Joint Optimization Method, Device and Storage Medium for Communication, Sensing and Computing Oriented to Crowdsensing

Abstract

The present invention discloses a joint optimization method, device and storage medium for communication, sensing and computing for crowdsensing, which relates to the field of crowdsensing, and includes the following steps: establishing a joint optimization algorithm for communication, sensing and computing based on a sensing platform; obtaining a communication, sensing and computing strategy for a sensing task by using the joint optimization algorithm based on current user state information and network resources; and collecting, transmitting and computing sensing data by a terminal participating in the sensing task based on the communication, sensing and computing strategy. The present invention is based on the technology of jointly optimizing and designing data sensing, computing and transmission strategies of this system, and also considers user selection and bandwidth allocation strategies at the same time to realize the effective utilization of network resources, greatly improving the performance of the system under the condition of limited network resources, and is very easy to be implemented in an actual system.

Description

Technical Field

[0001] The present invention relates to the field of crowd sensing, and in particular to a joint optimization method, device, and storage medium for communication, sensing, and computing for crowd sensing. Background Art

[0002] In recent years, with the widespread popularity of intelligent mobile terminal devices (such as smart phones, tablets, wearable devices, etc.) and the rapid development of embedded sensing technologies, a new sensing paradigm that uses the sensors embedded in the mobile terminal devices carried by people to collect environment-related information, namely mobile crowd sensing (MCS), has attracted wide attention in the industrial and academic fields. Compared with the traditional sensing mode formed by fixedly deployed sensors, MCS has the advantages of a wide sensing range, a variety of sensed data types, low deployment costs, and high scalability. Based on these advantages, MCS is widely used in various fields closely related to human life, such as environmental monitoring, intelligent transportation, autonomous driving, and information sharing.

[0003] Although MCS provides many advantages compared with the traditional sensing mode, it also faces new challenges when deployed in a wireless edge network. For example, in a wireless edge network, due to limited communication resources and / or unstable wireless channels, the wireless transmission of sensed data may fail. Although the overhead of communication resources can be reduced through the local (pre-)processing mode of sensed data, this will also consume the computing and energy resources of local devices. And since the resources of local devices are also limited, the preprocessing of sensed data and the wireless transmission of processing results may also fail. Most of the current task allocation-related work ignores the various limitations of wireless network resources and the computing process of sensed data, so the frameworks or algorithms proposed by them cannot guarantee that the sensed data can be effectively transmitted and processed in a wireless communication network.

[0004] In view of this, the present application is specifically proposed. Summary of the Invention

[0005] To solve the deficiencies of the prior art, the present invention provides a joint optimization method, device, and storage medium for communication, sensing, and computing for crowd sensing, aiming to jointly consider the sensing, communication, and computing strategies involved in the implementation process of sensing tasks under the conditions of multi-dimensional network resource limitations, so as to optimize the performance of the MCS system.

[0006] The present invention is realized through the following technical solutions:

[0007] In a first aspect, the present invention provides a joint optimization method for communication, sensing, and computing for crowd sensing, including the following steps:

[0008] Establish a joint optimization algorithm for communication, sensing, and computing based on a sensing platform;

[0009] Based on the current user status information and network resources, a joint sensing-communication-computation optimization algorithm is used to obtain the sensing-communication-computation strategy for the sensing task;

[0010] Based on the sensing-communication-computation strategy, the terminals participating in the sensing task collect, transmit, and calculate the sensing data.

[0011] Further, before the sensing platform establishes the joint sensing-communication-computation optimization algorithm, it includes:

[0012] The user reports the current status information to the sensing platform;

[0013] The sensing platform uses the sensing platform to obtain the sensing task with time constraints.

[0014] Further, in the user's reporting of the current status information to the sensing platform, the status information includes channel status information and user ability information, where:

[0015] The terminal sends a sounding reference signal to help the base station obtain the channel status information of each user, and the base station reports the channel status information to the sensing platform;

[0016] The sensing platform periodically queries the terminals through the base station to obtain the user ability information reported by the terminals.

[0017] Further, the joint sensing-communication-computation optimization algorithm includes:

[0018] The first algorithm is used to solve the data sensing strategy, data transmission strategy, and data calculation strategy in the partial offloading calculation mode based on the given user selection strategy and bandwidth allocation strategy;

[0019] The second algorithm is used to solve the optimal number of bandwidth units allocated to each user based on the dynamic programming method;

[0020] The third algorithm is used to solve the data sensing strategy, data transmission strategy, and data calculation strategy in the binary offloading calculation mode based on the given user selection strategy and bandwidth allocation strategy.

