Noma-based sensing data collection and computation offloading method

Abstract

The application provides a kind of based on NOMA's perception data acquisition and computing offloading method, by establishing the optimization problem model based on NOMA's perception data acquisition and computing offloading with the goal of maximizing offloading data volume;Simplify optimization problem model;The simplified optimization problem model is converted into two-layer optimization problem of inner and outer nesting;Get terminal scheduling variables and acquisition and offloading time allocation variables;According to terminal scheduling variables, schedule acquisition terminals in turn according to acquisition and offloading time allocation variables within the specified time based on NOMA for data offloading, and the acquisition terminals that have not been scheduled continue to carry out data acquisition;Base station calculates data according to the calculation resource allocation within the set time;Compared with the existing method, the method can effectively improve the data acquisition and computing efficiency, can maximize the offloading data volume, and the performance improvement is more obvious when the total computing resource is larger.

Description

Technical Field

[0001] This invention relates to a method for sensor data acquisition and computation offloading based on NOMA. Background Technology

[0002] Wireless communication technology has become a research focus in recent years, and the issues of sensor data acquisition and computation offloading in wireless communication are also hot research topics. The rapid development of the Internet of Things (IoT) in wireless communication has spurred the emergence of many new services, such as smart homes, smart fire protection, and smart cities. However, the performance of IoT terminals is generally limited by size and power consumption. Therefore, if the data collected by IoT terminals is processed locally, it is difficult to meet the service quality requirements of various emerging IoT services.

[0003] Therefore, mobile edge computing can effectively overcome the problem of insufficient terminal computing power by providing computing services to terminals at the network edge. Thus, it is necessary to offload the data collected by the terminal to edge servers at the network edge for computation and processing.

[0004] Large-scale data offloading by terminals involves terminal scheduling and access resource allocation issues. Furthermore, Non-Orthogonal Multiple Access (NOMA) technology can significantly improve data offloading efficiency by allowing multiple terminals to utilize the same time and frequency resources.

[0005] However, traditional NOMA technology performs data acquisition and offloading sequentially, which cannot significantly improve the overall efficiency of data acquisition and offloading; its performance is even lower than that of Time Division Multiple Access (TDMA) schemes. In conclusion, to improve the efficiency of terminal data acquisition and computation offloading, it is necessary to optimize terminal scheduling, time allocation, and computational resource allocation.

[0006] The above-mentioned issues should be considered and resolved during the NOMA-based sensing data acquisition and computation offloading process. Summary of the Invention

[0007] The purpose of this invention is to provide a NOMA-based method for sensing data acquisition and computation offloading to address the problem that the overall efficiency of data acquisition and offloading in existing technologies needs to be improved.

[0008] The technical solution of this invention is:

[0009] A NOMA-based method for sensing data acquisition and computation offloading is applied to a data acquisition system in the Internet of Things (IoT). The data acquisition system includes a base station equipped with a mobile edge computing server and several acquisition terminals, and includes the following steps:

[0010] S1. Establish an optimization problem model for sensing data acquisition and computation unloading based on NOMA with the goal of maximizing the amount of unloaded data. The optimization variables of the optimization problem model are terminal scheduling variables, computation resource allocation variables, and computation time allocation variables. Among them, the computation time allocation variables include acquisition and unloading time and computation time.

[0011] S2. By transforming the computational resource allocation variables and computational time allocation variables into functions of terminal scheduling variables and acquisition and unloading time allocation variables, the optimization problem model is simplified. The optimization variables of the simplified optimization problem model are terminal scheduling variables and acquisition and unloading time allocation variables.

[0012] S3. The simplified optimization problem model is transformed into a two-layer nested optimization problem, in which the outer optimization problem optimizes the terminal scheduling variable, and the inner optimization problem optimizes the acquisition and unloading time allocation variable.

[0013] S4. Solve the two-layer optimization problem to obtain the terminal scheduling variables and the acquisition and unloading time allocation variables, respectively;

[0014] S5. Based on the data acquisition and unloading time allocation variables, all acquisition terminals are made to acquire data within a specified time. Then, according to the terminal scheduling variables, the acquisition terminals are scheduled to unload data in sequence according to the data acquisition and unloading time allocation variables within a specified time based on NOMA. Acquisition terminals that have not yet been scheduled continue to acquire data.

