Offloading Decision and Resource Allocation Method Based on the Integration of Sensing and Computing

Abstract

The present invention discloses an offloading decision and resource allocation method based on integrated communication, sensing and computing. A system cost function is defined by comprehensively considering energy consumption, latency and migration cost. Under the constraints of user perception failure rate and maximum task completion latency, a problem of minimizing the long-term average cost of users is established. Based on the Lyapunov optimization theory, a virtual queue is established for the perception performance of users, and the long-term stochastic optimization problem is transformed into a single-slot deterministic optimization problem by using the Lyapunov drift-plus-penalty function. The transformed problem is divided into an inner layer and an outer layer. The optimal task offloading ratio under each user-base station selection and channel allocation is obtained through theoretical derivation in the inner layer. According to the results solved by the inner layer problem, the user-base station selection and channel allocation are determined through the matching theory in the outer layer.

Description

Technical Field

[0001] The present invention relates to the field of communication technologies, and particularly to an offloading decision and resource allocation method based on integrated communication, sensing, and computing. Background Art

[0002] Mobile Edge Computing (MEC for short) provides computing, communication, and storage capabilities for users at the network edge, significantly reducing the transmission delay of users and enhancing the computing power of users, which helps to meet the requirements of delay-sensitive and computing-intensive applications, such as video stream analysis, virtual reality, autonomous driving, etc.

[0003] In a Mobile Edge Computing network, the offloading decision mainly solves the problems of what to offload, how to offload, and how much to offload. Resource allocation mainly focuses on the joint management of the system's computing resources, storage resources, and communication resources. The offloading decision and resource allocation of users will directly affect the system performance, so it has become a research hotspot. Summary of the Invention

[0004] Due to the development of millimeter-wave communication, the spectra of communication and radar sensing are gradually overlapping. As a key technology in future mobile communications, Integrated Sensing and Communication (ISAC) integrates the hardware of sensing and communication, multiplexes the same frequency band, and uses the designed integrated signal to perform sensing and communication simultaneously, so that the tense spectrum resources can be further utilized.

[0005] Therefore, in order to overcome the deficiencies in the prior art, the present invention proposes an offloading decision and resource allocation method based on integrated communication, sensing, and computing. In a multi-user and multi-edge-node MEC network, this method takes into account the impact of user mobility on performance and integrated communication and sensing, and minimizes user energy consumption, task delay, and migration cost.

[0006] Technical Solution: To solve the above technical problems, the technical solution adopted by the present invention is as follows:

[0007] In a first aspect, the present invention provides an offloading decision and resource allocation method based on integrated communication, sensing, and computing, which performs the following steps in a preset cycle:

[0008] S1. Obtain the position, channel state, task data volume, virtual queue backlog of each user in the current time slot, and the user base station selection in the previous time slot; where each channel of each base station is recorded as a base station-channel pair, and a user accessing a channel of a base station is recorded as the user matching with this base station-channel pair;

[0009] S2. Based on the positions, channel states, task data volumes, virtual queue backlogs of each user in the current time slot, and the user-base station selection in the previous time slot, perform an initial matching according to the preference value of the user for the base station-channel pair and the preference value of the base station-channel pair for the user, and obtain an initial matching result;

[0010] S3. Perform an exchange matching based on the initial matching result to obtain an exchange matching result;

[0011] S4. Based on the exchange matching result, determine the user-base station selection, channel allocation, and task offloading ratio in the current time slot.

[0012] In some embodiments, S2. Performing an initial matching according to the preference value of the user for the base station-channel pair and the preference value of the base station-channel pair for the user includes:

[0013] S2.1. Set the initial set of matched users to be an empty set, the initial set of unmatched users to be the set composed of all users, the initial current matching to be an empty set, the initial set of unmatched base station-channel pairs to be the set composed of each channel of each base station, and the initial set of matched base station-channel pairs to be an empty set;

[0014] S2.2. Calculate the preference value of the user for all unmatched base station-channel pairs under the current matching in a certain order, and the user sends a matching request to the unmatched base station-channel pair with the highest preference value;

[0015] S2.3. Calculate the preference value of all unmatched base station-channel pairs for each requesting user under the current matching in a certain order, and the unmatched base station-channel pair accepts the request of the user with the highest preference value, and both sides reach an initial matching;

[0016] S2.4. Update the current matching, the set of matched users, the set of unmatched users, the unmatched base station-channel pairs, and the set of matched base station-channel pairs;

[0017] S2.5. Repeat steps S2.2 - S2.4 until all users are matched with a certain base station-channel pair, the initial matching stage ends, and an initial matching result is obtained, where the initial matching result includes: the current matching, the set of matched users, the set of unmatched users, the unmatched base station-channel pairs, and the set of matched base station-channel pairs.

[0018] In some embodiments, S3. Performing an exchange matching based on the initial matching result to obtain an exchange matching result includes:

[0019] S3.1. Record the current matching as the pre-exchange matching, and start a round of exchange matching;

[0020] S3.2. Detect the exchange closure pairs for all currently matched users of the base station-channel pairs in a certain order;

[0021] S3.3. If any two users satisfy the exchange closure pair, then exchange the base station-channel pairs matched by these two users;

[0022] S3.4. Update the current matching;

[0023] S3.5. After all users are detected and one round of exchange matching ends, record the current matching as the post-exchange matching;

[0024] S3.6. Repeat steps S3.1 - S3.5 until the pre-exchange matching and the post-exchange matching are the same in each round, and the exchange matching is completed to obtain the exchange matching result.