[0021] Further, in obtaining the sensing-communication-computation strategy for the sensing task by using the joint sensing-communication-computation optimization algorithm, it includes:

[0022] Use the second algorithm to obtain the optimal bandwidth allocation strategy in the partial offloading calculation mode;

[0023] Determine the user selection strategy according to the bandwidth allocation strategy; where, when a user is allocated a certain amount of bandwidth, it means that this user is selected to execute this sensing task; otherwise, it is not selected;

[0024] According to the bandwidth allocation and user selection strategies, use the first algorithm to obtain the data sensing strategy, data transmission strategy, and data processing strategy.

[0025] Furthermore, according to the bandwidth allocation and user selection strategies, determining the data sensing strategy, data transmission strategy, and data processing strategy further includes:

[0026] Use the second algorithm to obtain the optimal bandwidth allocation strategy in the binary offloading computing mode;

[0027] Determine the user selection strategy according to the bandwidth allocation strategy; where, when a user is allocated a certain amount of bandwidth, it means this user is selected to execute this sensing task; otherwise, it is not selected;

[0028] According to the bandwidth allocation and user selection strategies, use the third algorithm to obtain the data sensing strategy, data transmission strategy, and data processing strategy.

[0029] Furthermore, in the process of the terminals participating in the sensing task collecting, transmitting, and calculating sensing data based on the communication-sensing-computing strategy, it includes:

[0030] The sensing platform issues the communication-sensing-computing strategy to the base station;

[0031] The base station accesses and authorizes the terminals participating in the sensing task;

[0032] The terminals participating in the sensing task determine information such as the sensing time, sensing data transmission time, transmission power, local processing time, and local processing result transmission time, and perform the collection, transmission, and calculation of sensing data;

[0033] Furthermore, after the terminals participating in the sensing task collect, transmit, and calculate the sensing data, it further includes:

[0034] The base station then transmits the sensing data or the sensing data processing result to the sensing platform to complete this sensing task.

[0035] In a second aspect, the present invention provides a computer device, including:

[0036] A memory for storing a computer program;

[0037] A processor for implementing the steps of the above communication-sensing-computing joint optimization method for group intelligence sensing when executing the computer program.

[0038] In a third aspect, the present invention provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, it implements the steps of the above communication-sensing-computing joint optimization method for group intelligence sensing.

[0039] Compared with the prior art, the present invention has the following advantages and beneficial effects:

[0040] A joint optimization method for communication, sensing, and computing for crowdsensing, a computer device, and a storage medium provided by the present invention establish a joint optimization algorithm for communication, sensing, and computing based on a sensing platform; based on the current user state information and network resources, use the joint optimization algorithm for communication, sensing, and computing to obtain the communication, sensing, and computing strategy for the sensing task; based on the communication, sensing, and computing strategy, the terminals participating in the sensing task collect, transmit, and calculate sensing data. Thus, in terms of the technology of jointly optimizing the data sensing, computing, and transmission strategies of this system, and at the same time considering the user selection and bandwidth allocation strategies, the effective utilization of network resources is realized, the performance of the system under limited network resource conditions is greatly improved, and it is very easy to be implemented in an actual system. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] In order to more clearly illustrate the technical solutions of the exemplary embodiments of the present invention, the following will briefly introduce the drawings required for the embodiments. It should be understood that the following drawings only show some embodiments of the present invention, and thus should not be regarded as limiting the scope. For those of ordinary skill in the art, other related drawings can be obtained based on these drawings without creative efforts. In the drawings:

[0042] Figure 1 It is the method flowchart of the joint optimization method for communication, sensing, and computing for crowdsensing provided in this embodiment;

[0043] Figure 2 It is the sub-step flowchart of step S12 of the joint optimization method for communication, sensing, and computing for crowdsensing provided in this embodiment;

[0044] Figure 3 It is the sub-step flowchart of step S13 of the joint optimization method for communication, sensing, and computing for crowdsensing provided in this embodiment;

[0045] Figure 4 It is the architecture diagram of the mobile crowdsensing system of the joint optimization method for communication, sensing, and computing for crowdsensing provided in this embodiment;