[0015] S6. According to the calculation time allocation variable and the calculation resource allocation variable, the base station performs calculations on the data according to the calculation resource allocation within the set time.

[0016] Furthermore, in step S1, an optimization problem model P1 based on NOMA for sensing data acquisition and computational unloading is established with the objective of maximizing the amount of unloaded data:

[0017] P1:

[0018] in, For all possible permutations of the terminal scheduling variable π, Assign variables to the collection and unloading times. For the first The time slot length is , and c is a variable for calculating resource allocation. The number of data collection terminals. For serial number The data acquisition speed of the acquisition terminal The length of the (i+1)th time slot;

[0019] Constraint C1, used to indicate that the data collected by each acquisition terminal must be offloaded to the base station: ;

[0020] Constraint C2 represents the computing power allocated to all data acquisition terminals and cannot exceed the total computing power of the mobile edge computing servers equipped at the base station. ,in, Assigning sequence numbers to base stations The number of computing resources for the data acquisition terminal. The total computing power of the base station;

[0021] Constraint C3 is used to indicate that the data collected by each acquisition terminal must be calculated at the base station: ,in, Assigning data to the base station to the data acquisition terminal Number of computing resources For the first Each time slot length;

[0022] Used to represent The constraint C4 is that the sum of the time slots cannot exceed the total time. The entire data acquisition, unloading, and computation time is divided into several parts. One time slot: The length of the first time slot is Data is collected from all acquisition terminals. For the first Each time slot length , Assigned to serial number The acquisition terminal performs data offloading and sequence numbering based on NOMA. The data acquisition terminal collects data, the first Each time slot length Allocate data to base stations for data processing. The total time for data acquisition and unloading calculation for all acquisition terminals;

[0023] Optimization variables: Terminal scheduling variables ,in, For serial number The data acquisition terminals are sorted according to a set sorting method; data acquisition and unloading time allocation variables. Calculate the time allocation variable, i.e., the first Each time slot length Calculate resource allocation variables ,in, Assigning sequence numbers to base stations The number of computing resources of the data acquisition terminal.

[0024] Further, in step S1, the sequence number The acquisition terminal at the unloading rate per time slot for:

[0025]

[0026] Where, i= , For serial number The transmission power of the acquisition terminal, For serial number The wireless channel gain from the data acquisition terminal to the base station, This represents noise power.

[0027] Furthermore, in step S2, the optimization problem model is simplified by transforming the computational resource allocation variables and computational time allocation variables into functions of terminal scheduling variables and acquisition and unloading time allocation variables. The optimization variables in the simplified optimization problem model are the terminal scheduling variables and the acquisition and unloading time allocation variables, specifically...

[0028] S21. By ensuring that constraints C2 and C3 hold true, the base station allocates data to the data acquisition terminal. Number of computing resources and the Each time slot length The expression is:

[0029]

[0030] ;

[0031] in, The number of data collection terminals. For serial number The data acquisition speed of the acquisition terminal The length of the (i+1)th time slot The total computing power of the base station;

[0032] S22, and then the optimization problem model P1 is simplified into the simplified optimization problem model P2:

[0033] P2:

[0034] in, For all possible permutations of the terminal scheduling variable π, The number of data collection terminals. For serial number The data acquisition speed of the acquisition terminal The length of the (i+1)th time slot;

[0035] Constraint C1: ;

[0036] Constraint C5: ;

[0037] Optimization variables: Terminal scheduling variables Data collection and unloading time allocation variables .

[0038] Further, in step S4, the terminal scheduling variables and the acquisition and unloading time allocation variables are obtained by solving the two-layer optimization problems. Specifically, the terminal scheduling variables are obtained by solving the outer-layer optimization problem using a one-dimensional search method. Solve for the given terminal scheduling variables The inner-layer optimization problem under the given conditions is solved using linear programming to obtain the acquisition and unloading time allocation variables. .