[0025] In some embodiments, in step S1, obtaining the location, channel state, task data volume, and virtual queue backlog of each user in the current time slot includes:

[0026] S1.1. There are N base stations, M users, and L channels in the system. The base station set is denoted as The user set is denoted as The channel set is denoted as The users are in a mobile state. To represent the movement of the users, the time is discretized into time slots, and the length of one time slot is τ. It is assumed that the position of a user remains unchanged within one time slot and is different in different time slots. The time slots are represented by the set T = {1, 2, 3, …, t, …}. It is assumed that each user has a service, and the service consists of a series of tasks. And there is a dedicated virtual machine or container built on the edge server to process the generated tasks. As the user moves, the virtual machine or container will migrate with the handover of the base station connected by the user. Each user generates a task in each time slot. The amount of data required for communication of the task generated by user m in time slot t is denoted by I m (t), the task computing load is denoted by C m (t), the maximum task delay requirement is denoted by D m (t), and it is required that D m (t) ≤ τ, the task offloading ratio is denoted by ρ m (t), and it is required that ρ m (t) ∈ [ρ min , 1], part of the task is computed locally and part is computed at the base station; each user connects to a base station in each time slot to offload task data; each user can use the communication-sensing integrated signal to sense the surrounding environment when transmitting task data to the base station in each time slot; all base stations share the same multiple channels, and both the large-scale and small-scale fading of the channels are considered;

[0027] S1.2, Indicates the association between user m and base station n at time slot t. When user m accesses base station n at time slot t otherwise it is 0; all base stations share L channels, and each channel has the same bandwidth W. The signals of users using the same channel will interfere with each other; Indicates the usage of channel l by user m at time slot t. When user m occupies channel l at time slot t otherwise it is 0; a user can only access one base station and occupy one channel in one time slot;

[0028] S1.3. Use the conditional mutual information between the echo signal and the channel impulse response to measure the performance of the user's perception of the surrounding environment. The conditional mutual information MI of user m at time slot t m (t) is expressed as:

[0029]

[0030] where Indicates the radar signal-to-dry ratio of user m on channel l at time slot t, expressed as:

[0031]

[0032] where P m Indicates the transmit power of user m, K indicates the number of symbols of the integrated signal, T s Indicates the duration of one symbol, Indicates the Fourier transform of the channel impulse response at frequency f of channel l l which follows a standard normal distribution; Indicates the channel gain between user m' and user m on channel l, σ 2 = N0W indicates the channel additive white Gaussian noise power;

[0033] When the user mutual information is less than the set threshold MI min then the perception fails, expressed as:

[0034]

[0035] where indicator {x} Indicates the indicator function, that is, when the expression x is true, the value of the indicator function is 1, otherwise it is 0;

[0036] S1.4. For the perception failure event S of user m at time slot t m (t), define a virtual queue with an initial value of 0, and express the backlog of each queue as Q m (t), where Q m (0) = 0, use Sm (t) represents the increase in the virtual queue at time slot t, while δ represents the decrease in the virtual queue; the virtual queue backlog of user m between time slot t and time slot t + 1 satisfies:

[0037] Q m (t + 1) = [Q m (t) + S m (t) - δ] + 。

[0038] In some embodiments, in step S2, the calculation method for the preference value of the user for the base station-channel pair and the preference value of the base station-channel pair for the user in the current match includes:

[0039] The data transmission rate at which user m offloads tasks to the base station at time slot t is expressed as:

[0040]

[0041] where represents the channel gain between user m' and base station n on channel l, represents the interference caused to user m by other users using the same channel;

[0042] The time for user m to transmit tasks at time slot t is expressed as:

[0043]

[0044] The energy consumption of user m for transmitting tasks at time slot t is expressed as:

[0045]

[0046] Using f m,local to represent the computing power of user m, the local computing delay of the task of user m at time slot t is:

[0047]

[0048] The energy consumption of user m for locally computing tasks at time slot t is:

[0049]

[0050] where refers to the power coefficient;

[0051] After receiving the tasks transmitted by the users, the base station evenly distributes all computing resources among the computing tasks of all users. Then, the computing resources allocated by base station n to user m at time slot t is:

[0052]

[0053] The computing resource f obtained by user m at time slot t m (t) is:

[0054]

[0055] The task computing delay of user m at time slot t is:

[0056]

[0057] The task processing delay T of user m at time slot t m (t) is the maximum of the local computing delay and the base station computing delay, expressed as:

[0058]

[0059] The energy consumption E of user m for processing tasks at time slot t m (t) is:

[0060]

[0061] The migration cost of user m at time slot t is expressed as:

[0062]

[0063] where ε represents the migration cost and is a fixed value; when the base station selected by the user in this time slot is different from that in the previous time slot, there will be a migration cost of size ε, otherwise there is none;

[0064] The cost function U of user m at time slot t m (t) is expressed as:

[0065]

[0066] where α and β are the weight coefficients of delay and energy consumption respectively;

[0067] Establish the optimization problem P1 of minimizing the long-term cost of the user:

[0068]

[0069]

[0070]

[0071]

[0072]

[0073]

[0074]

[0075]