[0046] Figure 5 It is the flowchart of the terminal ability detection of the joint optimization method for communication, sensing, and computing for crowdsensing provided in this embodiment;

[0047] Figure 6 It is the P1 equivalent inner-outer sub-problem formula diagram of the joint optimization method for communication, sensing, and computing for crowdsensing provided in this embodiment;

[0048] Figure 7 It is the first algorithm schematic diagram of the joint optimization method for communication, sensing, and computing for crowdsensing provided in this embodiment;

[0049] Figure 8 This is the second algorithm schematic diagram of the joint optimization method of sensing, communication and computing for crowd-sourced sensing provided in this embodiment;

[0050] Figure 9 This is the first schematic diagram of the P2 equivalent inner-outer sub-problem of the joint optimization method of sensing, communication and computing for crowd-sourced sensing provided in this embodiment;

[0051] Figure 10 This is the third algorithm schematic diagram of the joint optimization method of sensing, communication and computing for crowd-sourced sensing provided in this embodiment. Detailed implementation manners

[0052] To make the objectives, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are some but not all of the embodiments of the present invention. Components of the embodiments of the present invention generally described and illustrated in the figures herein may be arranged and designed in a variety of different configurations.

[0053] Therefore, the following detailed description of the embodiments of the present invention provided in the drawings is not intended to limit the scope of the claimed present invention, but merely represents selected embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts fall within the scope of protection of the present invention.

[0054] It should be noted that like reference numerals and letters denote like items in the following figures, and thus, once an item is defined in one figure, it need not be further defined and explained in subsequent figures.

[0055] Currently, it is important to jointly consider sensing, communication and computation (JSCC) in the design of MCS systems. Some existing technologies have proposed a wireless power supply-based MCS framework to jointly control the wireless power supply, data sensing, compression and transmission processes under this framework. However, it ignores the impact of the computing process of sensing data and the limited computing resources in the network on the performance of MCS. For example, in the joint optimization scheme of data sensing and task offloading in a drone-based MCS system, existing technologies have provided a joint framework for crowd-sourced sensing and computing. However, they all ignore the impact of the transmission process of sensing data and the limited communication resources.

[0056] Therefore, in view of the above problems, the present invention aims to jointly consider the sensing, communication and computing strategies involved in the implementation process of sensing tasks under the conditions of multi-dimensional network resource constraints, so as to optimize the performance of the MCS system.

[0057] The present invention provides a joint optimization method for communication, sensing and computing for crowd-sensing, as Figure 1 shown, which includes the following steps:

[0058] S11: Establish a joint optimization algorithm for communication, sensing and computing based on the sensing platform;

[0059] Among them, this solution will first mathematically model the sensing, communication and computing problems in the MCS system, so as to establish a JSCC (joint sensing, communication and computation) framework for the multi-dimensional resource-constrained MCS system. Compared with the traditional sensing mode, MCS provides many advantages. As Figure 4 shown, the MCS system is mainly composed of multiple mobile users providing sensing services and a sensing platform managing sensing tasks. It is generally deployed in a wireless edge network, where the sensing platform is deployed on an edge server, and the edge server is deployed near a base station (BS). When a given sensing task is given, the sensing platform first needs to select appropriate users to participate. The selected users then collect environment-related information and transmit the collected sensing data to the sensing platform through the wireless network for corresponding data processing, such as big data analysis and machine learning model training. However, it is generally difficult for existing technologies to pay attention to the joint optimization design of data sensing, transmission and computing in the MCS system. Most of the existing technologies currently focus on solving the user incentive problem in the MCS framework, that is, how to motivate mobile users to participate in crowd-sensing, and the user selection (or task allocation) problem, that is, how to allocate sensing tasks to appropriate users. In particular, most of the current task allocation-related work ignores the multi-faceted limitations of wireless network resources and the computing process of sensing data. Therefore, the frameworks or algorithms they propose cannot guarantee that the sensing data can be effectively transmitted and processed in the wireless communication network.