[0039] Furthermore, in step S4, the terminal scheduling variables are obtained by solving the outer optimization problem using a one-dimensional search method. Solve for the given terminal scheduling variables The inner-layer optimization problem under the given conditions is solved using linear programming to obtain the acquisition and unloading time allocation variables. Specifically,

[0040] S41. Initialize search step size The number of searches during a one-dimensional search ;

[0041] S42. Let the weighting factor for the tradeoff between acquisition rate and channel gain be... equal The smaller of the two;

[0042] S43, Set the terminal scheduling variable For the data acquisition terminal The order is descending from largest to smallest, where... The acquisition rate is normalized to the maximum acquisition rate. It is the reciprocal of the channel gain normalized to the minimum channel gain;

[0043] S44. Solve for the given terminal scheduling variables. The inner-layer optimization problem under the condition of value is obtained by allocating variables for data acquisition and unloading time. And obtain the objective function value of the inner optimization problem;

[0044] S45, if If so, proceed to the next step S46; otherwise, let And return to step S42;

[0045] S46. Select the largest objective function value from all obtained objective function values, and set the terminal scheduling variable corresponding to the largest objective function value as... And the allocation of variables for collection and unloading time This is the final solution to the simplified optimization problem model P2.

[0046] Further, in step S5, all acquisition terminals are configured to acquire data within a specified time according to the acquisition and unloading time allocation variables; then, according to the terminal scheduling variables, the acquisition terminals are scheduled to unload data sequentially within the specified time based on NOMA, while the acquisition terminals that have not yet been scheduled continue to acquire data. Specifically,

[0047] S51, ensure all acquisition terminals are in the first time slot length Data collection is conducted internally;

[0048] S52, the sequence number of the i-th scheduled item. data acquisition terminal The length of the (k+1)th time slot and Data offloading based on NOMA, and unscheduled acquisition terminals. Data collection is performed;

[0049] S53, All acquisition terminals are based on NOMA in the (K+1)th time slot length. Data is unloaded internally.

[0050] Further, in step S6, according to the calculation time allocation variable and the calculation resource allocation variable, the base station performs calculations on the data according to the allocated calculation resources within a set time period. Specifically,

[0051] S61, Calculate the resource allocation variables for base stations and data acquisition terminals. Number of computing resources : ,in, The number of data collection terminals. For serial number The data acquisition speed of the acquisition terminal The length of the (i+1)th time slot The total computing power of the base station;

[0052] S62, Base station allocates variables during calculation time, i.e., the... Each time slot length Data calculations are performed within a given time period, where... ,in, For all possible permutations of the terminal scheduling variable π, The number of data collection terminals. For serial number The data acquisition speed of the acquisition terminal It is the length of the (i+1)th time slot.

[0053] The beneficial effects of this invention are as follows: This NOMA-based sensing data acquisition and computation offloading method effectively improves data acquisition and computation efficiency by optimizing terminal scheduling, time allocation, and computing resource allocation, thereby maximizing the amount of data offloaded. Addressing the drawback of traditional NOMA technology where data acquisition and offloading are performed sequentially, this method allows some terminals to offload data while others continue data acquisition, effectively improving the overall efficiency of data acquisition and computation offloading and achieving a higher amount of offloaded data. Experimental simulations verify that, compared with existing methods, this NOMA-based sensing data acquisition and computation offloading method achieves higher offloading throughput, with a more significant performance improvement when total computing resources are larger. Attached Figure Description

[0054] Figure 1 This is a flowchart illustrating the sensing data acquisition and computation offloading method based on NOMA according to an embodiment of the present invention.

[0055] Figure 2 This is an illustrative diagram illustrating the data acquisition system in the Internet of Things (IoT) embodiment.

[0056] Figure 3 This is a schematic diagram illustrating the timing of data acquisition and computational unloading at each acquisition terminal in the embodiment.