[0076] Among them, the variable set represents the access situation of each user to each base station in each time slot, represents the occupancy situation of each user for each channel in each time slot, represents the task division ratio of each user in each time slot; Constraints (C1) and (C2) mean that in each time slot, a user can only select one base station to access and can only occupy one channel; Constraint (C3) means that one channel of one base station can be used by at most one user in one time slot; Constraint (C4) means that the time average of the user's sensing failure event expectation E[S m (t)] is less than a certain set value δ, that is, the long-term sensing failure rate of the user should be below a certain threshold; Constraint (C5) means that each user's task should be completed within the maximum delay D m (t) requirement; Constraint (C6) means that the task of each user in each time slot is randomly divided between the minimum value ρ min and 1; Constraint (C7) means that the base station selection and channel allocation variables are binary variables;

[0077] The optimization problem P1 is transformed into the optimization problem P2 through the Lyapunov optimization theory:

[0078]

[0079]

[0080]

[0081]

[0082]

[0083]

[0084]

[0085] Among them, Q m (t) is the virtual queue backlog of user m at time slot t, S m (t) is the sensing failure event of user m at time slot t, is a constant, and V is the control parameter for balancing the stability of the virtual queue and the sum of the user cost functions;

[0086] Solve the optimization problem P2 to obtain the base station selection and channel allocation value corresponding to the current user matching;

[0087] When the base station selection of the user and channel allocation are determined, the optimization problem P2 is simplified to the optimization problem P2.1

[0088]

[0089]

[0090]

[0091] Solve the optimization problem P2.1 to obtain the user task offloading ratio under the current matching;

[0092] Let F0 = 1 - D m (t)f m,local / C m (t), When F0 > F1 or F1 < ρ min the problem P2.1 has no solution. Conversely, the optimal solution of P2.1, that is, the user task offloading ratio is:

[0093]

[0094] where

[0095]

[0096]

[0097]

[0098]

[0099]

[0100]

[0101]

[0102] According to the obtained user task offloading ratio, calculate the preference value of the user for the base station-channel pair and the preference value of the base station-channel pair for the user under the current matching;

[0103] User Preference value for the base station-channel pair under the current match v at time slot t is:

[0104]

[0105] where Iv m (t) = 1 indicates that the inner problem of user m has a solution at time slot t, that is, the maximum delay constraint can be satisfied. Conversely, Iv m (t) takes 0, and η is a positive number with a very large order of magnitude;

[0106] Preference value for the user under the current match v of the base station-channel pair at time slot t is:

[0107]

[0108] In some embodiments, in step S3.2, the detection method for the exchange closed pair includes:

[0109] Under the current match, assume that two users exchange the matched base station-channel pairs, and other users keep the currently matched base station-channel pairs;

[0110] If after the two users exchange the matched base station-channel pairs, the preference values of the users for the base station-channel pairs and the preference values of the base station-channel pairs for the users both increase, these two users satisfy the exchange closed pair of the current match.

[0111] In some embodiments, the preset period is the length of one time slot.

[0112] In a second aspect, the present invention provides an offloading decision and resource allocation device based on integrated communication and sensing computing, including a processor and a storage medium;

[0113] The storage medium is used to store instructions;

[0114] The processor is used to operate according to the instructions to execute the steps of the method according to the first aspect.

[0115] In a third aspect, the present invention provides a storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the steps of the method according to the first aspect are implemented.

[0116] In a fourth aspect, the present invention provides a computing device, including,

[0117] one or more processors, one or more memories, and one or more programs, where one or more programs are stored in the one or more memories and are configured to be executed by the one or more processors, and the one or more programs include instructions for executing any of the methods in the method according to the first aspect.

[0118] Beneficial effects: Compared with the prior art, the advantages of the present invention include:

[0119] 1. In the MEC network with multiple users and multiple edge nodes of the present invention, considering that users have mobility, when a user switches from one base station to another, the user's service can be migrated to the computing node of the new base station for processing, and the task adopts a partial offloading method. During the communication process of task offloading, an integrated signal is used to analyze the echo of the integrated signal reflected by the surrounding environment to sense the environment. Under the constraints of the long-term perception failure rate of the user and the maximum task completion delay, a system cost function is defined by comprehensively considering energy consumption, delay, and migration cost, and a problem of minimizing the long-term average cost of the user is established.

[0120] 2. The established optimization problem is a long-term stochastic optimization problem. Based on the Lyapunov optimization theory, a virtual queue is established for the perception performance of the user, a relevant Lyapunov drift-plus-penalty function is defined, and the long-term stochastic optimization problem is transformed into a single-slot deterministic optimization problem through its upper bound.

[0121] 3. The single-slot problem is solved by dividing it into an inner layer and an outer layer. The inner layer obtains the optimal task offloading ratio under each user base station selection and channel allocation through theoretical derivation, and the outer layer designs a "preference value" to evaluate the current user base station selection and channel allocation according to the solution result of the inner layer problem, and iteratively converges to the final base station selection and channel allocation based on the matching theory. Description of the Drawings

[0122] Figure 1 is a schematic flowchart of the method according to an embodiment of the present invention;

[0123] Figure 2 is a schematic flowchart of the initial matching stage in an embodiment of the present invention;

[0124] Figure 3 is a schematic flowchart of the exchange matching stage in an embodiment of the present invention;

[0125] Figure 4 is a schematic diagram of the system model provided by an embodiment of the present invention. Detailed Embodiments

[0126] The present invention will be further described below with reference to the drawings. The following embodiments are only used to more clearly illustrate the technical solutions of the present invention and cannot be used to limit the protection scope of the present invention.

