:

[0105]

[0106] In addition, considering the mutual interference between multiple IPPO base stations, in the multi-base station scenario, the perception task model includes interference plus noise power as an additional indicator of radar perception; the interference plus noise power is expressed as Θ k ;

[0107]

[0108] where K represents the set of all IPPO base stations, M j represents the set of AGVs that IPPO base station j is responsible for sensing; represents the sensing beam weight vector of IPPO base station j to AGV m, represents the communication beam weight matrix; represents the sensing link channel matrix from IPPO base station j to IPPO base station k in the radar echo channel, i.e., Radar-to-Radar interference; H kj represents the communication link channel matrix from IPPO base station j to IPPO base station k in the signal interference path, i.e., Comm-to-radar interference; represents the noise power (Gaussian white noise) at the receiving end of IPPO base station k.

[0109] These indicators jointly measure the performance of radar sensing in a multi-cell cooperative IPPO system, and they are crucial for designing effective beamforming strategies and optimizing overall system performance.

[0110] The sensing task decomposition of the integrated sensing and communication network is a key step to realize the fusion of communication and sensing functions.

[0111] In the step S2, the complex sensing task is decomposed into subtasks that can be executed by each node in the network through a task model, while ensuring efficient allocation and execution of the task;

[0112] The sensing task consists of Y subtasks with dependencies, represented by the set Y; based on this, a weighted directed acyclic graph DAG is used to simulate multi-task service, denoted as G = {V, E}, where the nodes in V = {1, 2,..., Y} represent subtasks, and the edges in E = {(i, j) | i, j ∈ V} represent the dependency relationship between subtasks; in addition, the dependency indicates that a subtask can only be executed after its predecessor tasks are completed and the results are received;

[0113] The nodes represent subtasks, with a computational attribute, denoted as w = {w i | i ∈ V} represents the number of CPU cycles required by the subtask; if subtask y is executed on resource node RNn, it is denoted as x y,n = 1, otherwise 0;

[0114] For the resource node RNn in the network, the total resource amount is f n (cycle / s), and the fixed computing power allocated to the subtask is defined as f n,y (cycle / s), y ∈ Y, then the computation duration is:

[0115]

[0116] where w i is the computation requirement of subtask i.

[0117] In addition to the computation resource requirement, the edge between nodes represents the predecessor and successor relationship of tasks, with a communication attribute, indicating that the transmission bandwidth between nodes needs to meet the transmission requirement r i,j between subtasks.

[0118] For a directed acyclic graph (DAG) task, since the subtasks have a predecessor and successor relationship, only after the successor task is completed and the data is transmitted to the node where the subtask is located, the subtask can start execution;

[0119] The start time and end time of subtask i are defined as t start,i and t end,i , and the allocation decision vector is defined as α = (α1, α2, …, α K ), where α k = n represents deploying subtask i to resource node RNn; the total execution time overhead T delay of the task is the end time of the last executed subtask, expressed as:

[0120]

[0121] In the step S2, the communication task model is responsible for implementing the data offloading process, and the transmission signal x k of the kth IPPO base station is expressed as:

[0122]

[0123] The channel coefficient from the kth IPPO base station to the macro base station is expressed as H k ∈ C L×L ; when allocating resources, the available spectrum B is allocated to different automatic guided vehicles AGV and tasks, and b ik is used to represent the spectrum resource allocated to the communication task between the automatic guided vehicle AGVm and the IPPO base station k, and the data offloading rate of the kth IPPO base station is expressed as:

[0124]

[0125] where I M represents an M-order unit matrix;

[0126] In addition, the M edge servers allocate computing resources to different IPPO base stations to process the perception task of the automatic guided vehicle AGV, and c ikm is used to represent the computing resource allocated to the IPPO base station n k ​The AGV v i The sensing task is offloaded to the edge server s m The allocation of computing resources for processing;

[0127] The IPPO base station allocates resources for each AGV. The total resource allocation for all AGVs’ perception and communication cannot exceed the bandwidth resource B of a single IPPO base station. total and total power P total , which is expressed as follows:

[0128]

[0129] If each IPPO base station provides the sensing data offloading function, the total sensing power cannot exceed the base station sensing power upper limit Cc, which is expressed as