[0057] Figure 4 The embodiment of the NOMA-based sensing data acquisition and computation offloading method differs from existing methods in that the amount of offloaded data increases with the total computing resources. A diagram illustrating the changes. Detailed Implementation

[0058] The preferred embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0059] like Figure 1 This embodiment provides a NOMA-based method for sensing data acquisition and computation offloading, applied to a data acquisition system in the Internet of Things (IoT). The data acquisition system includes a base station equipped with a mobile edge computing server and several acquisition terminals, such as... Figure 2 This includes the following steps:

[0060] S1. Establish an optimization problem model for sensing data acquisition and computation unloading based on NOMA with the goal of maximizing the amount of unloaded data. The optimization variables of the optimization problem model are terminal scheduling variables, computation resource allocation variables, and computation time allocation variables. Among them, the computation time allocation variables include acquisition and unloading time and computation time.

[0061] In step S1, an optimization problem model P1 based on NOMA for sensing data acquisition and computational unloading is established with the objective of maximizing the amount of unloaded data.

[0062] P1:

[0063] in, For all possible permutations of the terminal scheduling variable π, Assign variables to the collection and unloading times. For the first The time slot length is , and c is a variable for calculating resource allocation. The number of data collection terminals. For serial number The data acquisition speed of the acquisition terminal The length of the (i+1)th time slot;

[0064] Constraint C1, used to indicate that the data collected by each acquisition terminal must be offloaded to the base station: ;

[0065] Constraint C2 represents the computing power allocated to all data acquisition terminals and cannot exceed the total computing power of the mobile edge computing servers equipped at the base station. ,in, Assigning sequence numbers to base stations The number of computing resources for the data acquisition terminal. The total computing power of the base station;

[0066] Constraint C3 is used to indicate that the data collected by each acquisition terminal must be calculated at the base station: ,in, Assigning data to the base station to the data acquisition terminal Number of computing resources For the first Each time slot length;

[0067] Used to represent The constraint C4 is that the sum of the time slots cannot exceed the total time. The entire data acquisition, unloading, and computation time is divided into several parts. One time slot: The length of the first time slot is Data is collected from all acquisition terminals. For the first Each time slot length , Assigned to serial number The acquisition terminal performs data offloading and sequence numbering based on NOMA. The data acquisition terminal collects data, the first Each time slot length Allocate data to base stations for data processing. The total time for data acquisition and unloading calculation for all acquisition terminals;

[0068] Optimization variables: Terminal scheduling variables ,in, For serial number The data acquisition terminals are sorted according to a set sorting method; data acquisition and unloading time allocation variables. Calculate the time allocation variable, i.e., the first Each time slot length Calculate resource allocation variables ,in, Assigning sequence numbers to base stations The number of computing resources of the acquisition terminal.

[0069] In step S1, the sequence number The acquisition terminal in the first unloading rate per time slot for:

[0070]

[0071] Where, i= , For serial number The transmission power of the acquisition terminal, For serial number The wireless channel gain from the data acquisition terminal to the base station, This represents noise power.

[0072] S2. By transforming the computational resource allocation variables and computational time allocation variables into functions of terminal scheduling variables and acquisition and unloading time allocation variables, the optimization problem model is simplified. The optimization variables of the simplified optimization problem model are terminal scheduling variables and acquisition and unloading time allocation variables.

[0073] In step S2, the optimization problem model is simplified by transforming the computational resource allocation variables and computational time allocation variables into functions of terminal scheduling variables and acquisition and unloading time allocation variables. The simplified optimization problem model then uses terminal scheduling variables and acquisition and unloading time allocation variables as its optimization variables. Specifically,

[0074] S21. By ensuring that constraints C2 and C3 hold true, the base station allocates data to the data acquisition terminal. Number of computing resources and the Each time slot length The expression is:

[0075]

[0076] ;

[0077] in, The number of data collection terminals. For serial number The data acquisition speed of the acquisition terminal The length of the (i+1)th time slot The total computing power of the base station;

[0078] S22, and then the optimization problem model P1 is simplified into the simplified optimization problem model P2:

[0079] P2:

[0080] in, For all possible permutations of the terminal scheduling variable π, The number of data collection terminals. For serial number The data acquisition speed of the acquisition terminal The length of the (i+1)th time slot;

[0081] Constraint C1: ;

[0082] Constraint C5: ;

[0083] Optimization variables: Terminal scheduling variables Data collection and unloading time allocation variables .