[0127] In the description of the present invention, the descriptions referring to terms such as "one embodiment", "some embodiments", "schematic embodiments", "examples", "specific examples", or "some examples" mean that the specific features, structures, materials, or characteristics described in connection with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic expressions of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials, or characteristics described can be combined in a suitable manner in any one or more embodiments or examples.

[0128] Embodiment 1

[0129] A method for offloading decision and resource allocation based on integrated communication, sensing, and computing, comprising:

[0130] S1. Obtain the location, channel state, task data volume, virtual queue backlog of each user in the current time slot, and the user - base station selection in the previous time slot; where each channel of each base station is denoted as a base - station - channel pair, and a user accessing a channel of a base station is denoted as the user matching with this base - station - channel pair;

[0131] S2. Based on the location, channel state, task data volume, virtual queue backlog of each user in the current time slot, and the user - base station selection in the previous time slot, perform an initial matching according to the preference value of the user for the base - station - channel pair and the preference value of the base - station - channel pair for the user to obtain an initial matching result;

[0132] S3. Perform an exchange matching based on the initial matching result to obtain an exchange matching result;

[0133] S4. Based on the exchange matching result, determine the user - base station selection, channel allocation, and task offloading ratio in the current time slot.

[0134] In some specific embodiments, as Figure 1 shown, a method for offloading decision and resource allocation based on integrated communication, sensing, and computing, comprising:

[0135] S1. Obtain the location, channel state, task data volume, virtual queue backlog of each user in the current time slot, and the user - base station selection in the previous time slot; where each channel of each base station is denoted as a base - station - channel pair, and a user accessing a channel of a base station is denoted as the user matching with this base - station - channel pair;

[0136] In some embodiments, obtaining the location, channel state, task data volume, virtual queue backlog of each user in the current time slot includes:

[0137] S1.1. As Figure 4 shown, there are N base stations, M users, and L channels in the system. The base - station set is denoted as The user set is represented as The channel set is represented as The user is in a mobile state. To represent the movement of the user, time is discretized into time slots, and the length of one time slot is τ. It is assumed that the position of the user remains unchanged within one time slot and is different in different time slots. The time slots are represented by the set T = {1, 2, 3, …, t, …}. It is assumed that each user has a service, and the service consists of a series of tasks. A dedicated virtual machine or container is set up on the edge server to process the generated tasks. As the user moves, the virtual machine or container will migrate as the base station connected by the user switches. The user generates a task in each time slot. The amount of data required for communication of the task generated by user m in time slot t is represented by I m (t), the task computing load is represented by C m (t), the maximum delay requirement of the task is represented by D m (t), and it is required that D m (t) ≤ τ, the task offloading ratio is represented by ρ m (t), and it is required that ρ m (t) ∈ [ρ min , 1], part of the task is computed locally and part is computed at the base station. The user connects to a base station in each time slot to offload task data. When the user transmits task data to the base station in each time slot, it can use the communication-sensing integrated signal to sense the surrounding environment. All base stations share the same multiple channels, and both the large-scale and small-scale fading of the channels are considered;

[0138] S1.2、 represents the association situation between user m and base station n in time slot t. When user m accesses base station n in time slot t otherwise it is 0; all base stations share L channels, and each channel has the same bandwidth, which is W. The signals of users using the same channel will interfere with each other; represents the usage situation of user m for channel l in time slot t. When user m occupies channel l in time slot t otherwise it is 0; the user can only access one base station and occupy one channel in one time slot;

[0139] S1.3、Use the conditional mutual information between the echo signal and the channel impulse response to measure the performance of the user's perception of the surrounding environment. The conditional mutual information MI m (t) of user m in time slot t is represented as:

[0140]

[0141] where represents the radar signal-to-dry ratio of user m on channel l in time slot t, which is represented as:

[0142]

[0143] where P m represents the transmit power of user m, K represents the number of symbols of the integrated signal, and T s represents the duration of one symbol, represents the Fourier transform of the channel impulse response at frequency f of channel l, which follows a standard normal distribution; l represents the channel gain between user m' and user m on channel l, and σ 2 =N0W represents the channel additive white Gaussian noise power;

[0144] When the mutual information of the user is less than the set threshold MI min , then the sensing fails, which is expressed as:

[0145]

[0146] where indicator {x} represents the indicator function, that is, when the expression x is true, the value of the indicator function is 1, otherwise it is 0;

[0147] S1.4. For the sensing failure event S m (t) of user m at time slot t, define a corresponding virtual queue with an initial value of 0, and represent the backlog of each queue as Q m (t), where Q m (0)=0, use S m (t) to represent the increase in the virtual queue at time slot t, and δ represents the decrease in the virtual queue; the backlog of the virtual queue of user m at time slots t and t + 1 satisfies:

[0148] Q m (t + 1)=[Q m (t)+S m (t)-δ] + .