[0130] The total throughput of the AGV represents the sum of the data transmission rates between all AGVs and the edge server, thereby improving the overall communication efficiency of the system; the average latency of the AGV Represents the average value of the data transmission delay between the automated guided vehicle (AGV) and the edge server;

[0131] Delay Constraints

[0132] Specifically, the end-to-end delay is expressed as: D proc (t) = D trans (t)+D comp (t)≤D max ,

[0133] Among them, D trans (t) represents the transmission delay, D comp (t) represents the computation delay, D max Indicates the maximum delay;

[0134]

[0135] Since the AGV has a certain movement speed, too long processing time will cause positioning deviation and affect positioning accuracy. Therefore, the performance of the IPPO base station for AGV positioning includes the end-to-end processing of the perception data, and the positioning error Use the Cramer-Rao lower bound CRLB constraint:

[0136]

[0137] Where c is the speed of light, SNR(t) is the signal-to-noise ratio;

[0138]

[0139] In the step S2, the target of the resource allocation task model is to maximize the system communication resource use efficiency under the premise of accurate positioning of the automated guided vehicle AGV; the resource allocation task model is constructed as follows:

[0140]

[0141] Wherein, G s (t) is the positioning accuracy of the automated guided vehicle AGV at time t, R c (t) is the spectrum efficiency at time t, η comp (t) is the resource efficiency at time t.

[0142] The constraint conditions are as follows:

[0143] C1: Ensure that the start time of the subtask i should be greater than or equal to zero, expressed as: t start,i ≥ 0.

[0144] C2: Ensure that the execution of the subtask is atomic, expressed as: t end , i = t start,i + t i,k .

[0145] C3: Ensure that the subtasks are allocated to different resource nodes Rn, because each subtask has different characteristics, expressed as:

[0146] C4: A subtask can only be executed after all its predecessor tasks are completed and the results are received, expressed as:

[0147]

[0148] C5: The total sum of resource allocation of all automated guided vehicle AGV perception and communication cannot exceed the bandwidth resource of a single IPPO base station, expressed as:

[0149]

[0150] C6: The total sum of resource allocation of all automated guided vehicle AGV perception and communication cannot exceed the total power of a single IPPO base station, expressed as:

[0151]

[0152] C7: The perception should satisfy that the perception deviation is less than the self-defined upper bound, expressed as:

[0153] The multi-terminal cooperative perception communication resource allocation modeling system is used for realizing the above method and comprises a data acquisition module, a perception task model construction module, a computing task model construction module, a communication task model construction module and a resource allocation task model construction module.

[0154] The multi-terminal cooperative perception communication resource allocation modeling computing device comprises:

[0155] One or more processors, one or more memories and one or more programs, wherein the one or more programs are stored in the one or more memories and configured to be executed by the one or more processors, and the one or more programs comprise instructions for executing any one of the above methods.

[0156] The computer readable storage medium storing one or more programs, the one or more programs comprising instructions that, when executed by a multi-terminal cooperative perception communication resource allocation modeling computing device, cause the multi-terminal cooperative perception communication resource allocation modeling computing device to perform any one of the above methods.

[0157] The multi-terminal cooperative perception communication resource allocation modeling method is oriented to a 5G smart factory, can not only provide accurate positioning capability for the smart factory, but also can guarantee accurate perception of a large number of automatic guided vehicles AGVs and resource efficiency of maximum perception and communication of efficient transmission of various heterogeneous business data in a factory area under limited resource conditions by establishing task models including a perception task model, a computing task model, a communication task model and a resource allocation task model.

[0158] The above embodiments are only one of the specific embodiments of the present application, and the usual changes and replacements made by those skilled in the art within the scope of the technical solutions of the present application should be included in the protection scope of the present application.

Claims

1. A multi-terminal cooperative sensing communication resource allocation modeling method, characterized by: The following steps are involved: Step S1: Real-time collection and analysis of environmental information Leverage the computing power of edge servers to collect and analyze environmental information in real time, including obstacle locations and AGV density; Step S2: Constructing a task model based on the synaesthesia integrated network In a smart factory system consisting of several IPPO base stations, a smart factory automated guided vehicle positioning resource allocation task model is constructed based on a synergistic integrated network, including a perception task model, a computation task model, a communication task model, and a resource allocation task model, to achieve collaborative perception resource allocation for complex tasks, thereby improving positioning accuracy, communication efficiency, and system performance.