[0084] S3. The simplified optimization problem model is transformed into a two-layer nested optimization problem, in which the outer optimization problem optimizes the terminal scheduling variable, and the inner optimization problem optimizes the acquisition and unloading time allocation variable.

[0085] S4. Solve the two-layer optimization problem to obtain the terminal scheduling variables and the acquisition and unloading time allocation variables, respectively;

[0086] In step S4, the terminal scheduling variables and the acquisition and unloading time allocation variables are obtained by solving the two-layer optimization problems. Specifically, the terminal scheduling variables are obtained by solving the outer-layer optimization problem using a one-dimensional search method. Solve for the given terminal scheduling variables The inner-layer optimization problem under the given conditions is solved using linear programming methods (such as the simplex method) to obtain the acquisition and unloading time allocation variables. .

[0087] In step S4, the terminal scheduling variables are obtained by solving the outer optimization problem using a one-dimensional search method. Solve for the given terminal scheduling variables The inner-layer optimization problem under the given conditions is solved using linear programming to obtain the acquisition and unloading time allocation variables. Specifically,

[0088] S41. Initialize search step size The number of searches during a one-dimensional search ;

[0089] S42. Let the weighting factor for the tradeoff between acquisition rate and channel gain be... equal The smaller of the two;

[0090] S43, Set the terminal scheduling variable For the data acquisition terminal The order is descending from largest to smallest, where... The acquisition rate is normalized to the maximum acquisition rate. It is the reciprocal of the channel gain normalized to the minimum channel gain;

[0091] In step S43, Smaller data acquisition terminals should be scheduled later, i.e., in The middle position is towards the back, because this allows for a longer collection time and results in more data being collected. Larger data acquisition terminals should be scheduled earlier, i.e., in The middle position is near the front, because The larger the value, the lower the unloading rate. Earlier scheduling allows more time for unloading, resulting in more data being unloaded. This is achieved by weighting these two factors. The scheduling variables are determined jointly in a manner that allows for their determination. .

[0092] S44. Solve for the given terminal scheduling variables. The inner-layer optimization problem under the condition of value is obtained by allocating variables for data acquisition and unloading time. And obtain the objective function value of the inner optimization problem;

[0093] S45, if If so, proceed to the next step S46; otherwise, let And return to step S42;

[0094] S46. Select the largest objective function value from all obtained objective function values, and set the terminal scheduling variable corresponding to the largest objective function value as... And the allocation of variables for collection and unloading time This is the final solution to the simplified optimization problem model P2.

[0095] S5. Based on the data acquisition and unloading time allocation variables, ensure all acquisition terminals acquire data within the specified time; then, according to the terminal scheduling variables, schedule the acquisition terminals to sequentially unload data based on NOMA within the specified time according to the data acquisition and unloading time allocation variables, while acquisition terminals that have not yet been scheduled continue data acquisition; for example... Figure 3 :

[0096] S51, ensure all acquisition terminals are in the first time slot length Data collection is conducted internally;

[0097] S52, the sequence number of the i-th scheduled item. data acquisition terminal The length of the (k+1)th time slot and Data offloading based on NOMA, and unscheduled acquisition terminals. Data collection is performed;

[0098] In step S52, taking i=1 and i=2 as examples, the explanation is as follows: Figure 3 When i=1, the data acquisition terminal In the Each time slot length Data is unloaded from the internal data collection terminal. Continue data collection; when i=2, the data collection terminal... In the Each time slot length Internally, data offloading and collection are performed based on NOMA, and the collection terminal... Continue data collection.

[0099] S53, All acquisition terminals are based on NOMA in the (K+1)th time slot length. Data is unloaded internally.

[0100] S6. According to the calculation time allocation variable and the calculation resource allocation variable, the base station calculates the data according to the allocated calculation resources within a set time; specifically,

[0101] S61, Calculate the resource allocation variables for base stations and data acquisition terminals. Number of computing resources : ,in, The number of data collection terminals. For serial number The data acquisition speed of the acquisition terminal The length of the (i+1)th time slot The total computing power of the base station;

[0102] S62, Base station allocates variables during calculation time, i.e., the... Each time slot length Data calculations are performed within a given time period, among which... ,in, For all possible permutations of the terminal scheduling variable π, The number of data collection terminals. For serial number The data acquisition speed of the acquisition terminal It is the length of the (i+1)th time slot.