[0149] S2. Based on the location, channel state, task data volume, virtual queue backlog of each user in the current time slot, and the user - base station selection in the previous time slot, perform an initial matching according to the preference value of the user for the base - station - channel pair and the preference value of the base - station - channel pair for the user to obtain an initial matching result;

[0150] In some embodiments, as Figure 2 shown, the initial matching according to the preference value of the user for the base - station - channel pair and the preference value of the base - station - channel pair for the user includes:

[0151] ​S2.1. Set the initial set of matched users to be an empty set, the initial set of unmatched users to be the set consisting of all users, the initial current match to be an empty set, the initial set of unmatched base station-channel pairs to be the set consisting of each channel of each base station, and the initial set of matched base station-channel pairs to be an empty set;

[0152] S2.2. Calculate, in a certain order, the preference values of users for all unmatched base station-channel pairs under the current match, and users send match requests to the unmatched base station-channel pair with the highest preference value;

[0153] S2.3. Calculate, in a certain order, the preference values of all unmatched base station-channel pairs for each requesting user under the current match, and the unmatched base station-channel pair accepts the request of the user with the highest preference value, and both sides reach an initial match;

[0154] S2.4. Update the current match, the set of matched users, the set of unmatched users, the unmatched base station-channel pairs, and the set of matched base station-channel pairs;

[0155] S2.5. Repeat steps S2.2 - S2.4 until all users are matched with a certain base station-channel pair, the initial matching stage ends, and an initial matching result is obtained, where the initial matching result includes: the current match, the set of matched users, the set of unmatched users, the unmatched base station-channel pairs, and the set of matched base station-channel pairs.

[0156] Among them, the calculation methods for the preference value of a user for a base station-channel pair and the preference value of a base station-channel pair for a user under the current match include:

[0157] The data transmission rate of user m unloading tasks to the base station in time slot t Is expressed as:

[0158]

[0159] Where Represents the channel gain between user m' and base station n on channel l, Represents the interference caused by other users using the same channel to user m;

[0160] The time for user m to transmit tasks in time slot t Is expressed as:

[0161]

[0162] The energy consumption of user m transmitting tasks in time slot t Is expressed as:

[0163]

[0164] Use fm,local Denotes the computing power of user m. The local computing delay of the task of user m in time slot t is:

[0165]

[0166] The energy consumption of user m in locally computing the task in time slot t is:

[0167]

[0168] where refers to the power coefficient;

[0169] After receiving the tasks transmitted by users, the base station evenly distributes all computing resources for the computing tasks of all users. Then, the computing resources allocated by base station n to user m in time slot t

[0170]

[0171] The computing resources f obtained by user m in time slot t m (t) is:

[0172]

[0173] The task computing delay of user m in time slot t is:

[0174]

[0175] The task processing delay T of user m in time slot t m (t) is the maximum of the local computing delay and the base station computing delay, expressed as:

[0176]

[0177] The energy consumption E of user m in processing the task in time slot t m (t) is:

[0178]

[0179] The migration cost of user m in time slot t is expressed as:

[0180]

[0181] where ε represents the migration cost and is a fixed value; when the base station selected by the user in this time slot is different from that in the previous time slot, there will be a migration cost of size ε, otherwise there is none;

[0182] The cost function U of user m in time slot t m(t) is expressed as:

[0183]

[0184] where α and β are the weight coefficients of time delay and energy consumption respectively;

[0185] An optimization problem P1 of establishing and minimizing the long-term cost of users is:

[0186]

[0187]

[0188]

[0189]

[0190]

[0191]

[0192]

[0193]

[0194] where the variable set represents the access situation of each user to each base station in each time slot, represents the occupancy situation of each user for each channel in each time slot, represents the task division ratio of each user in each time slot; Constraints (C1) and (C2) mean that a user can only select one base station to access and can only occupy one channel in each time slot; Constraint (C3) means that at most one user can use one channel of one base station in one time slot; Constraint (C4) means that the time average of the user's perception failure event expectation E[S m (t)] is less than a certain set value δ, that is, the long-term perception failure rate of the user is below a certain threshold; Constraint (C5) means that the task of each user must be completed within the maximum time delay D m (t) requirement; Constraint (C6) means that the task of each user in each time slot is randomly divided between the minimum value ρ min and 1; Constraint (C7) means that the base station selection and channel allocation variables are binary variables;

[0195] The optimization problem P1 is transformed into an optimization problem P2 through the Lyapunov optimization theory:

[0196]

[0197]

[0198]

[0199]

[0200]

[0201]

[0202]

[0203] Among them, Q m (t) is the virtual queue backlog of user m at time slot t, and S m (t) is the sensing failure event of user m at time slot t, is a constant, and V is the control parameter for balancing the virtual queue stability and the sum of the user cost functions;

[0204] Solve the optimization problem P2 to obtain the base station selection corresponding to the current match of the user and channel allocation value;

[0205] When the base station selection of the user and channel allocation are determined, the optimization problem P2 is simplified to the optimization problem P2.1

[0206]

[0207]

[0208]

[0209] Solve the optimization problem P2.1 to obtain the user task offloading ratio under the current match;

[0210] Let F0 = 1 - D m (t)f m,local / C m (t), When F0 > F1 or F1 < ρ min at this time, the problem P2.1 has no solution. Conversely, the optimal solution of P2.1, that is, the user task offloading ratio is:

[0211]

[0212] Among them

[0213]

[0214]

[0215]

[0216]

[0217]

[0218]

[0219]

[0220] According to the obtained user task offloading ratio, calculate the preference value of the user for the base station-channel pair and the preference value of the base station-channel pair for the user under the current matching;

[0221] User The preference value of the user for the base station-channel pair under the current matching v at time slot t is:

[0222]