2. The multi-terminal cooperative sensing communication resource allocation modeling method according to claim 1, characterized in that: In step S2, as a perception target, the task model of the automatic guided vehicle AGV includes a communication task model and a perception task model; The perception task model is constructed using two indicators: beam pattern error and mean square cross correlation pattern. 1) Beam pattern error The beam pattern error is used to evaluate the quality of the radar beam, with the goal of matching the desired beam pattern and optimizing the transmit power of interest in multiple directions. The beam pattern error of the kth IPPO base station is denoted as η k , calculated as the mean square error between the desired beam pattern and the designed beam pattern, expressed as: Among them, θ m represents the azimuth of the target automated guided vehicle AGV m in the polar coordinates of the IPPO base station k, ρ k (θ m ) represents a given desired beam pattern, a k (θ m ) represents the guidance vector of IPPO base station k at the automatic guided vehicle AGV m, is a k (θ m )’s conjugate transposed matrix; M k It is a collection of automatic guided vehicles (AGVs) served by IPPO base station k. k | represents the set M k The number of elements in ; S x,k represents the covariance matrix of the transmitted signal, which is expressed as: Among them, x k represents the base station transmission signal vector of IPPO base station k, constructed according to the array structure, is x k The conjugate transposed matrix of ; 2) Mean square cross correlation mode: In order to design a high-quality sounding beam, it is necessary to reduce the cross-correlation between the considered directions. The mean square cross-correlation pattern φ of the kth IPPO base station k Expressed as: In addition, considering the mutual interference between multiple IPPO base stations, in the multi-base station scenario, the perception task model includes interference plus noise power as an additional indicator of radar perception; the interference plus noise power is expressed as Θ k ; Among them, K represents the set of all IPPO base stations, M j represents the set of automated guided vehicles (AGVs) that IPPO base station j is responsible for sensing; represents the perception beam weight vector of IPPO base station j to the automatic guided vehicle AGV m, represents the communication beam weight matrix; represents the sensing link channel matrix from IPPO base station j to IPPO base station k in the radar echo channel, i.e., Radar-to-Radar interference; H kj represents the communication link channel matrix from IPPO base station j to IPPO base station k in the signal interference path, i.e., Comm-to-radar interference; represents the noise power at the receiving end of IPPO base station k.

3. The multi-terminal cooperative sensing communication resource allocation modeling method according to claim 2, characterized in that: In step S2, the complex sensing task is decomposed into subtasks that can be performed by each node in the network by computing the task model, while ensuring efficient distribution and execution of the tasks; The perception task consists of Y subtasks with dependencies, represented by the set Y. Based on this, a weighted directed acyclic graph (DAG) is used to simulate the multitask service, denoted as G = {V, E}, where the nodes in V = {1, 2, ..., Y} represent subtasks, and the edges in E = {(i, j) | i, j ∈ V} represent the dependencies between subtasks. In addition, the dependency indicates that a subtask can only be executed after its predecessor task completes and receives its result. Nodes represent subtasks and have computational attributes, using w = {w i |i∈V} represents the number of CPU cycles required for the subtask; if the subtask y is executed on the resource node RNn, it is represented as x y,n =1, otherwise 0; For the resource node RNn in the network, the total resource amount is f n (cycle / s), the fixed computing power allocated to the subtask is defined as f n,y (cycle / s), y∈Y, then the calculation duration is: Among them, w i is the computational requirement of subtask i.

4. The multi-terminal cooperative sensing communication resource allocation modeling method according to claim 3, characterized in that: In addition to computing resource requirements, the edges between nodes represent the successor relationship of tasks and have communication attributes, indicating that the subtasks maintain smooth collaboration and need to ensure that the transmission bandwidth between nodes meets the transmission requirements between subtasks. i,j ,(i,j)∈E; For DAG tasks, since subtasks have a successor relationship, a subtask can only be executed after the current successor task is completed and the data is transmitted to the node where the corresponding subtask is located; The start and end time of subtask i is defined as t start,i and t end,i , the allocation decision vector is defined as α=(α1,α2,…,α K ), where α k =n, which means deploying subtask i to resource node RNn; the total execution time overhead of the task is T delay Expressed as: It is the end time of the last subtask that is completed.