[0103] This NOMA-based sensor data acquisition and computation offloading method effectively improves data acquisition and computation efficiency by optimizing terminal scheduling, time allocation, and computational resource allocation, thereby maximizing the amount of data offloaded. Addressing the sequential nature of traditional NOMA technology's data acquisition and offloading phases, this method allows some terminals to offload data while others continue acquisition, effectively improving the overall efficiency of data acquisition and computation offloading and achieving a higher offloaded data volume. Experimental simulations demonstrate that, compared to existing methods, this NOMA-based sensor data acquisition and computation offloading method achieves higher offload throughput, with a more significant performance improvement when total computational resources are abundant.

[0104] This NOMA-based sensing data acquisition and computation offloading method establishes an optimization problem model for sensing data acquisition and computation offloading based on NOMA. The optimization variables are terminal scheduling, time allocation (including acquisition and offloading time, and computation time), and computation resource allocation variables. The optimization problem model is simplified by transforming the computation resource allocation variables and computation time allocation variables into functions of terminal scheduling variables and acquisition and offloading time allocation variables. The simplified optimization problem is then transformed into a two-layer nested optimization problem. The outer optimization problem optimizes the terminal scheduling variables, while the inner optimization problem optimizes the acquisition and offloading time allocation variables. The outer problem is optimized using a one-dimensional search, and the inner problem is optimized using linear programming. Based on the acquisition and offloading time allocation variables, all terminals acquire data within a specified time. Then, according to the terminal scheduling variables, scheduled terminals sequentially offload data based on NOMA within the specified time according to the acquisition and offloading time allocation variables. Terminals that have not yet been scheduled continue to acquire data. Finally, according to the computation time allocation variables, the base station performs data computation within a set time. This NOMA-based sensing data acquisition and computation offloading method optimizes terminal scheduling, time allocation, and computing resource allocation variables to maximize the amount of data offloaded, thereby effectively improving data acquisition and computation efficiency.

[0105] The NOMA-based sensing data acquisition and computation offloading method described in this embodiment is verified through experimental simulation as follows:

[0106] The NOMA-based sensing data acquisition and computation offloading method of this embodiment is compared with the selected existing methods, namely the traditional NOMA scheme (all terminals perform data offloading simultaneously) and the TDMA scheme (each terminal is allocated a dedicated time for data offloading, while other terminals that have not offloaded continue to acquire data).

[0107] The simulation parameters are set as follows: 6 terminals are randomly distributed within a distance of 10 to 100 meters around the base station, and the channel gain is... ,in, For distance, The system follows an exponential distribution with a mean of 1, and the system bandwidth is 1MHz. Uniformly distributed between 0.1 Mbits / s and 1 Mbits / s s, W, bits / s W.

[0108] Figure 4 The embodiment of the NOMA-based sensing data acquisition and computation offloading method differs from existing methods in that the amount of offloaded data increases with the total computing resources. A diagram illustrating the changes. (From...) Figure 3 It can be seen that when total computing resources When increasing, the amount of unloaded data increases first, and then the total computing resources increase. It tends to level off when the total computing resources are large. This is because when the total computing resources... When the data volume is relatively small, the amount of data offloaded is limited by the base station's computing resources, thus increasing the total computing resources required. This allows base stations to compute more data offloaded from terminals, and when total computing resources... When the data volume is large, the limited data acquisition capability prevents further increases in the offloaded data. Compared to the traditional NOMA scheme, both the proposed method and the TDMA method in this embodiment achieve higher offload throughput, and when the total computing resources... The performance improvement is more significant when the value is larger. Compared with the TDMA method, this NOMA-based sensing data acquisition and computation offloading method in this embodiment shows that... The amount of unloaded data achieved at smaller scales is roughly the same, but as the total computing resources increase... As the amount of data increases, the amount of unloading data achieved by the method in the embodiments becomes larger.