[0223] where, Iv m (t) = 1 indicates that the inner problem of user m has a solution at time slot t, that is, the maximum delay constraint can be satisfied. Conversely, Iv m (t) takes 0, and η is a positive number with a very large order of magnitude;

[0224] The preference value of the base station-channel pair for the user under the current matching v at time slot t is:

[0225]

[0226] S3. Perform exchange matching based on the initial matching result to obtain the exchange matching result;

[0227] In some embodiments, as Figure 3 shown, performing exchange matching specifically includes:

[0228] S3.1. Record the current matching as the pre-exchange matching, and start a round of exchange matching;

[0229] S3.2. Detect the exchange-closed pairs of the users currently matched by all base station-channel pairs in a certain order;

[0230] S3.3. If any two users satisfy the exchange-closed pair, exchange the base station-channel pairs they are matched with;

[0231] S3.4. Update the current matching;

[0232] S3.5. After all users are detected, a round of exchange matching ends, and record the current matching as the post-exchange matching;

[0233] S3.6. Repeat steps S3.1 - S3.5 until the matches before and after each round of exchange are the same, indicating the completion of the exchange match and obtaining the exchange match result.

[0234] S4. Based on the exchange match result, determine the user - base station selection, channel allocation, and task offloading ratio for the current time slot.

[0235] Further, in step S3.2, the method for detecting exchange - closed pairs includes:

[0236] Under the current match, assume that two users exchange the base - station - channel pairs of the match, and other users maintain the base - station - channel pairs of the current match.

[0237] If, after the two users exchange the base - station - channel pairs of the match, the preference values of the users for the base - station - channel pairs and the preference values of the base - station - channel pairs for the users both increase, then these two users satisfy the exchange - closed pairs of the current match.

[0238] It should be specifically noted that, in some embodiments, the preset period is the length of one time slot.

[0239] An offloading decision and resource allocation method based on integrated communication, sensing, and computing provided by an embodiment of the present invention comprehensively considers energy consumption, delay, and migration cost to define a system cost function. Under the constraints of the user sensing failure rate and the maximum task completion delay, a problem of minimizing the long - term average cost of users is established. Based on the Lyapunov optimization theory, a virtual queue is established for the sensing performance of users, and the long - term stochastic optimization problem is transformed into a single - time - slot deterministic optimization problem by using the Lyapunov drift - plus - penalty function. The transformed problem is divided into an inner layer and an outer layer. The inner layer obtains the optimal task offloading ratio under each user - base - station selection and channel allocation through theoretical derivation. The outer layer determines the user - base - station selection and channel allocation according to the results of solving the inner - layer problem through the matching theory. While reducing the task completion delay, it effectively reduces the user energy consumption and also ensures the sensing performance.

[0240] Embodiment 2

[0241] Second, this embodiment provides an offloading decision and resource allocation device based on integrated communication, sensing, and computing, including a processor and a storage medium;

[0242] The storage medium is used to store instructions;

[0243] The processor is used to operate according to the instructions to execute the steps of the method according to Embodiment 1.

[0244] Embodiment 3

[0245] In a third aspect, this embodiment provides a storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the steps of the method described in Embodiment 1 are implemented.

[0246] Embodiment 4

[0247] In a fourth aspect, the present invention provides a computing device, including,

[0248] 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 are configured to be executed by the one or more processors, and the one or more programs include instructions for performing any of the methods described in the method of Embodiment 1.

[0249] Those skilled in the art should understand that the embodiments of the present application can be provided as a method, a system, or a computer program product. Therefore, the present application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present application can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk memories, CD-ROMs, optical memories, etc.) containing computer-usable program code.

[0250] The present application is described with reference to the flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each flow and / or block in the flowcharts and / or block diagrams, as well as the combination of flows and / or blocks in the flowcharts and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to the processors of general-purpose computers, special-purpose computers, embedded processors, or other programmable data processing devices to generate a machine, so that the instructions executed by the processors of the computer or other programmable data processing devices generate means for implementing the specified functions in Figure 1 one flow or multiple flows and / or blocks Figure 1 one block or multiple blocks.

[0251] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer-readable memory generate a manufactured article including instruction means, and the instruction means implements the specified functions in Figure 1 one flow or multiple flows and / or blocks Figure 1 one block or multiple blocks.

[0252] These computer program instructions can also be loaded onto a computer or other programmable data processing apparatus, so that a series of operation steps are executed on the computer or other programmable apparatus to produce a computer-implemented process, thereby the instructions executed on the computer or other programmable apparatus provide steps for realizing the functions specified in one process or a plurality of processes and / or blocks Figure 1 one process or a plurality of processes and / or blocks Figure 1 steps for realizing the functions specified in one block or a plurality of blocks.

[0253] The embodiments of the present invention have been described in detail above in conjunction with the accompanying drawings. However, the present invention is not limited to the above embodiments, and various changes can be made without departing from the spirit of the present invention within the scope of knowledge possessed by those of ordinary skill in the art.

Claims

1. A method for offloading decision-making and resource allocation based on integrated communication and sensing computing, characterized in that, Execute the following steps at a preset period: S1. Obtain the location, channel state, task data volume, virtual queue backlog of each user in the current time slot, and the user-base station selection in the previous time slot; Wherein each channel of each base station is recorded as a base station-channel pair, and a user accessing a channel of a base station is recorded as the user matching with this base station-channel pair; S2. Based on the location, channel state, task data volume, virtual queue backlog of each user in the current time slot, and the user-base station selection in the previous time slot, perform an initial matching according to the preference value of the user for the base station-channel pair and the preference value of the base station-channel pair for the user to obtain an initial matching result; S3. Perform a swapping matching based on the initial matching result to obtain a swapping matching result; S4. Based on the swapping matching result, determine the user-base station selection, channel allocation, and task offloading ratio in the current time slot.