5. The multi-terminal cooperative sensing communication resource allocation modeling method according to claim 4, characterized in that: In step S2, the communication task model is responsible for implementing the data unloading process. The transmission signal x of the kth IPPO base station k Expressed as: The channel coefficient from the kth IPPO base station to the macro base station is expressed as H k ∈C L×L ; When allocating resources, allocate the available spectrum B to different AGVs and tasks, using b ik represents the spectrum resources allocated to the communication task between the automatic guided vehicle AGVm and the IPPO base station k, and the data offloading rate of the kth IPPO base station Expressed as: Among them, I M represents the M-order identity matrix; In addition, the computing resources of M edge servers are allocated to different IPPO base stations to process the perception tasks of the automatic guided vehicle AGV, using c ikm Indicates that the IPPO base station n k The AGV v i The sensing task is offloaded to the edge server s m The allocation of computing resources for processing; The IPPO base station allocates resources for each AGV. The total resource allocation for all AGVs’ perception and communication cannot exceed the bandwidth resource B of a single IPPO base station. total and total power P total , which is expressed as follows: If each IPPO base station provides the sensing data offloading function, the total sensing power cannot exceed the base station sensing power upper limit Cc, which is expressed as: The total throughput of the AGV represents the sum of the data transmission rates between all AGVs and the edge server, thereby improving the overall communication efficiency of the system; the average latency of the AGV Represents the average value of the data transmission delay between the automated guided vehicle (AGV) and the edge server; Delay Constraints Specifically, the end-to-end delay is expressed as: D proc (t) = D trans (t)+D comp (t)≤D max Among them, D trans (t) represents the transmission delay, D comp (t) represents the computation delay, D max Indicates the maximum delay; Since the AGV has a certain movement speed, too long processing time will cause positioning deviation and affect positioning accuracy. Therefore, the performance of the IPPO base station for AGV positioning includes the end-to-end processing of the perception data, and the positioning error Use the Cramer-Rao lower bound CRLB constraint: Where c is the speed of light and SNR(t) is the signal-to-noise ratio, expressed as:

6. The multi-terminal cooperative sensing communication resource allocation modeling method according to claim 5, characterized in that: In step S2, the goal of the resource allocation task model is to maximize the efficiency of system communication resource utilization under the premise of accurate positioning of the automated guided vehicle (AGV). The resource allocation task model is constructed as follows: Among them, G s (t) is the positioning accuracy of the automatic guided vehicle AGV at time t, R c (t) is the spectrum efficiency at time t, η comp (t) is the resource efficiency at time t; The constraints are as follows: C1: Ensure that the start time of subtask i should be greater than or equal to zero, expressed as: t start,i ≥0; C2: Ensure that the execution of subtasks is atomic, expressed as: t end,i =t start,i +t i,k ; C3: Ensure that subtasks are assigned to different resource nodes Rn, because each subtask has different characteristics, expressed as: C4: A subtask can only be executed after all its predecessor tasks have completed and received their results, which is expressed as: C5: The total resource allocation for all AGV perception and communication cannot exceed the bandwidth resources of a single IPPO base station, expressed as: C6: The total resource allocation for all AGV perception and communication cannot exceed the total power of a single IPPO base station, expressed as: C7: The perception deviation must be less than the custom upper bound, expressed as:

7. A multi-terminal collaborative sensing communication resource allocation modeling system, characterized by: The method for implementing any one of claims 1 to 6 comprises a data acquisition module, a perception task model construction module, a computing task model construction module, a communication task model construction module and a resource allocation task model construction module.

8. A multi-terminal collaborative sensing communication resource allocation modeling and computing device, characterized by: include: One or more processors, one or more memories, and one or more programs, wherein the one or more programs are stored in the one or more memories and configured to be executed by the one or more processors, and the one or more programs include instructions for executing the method according to any one of claims 1 to 6.

9. A computer-readable storage medium storing one or more programs, characterized in that: The one or more programs include instructions, which, when executed by a multi-terminal cooperatively aware communication resource allocation modeling computing device, enable the multi-terminal cooperatively aware communication resource allocation modeling computing device to execute the method according to any one of claims 1 to 6.

Images