[0109] The above description is only a preferred embodiment of the present invention. The scope of protection of the present invention is not limited to the above embodiments. Any equivalent modifications or changes made by those skilled in the art based on the content disclosed in the present invention should be included within the scope of protection set forth in the claims.

Claims

1. A method for sensing data acquisition and computation offloading based on NOMA, applied to a data acquisition system in the Internet of Things (IoT), the data acquisition system comprising a base station equipped with a mobile edge computing server and several acquisition terminals, characterized in that: Includes the following steps, S1. Establish an optimization problem model for sensing data acquisition and computational offloading based on NOMA with the goal of maximizing the amount of unloaded data. The optimization variables of the optimization problem model are terminal scheduling variables, computational resource allocation variables, and time allocation variables. Among them, the time allocation variables include acquisition and offloading time allocation variables and computation time allocation variables. In step S1, an optimization problem model P1 based on NOMA for sensing data acquisition and computational unloading is established with the objective of maximizing the amount of unloaded data. P1： ， in, For all possible permutations of the terminal scheduling variable π, Assign variables to the collection and unloading times. For the first The time slot length is , and c is a variable for calculating resource allocation. The number of data collection terminals. For serial number The data acquisition speed of the acquisition terminal The length of the (i+1)th time slot; Constraint C1, used to indicate that the data collected by each acquisition terminal must be offloaded to the base station: ; Constraint C2 represents the computing power allocated to all data acquisition terminals and cannot exceed the total computing power of the mobile edge computing servers equipped at the base station. ,in, Assigning sequence numbers to base stations The number of computing resources for the data acquisition terminal. The total computing power of the base station; Constraint C3 is used to indicate that the data collected by each acquisition terminal must be calculated at the base station: , Assigning sequence numbers to base stations The number of computing resources for the data acquisition terminal. For the first Each time slot length; Used to represent The constraint C4 is that the sum of the time slots cannot exceed the total time. The entire data acquisition, unloading, and computation time is divided into several parts. One time slot: The length of the first time slot is Data is collected from all acquisition terminals. For the first Each time slot length , Assigned to serial number The acquisition terminal performs data offloading and sequence numbering based on NOMA. The data acquisition terminal collects data, the first Each time slot length Allocate data to base stations for data processing. The total time for data acquisition and unloading calculation for all acquisition terminals; Optimization variables: Terminal scheduling variables ,in, For serial number The data acquisition terminals are sorted according to a set sorting method; data acquisition and unloading time allocation variables. Calculate the time allocation variable, i.e., the first Each time slot length Calculate resource allocation variables ,in, Assigning sequence numbers to base stations The number of computing resources for the acquisition terminal; S2. By transforming the computational resource allocation variables and computational time allocation variables into functions of terminal scheduling variables and acquisition and unloading time allocation variables, the optimization problem model is simplified. The optimization variables of the simplified optimization problem model are the terminal scheduling variables and the acquisition and unloading time allocation variables; specifically, S21. By ensuring that constraints C2 and C3 hold true, the sequence number assigned to the base station is obtained. The number of computing resources of the acquisition terminal and the Each time slot length The expression is: ， ； in, The number of data collection terminals. For serial number The data acquisition speed of the acquisition terminal The length of the (i+1)th time slot The total computing power of the base station; S22, and then the optimization problem model P1 is simplified into the simplified optimization problem model P2: P2： ， in, For all possible permutations of the terminal scheduling variable π, The number of data collection terminals. For serial number The data acquisition speed of the acquisition terminal The length of the (i+1)th time slot; Constraint C1: ; Constraint C5: ; Optimization variables: Terminal scheduling variables Data collection and unloading time allocation variables ; S3. The simplified optimization problem model is transformed into a two-layer nested optimization problem, in which the outer optimization problem optimizes the terminal scheduling variable, and the inner optimization problem optimizes the acquisition and unloading time allocation variable. S4. Solve the two-layer optimization problem to obtain the terminal scheduling variables and the acquisition and unloading time allocation variables, respectively; S5. Based on the data acquisition and unloading time allocation variables, all acquisition terminals are made to acquire data within a specified time. Then, according to the terminal scheduling variables, the acquisition terminals are scheduled to unload data in sequence according to the data acquisition and unloading time allocation variables within a specified time based on NOMA. Acquisition terminals that have not yet been scheduled continue to acquire data. S6. According to the calculation time allocation variable and the calculation resource allocation variable, the base station performs calculations on the data according to the calculation resource allocation within the set time.