2. The offloading decision and resource allocation method based on integrated communication and sensing computing according to claim 1, wherein S2. Perform an initial matching according to the preference value of the user for the base station-channel pair and the preference value of the base station-channel pair for the user, including: S2.

1. Set the initial set of matched users to be an empty set, the initial set of unmatched users to be the set composed of all users, the initial current matching to be an empty set, the initial set of unmatched base station-channel pairs to be the set composed of each channel of each base station, and the initial set of matched base station-channel pairs to be an empty set; S2.

2. Calculate the preference value of the user for all unmatched base station-channel pairs under the current matching in a certain order, and the user sends a matching request to the unmatched base station-channel pair with the highest preference value; S2.

3. Calculate the preference value of all unmatched base station-channel pairs for each requesting user under the current matching in a certain order, and the unmatched base station-channel pair accepts the request of the user with the highest preference value, and both sides reach an initial matching; S2.

4. Update the current matching, the set of matched users, the set of unmatched users, the unmatched base station-channel pairs, and the set of matched base station-channel pairs; S2.

5. Repeat steps S2.2 - S2.4 until all users are matched with a certain base station-channel pair, the initial matching stage ends, and an initial matching result is obtained, where the initial matching result includes: the current matching, the set of matched users, the set of unmatched users, the unmatched base station-channel pairs, and the set of matched base station-channel pairs.

3. The offloading decision and resource allocation method based on integrated communication and sensing computing according to claim 1, characterized in that, S3. Perform a swapping matching based on the initial matching result to obtain a swapping matching result, including: S3.

1. Record the current matching as the pre-swapping matching, and start a round of swapping matching; S3.

2. Detect the swapping closed pairs of the users currently matched by all base station-channel pairs in a certain order; S3.

3. If any two users satisfy the swapping closed pair, swap the base station-channel pairs they are matched with; S3.

4. Update the current matching; S3.

5. After all users are detected, a round of swapping matching ends, and record the current matching as the post-swapping matching; S3.

6. Repeat steps S3.1 - S3.5 until the pre-swapping matching and the post-swapping matching in each round are the same, the swapping matching is completed, and a swapping matching result is obtained.

4. The offloading decision and resource allocation method based on integrated communication and sensing computing according to claim 1, wherein In the step S1, obtaining the location, channel state, task data volume, and virtual queue backlog of each user in the current time slot includes: S1.

1. There are N base stations, M users, and L channels in the system. The base station set is denoted as The user set is denoted as The channel set is denoted as Users are in a mobile state. To represent the movement of users, time is discretized into time slots, and the length of one time slot is τ. It is assumed that the position of a user remains unchanged within one time slot and is different in different time slots. The time slot is represented by the set T = {1, 2, 3, …, t, …}. It is assumed that each user has a service, and the service consists of a series of tasks. There is a dedicated virtual machine or container built on the edge server to process the generated tasks. As the user moves, the virtual machine or container will migrate as the base station connected by the user switches. Each user generates a task in each time slot. The amount of communication data required for the task generated by user m in time slot t is denoted by I m (t), the task computing load is denoted by C m (t), the maximum delay requirement of the task is denoted by D m (t), and it is required that D m (t) ≤ τ, the task offloading ratio is denoted by ρ m (t), and it is required that ρ m (t) ∈ [ρ min , 1], part of the task is computed locally and part is computed at the base station. Each user connects to a base station in each time slot to offload task data. When each user transmits task data to the base station in each time slot, it can use the communication-sensing integrated signal to sense the surrounding environment. All base stations share the same multiple channels, and both the large-scale and small-scale fading of the channels are considered. S1.2, Indicates the association between user m and base station n at time slot t. When user m accesses base station n at time slot t Otherwise it is 0; All base stations share L channels, and each channel has the same bandwidth of W. Signals of users using the same channel will interfere with each other; Indicates the usage of channel l by user m at time slot t. When user m occupies channel l at time slot t Otherwise it is 0; A user can only access one base station and occupy one channel in one time slot; S1.

3. Measure the performance of the user's perception of the surrounding environment using the conditional mutual information between the echo signal and the channel impulse response. The conditional mutual information MI m (t) of user m at time slot t is expressed as: Among them represents the radar signal-to-interference-plus-noise ratio of user m on channel l at time slot t, expressed as: where P m represents the transmit power of user m, K represents the number of symbols of the integrated signal, and T s represents the duration of one symbol, represents the Fourier transform of the channel impulse response at frequency f of channel l, following a standard normal distribution; l and represents the channel gain between user m' and user m on channel l, and σ 2 = N0W represents the channel additive white Gaussian noise power; When the user mutual information is less than the set threshold value MI min then the sensing fails, which is expressed as: where indicator {x} represents the indicator function, that is, when the expression x is true, the value of the indicator function is 1, otherwise it is 0; S1.

4. Define a corresponding virtual queue with an initial value of 0 for the sensing failure event S of user m at time slot t, and represent the backlog of each queue as Q m (t), where Q m (t) satisfies Q m (0) = 0. Use S m (t) to represent the increase in the virtual queue at time slot t, and δ to represent the decrease in the virtual queue. The backlog of the virtual queue of user m at time slots t and t + 1 satisfies: Q m (t + 1) = [Q m (t) + S m (t) - δ] + 。 5. The offloading decision and resource allocation method based on integrated communication and sensing calculation according to claim 4, characterized in that, In the step S2, the calculation methods for the preference value of the user for the base station-channel pair and the preference value of the base station-channel pair for the user in the current matching include: The data transmission rate at which user m offloads tasks to the base station in time slot t It is expressed as: wherein represents the channel gain between user m' and base station n on channel l, represents the interference caused by other users using the same channel to user m; The time when user m transmits a task in time slot t It is expressed as: The energy consumption of user m for transmitting tasks in time slot t It is expressed as: Use f m,local to represent the computing power of user m. The local computing delay of the task of user m in time slot t is as follows: The energy consumption of user m for the local computing task in time slot t is as follows: wherein refers to the power coefficient; After receiving the tasks transmitted by the users, the base station evenly distributes all computing resources among all users' computing tasks. Then, the computing resources allocated by base station n to user m at time slot t are as follows: The computing resources f obtained by user m at time slot t m (t) is as follows: The task computing delay of user m at time slot t is as follows: The task processing delay \(T\) of user \(m\) at time slot \(t\) m (t) is the maximum of the local computing delay and the base station computing delay, expressed as: The energy consumption E of user m for processing tasks in time slot t m (t) is as follows: The migration cost of user m in time slot t is expressed as: where ε represents the migration cost, which is a fixed value; when the base station selected by the user in this time slot is different from that in the previous time slot, there will be a migration cost of size ε, otherwise there is none; The cost function U of user m at time slot t m is expressed as: where α and β are the weight coefficients of delay and energy consumption respectively; Establish an optimization problem P1 for minimizing the long-term cost of the user: Among them, the variable set represents the access situation of each user to each base station in each time slot, represents the occupancy situation of each user for each channel in each time slot, represents the task division ratio of each user in each time slot; Constraints (C1) and (C2) mean that a user can only select one base station to access and can only occupy one channel in each time slot; Constraint (C3) means that at most one user can use one channel of a base station in one time slot; Constraint (C4) means that the time average of the user's sensing failure event expectation E[S m (t)] should be less than a certain set value δ, that is, the long-term sensing failure rate of the user should be below a certain threshold; Constraint (C5) means that the task of each user should be completed within the maximum delay D m (t) requirement; Constraint (C6) means that the task of each user in each time slot is randomly divided between the minimum value ρ min and 1; Constraint (C7) means that the base station selection and channel allocation variables are both binary variables; Transform the optimization problem P1 into an optimization problem P2 through the Lyapunov optimization theory: Among them, Q m (t) is the virtual queue backlog of user m at time slot t, and S m (t) is the sensing failure event of user m at time slot t, is a constant, and V is the control parameter for balancing the virtual queue stability and the sum of the user cost functions; Solve the optimization problem P2 to obtain the base station selection corresponding to the current match of the user and channel allocation value; When the user's base station selection and channel allocation are determined, the optimization problem P2 is simplified to the optimization problem P2.1 Solve the optimization problem P2.1 to obtain the user task offloading ratio in the current matching; Let F0 = 1 - D m (t)f m,local / C m (t), When F0 > F1 or F1 < ρ min then, problem P2.1 has no solution. Conversely, the optimal solution of P2.1, i.e., the user task offloading ratio is: where According to the obtained user task offloading ratio, calculate the preference value of the user for the base station-channel pair and the preference value of the base station-channel pair for the user in the current matching; User Preference value for the base station-channel pair under the current match v at time slot t is as follows: Among them, Iv m (t) = 1 indicates that the inner problem of user m has a solution at time slot t, that is, the maximum delay constraint can be satisfied. Conversely, Iv m (t) takes 0, and η is a positive number with an extremely large order of magnitude; The preference value of the base station-channel pair for the user under the current match v at time slot t is as follows:

6. The offloading decision and resource allocation method based on integrated communication and sensing computing according to claim 1, characterized in that In the step S3.2, the detection method for the exchange closed pair includes: In the current matching, assume that two users exchange the matched base station-channel pairs, and other users keep the currently matched base station-channel pairs; If after the two users exchange the matched base station-channel pairs, the preference value of the user for the base station-channel pair and the preference value of the base station-channel pair for the user both increase, these two users satisfy the exchange closed pair of the current matching.

7. The offloading decision and resource allocation method based on integrated communication and sensing computing according to claim 1, characterized in that The preset period is the length of one time slot.

8. An offloading decision and resource allocation device based on integrated communication and sensing computing, characterized in that, Including a processor and a storage medium; The storage medium is used to store instructions; The processor is used to operate according to the instructions to execute the steps of the method according to any one of claims 1 to 7.

9. A storage medium, on which a computer program is stored, characterized in that, When the computer program is executed by the processor, it implements the steps of the method according to any one of claims 1 to 7.

10. A computing device, characterized in that: Including, One or more processors, one or more memories, and one or more programs, where one or more programs are stored in the one or more memories and are configured to be executed by the one or more processors, and the one or more programs include instructions for executing any one of the methods according to claims 1 to 7.

Images