2. The NOMA-based sensing data acquisition and computation offloading method as described in claim 1, characterized in that: In step S1, the sequence number The acquisition terminal at the unloading rate per time slot for: ， Where, i= , For serial number The transmission power of the acquisition terminal, For serial number The wireless channel gain from the data acquisition terminal to the base station, This represents noise power.

3. The NOMA-based sensing data acquisition and computation offloading method as described in claim 1 or 2, characterized in that: In step S4, the terminal scheduling variables and the acquisition and unloading time allocation variables are obtained by solving the two-layer optimization problems. Specifically, the terminal scheduling variables are obtained by solving the outer-layer optimization problem using a one-dimensional search method. Solve for the given terminal scheduling variables The inner-layer optimization problem under the given conditions is solved using linear programming to obtain the acquisition and unloading time allocation variables. .

4. The NOMA-based sensing data acquisition and computation offloading method as described in claim 3, characterized in that: In step S4, the terminal scheduling variables are obtained by solving the outer optimization problem using a one-dimensional search method. Solve for the given terminal scheduling variables The inner-layer optimization problem under the given conditions is solved using linear programming to obtain the acquisition and unloading time allocation variables. Specifically, S41. Initialize search step size The number of searches during a one-dimensional search ; S42. Let the weighting factor for the tradeoff between acquisition rate and channel gain be... equal The smaller of the two; S43, Set the terminal scheduling variable For the data acquisition terminal The order is descending from largest to smallest, where... The acquisition rate is normalized to the maximum acquisition rate. It is the reciprocal of the channel gain normalized to the minimum channel gain; S44. Solve for the given terminal scheduling variables. The inner-layer optimization problem under the condition of value is obtained by allocating variables for data acquisition and unloading time. And obtain the objective function value of the inner optimization problem; S45, if If so, proceed to the next step S46; otherwise, let And return to step S42; S46. Select the largest objective function value from all obtained objective function values, and set the terminal scheduling variable corresponding to the largest objective function value as... And the allocation of variables for collection and unloading time This is the final solution to the simplified optimization problem model P2.

5. The NOMA-based sensing data acquisition and computation offloading method as described in claim 1 or 2, characterized in that: In step S5, variables are allocated based on the collection and unloading time to ensure that all collection terminals collect data within a specified time. Then, according to the terminal scheduling variables, the scheduled acquisition terminals sequentially perform data offloading based on NOMA within the specified time according to the acquisition and offloading time allocation variables. Acquisition terminals that have not yet been scheduled continue to acquire data. Specifically... S51, ensure all acquisition terminals are in the first time slot length Data collection is conducted internally; S52, the sequence number of the i-th scheduled item. data acquisition terminal The length of the (k+1)th time slot and Data offloading based on NOMA, and unscheduled acquisition terminals. Data collection is performed; S53, All acquisition terminals are based on NOMA in the (K+1)th time slot length. Data is unloaded internally.

6. The NOMA-based sensing data acquisition and computation offloading method as described in claim 1, characterized in that: In step S6, according to the calculation time allocation variable and the calculation resource allocation variable, the base station performs calculations on the data according to the calculation resource allocation at a set time. Specifically, S61, Calculate the resource allocation variable where the base station is assigned to the sequence number. The number of computing resources of the acquisition terminal : ,in, The number of data collection terminals. For serial number The data acquisition speed of the acquisition terminal The length of the (i+1)th time slot The total computing power of the base station; S62, Base station allocates variables during calculation time, i.e., the... Each time slot length Data calculations are performed within a given time period, among which... ,in, For all possible permutations of the terminal scheduling variable π, The number of data collection terminals. For serial number The data acquisition speed of the acquisition terminal It is the length of the (i+1)th time slot.

Images