[0060] Based on the above discussion, the present invention aims to jointly consider the sensing, communication, and computing strategies involved in the implementation of sensing tasks under the conditions of multi-dimensional network resource constraints, so as to optimize the performance of the MCS system. The present invention jointly optimizes the five major strategies involved in each sensing task based on JSCC, namely user selection, bandwidth allocation, data sensing, transmission, and computing. This solution models the above problem as a non-convex joint optimization problem. To solve this problem, this solution first rewrites it into an equivalent inner-outer sub-problem form, where the inner sub-problem is to jointly optimize the data sensing, transmission, and computing strategies of each selected user, and the outer sub-problem is to further determine the user selection and bandwidth allocation strategies. This solution uses an iterative optimization algorithm to solve the inner sub-problem, and based on this iterative algorithm, uses a dynamic programming algorithm to solve the outer sub-problem. By jointly optimizing the sensing, communication, and computing strategies involved in the implementation of sensing tasks, the JSCC mechanism proposed in this solution greatly improves the performance of the MCS system under the conditions of limited network resources compared with the existing solutions, and is very easy to be implemented in an actual system.

[0061] S12: Based on the current user status information and network resources, use the joint sensing, communication, and computing optimization algorithm to obtain the joint sensing, communication, and computing strategy for the sensing task;

[0062] In this solution, the joint sensing, communication, and computing strategy in this embodiment includes: user selection strategy (i.e., determining the users participating in this sensing task), bandwidth allocation strategy (i.e., determining the uplink bandwidth of each selected user), data sensing strategy (i.e., determining the sensing data collection time of each selected user), data transmission strategy (i.e., determining the sensing data transmission time and transmission power of each selected user), and data processing strategy (i.e., determining whether the sensing data is processed locally, as well as the local processing time and the processing time of the sensing platform). According to the data type requirements and time limits of the sensing task, the sensing platform executes the joint sensing, communication, and computing optimization algorithm to output the joint sensing, communication, and computing strategy for this sensing task, so as to maximize the amount of sensing data processed within the specified time of the sensing task. The purpose is to achieve the effective processing of sensing data by designing the transmission and computing strategies of sensing data. Further, this solution will simultaneously consider the partial offloading and binary offloading computing modes. In the partial offloading mode, the computing task of sensing data can be finitely divided, that is, the sensing data can be processed locally and on the server at the same time, while in the binary offloading mode, the processing of sensing data can only be performed locally or on the server.

[0063] S13: Based on the joint sensing, communication, and computing strategy, the terminals participating in the sensing task collect, transmit, and compute the sensing data;

[0064] In this embodiment, as those skilled in the art should know, the sensing task can be a task in the field of intelligent transportation, and its sensing data may include the speed of vehicles, photo information of roads, etc. It can also be a task in the field of environmental monitoring, and its sensing data may be temperature, noise level (numerical sensing data), etc. The present invention does not make further limitations. An embedded sensor for collecting different types of sensing data can be provided inside the terminal participating in the sensing task. For example, an acceleration sensor can be used to collect sensing data related to motion, a GPS sensor can be used to collect sensing data related to position and speed, and a camera can be used to collect sensing data related to images. Due to the requirements of the sensing task and the limitation of wireless bandwidth resources, only a suitable terminal can be selected to execute the sensing task. Therefore, this solution will also consider user selection and bandwidth allocation strategies to achieve the effective utilization of network resources.

[0065] In this embodiment, the sensing data can be processed in a centralized or distributed manner. That is to say, the sensing data can either be transmitted to the server for unified processing or the local computing power of the sensing device can be used to process the sensing data, and then the processing result is transmitted to the server for fusion. The computing strategy of the sensing data is to determine how each sensing device processes the collected sensing data, and there are three possible situations: 1) local processing and then transmitting the result to the server; 2) completely processed by the server; 3) part of the processing is done locally and part is done by the server. Note that this solution will consider two computing task offloading modes, namely partial offloading and binary offloading. Since the processing of sensing data is indivisible in the binary offloading mode, the above third situation will not appear in the binary offloading mode. Since different sensing strategies (such as sensing time) will result in different amounts of sensing data, thus leading to different computing resource requirements; and different computing strategies will also result in different communication resource overheads (because the amount of data transmitted is different), it is very necessary to jointly optimize and design the sensing, communication, and computing strategies of the MCS system to maximize the amount of data processed in the sensing task. This is because when more data is processed, the processing result is always more accurate and general. For example, in an intelligent transportation task, the more photos are processed, the more accurate the traffic information obtained.

[0066] In an alternative embodiment of the present application, before establishing the joint optimization algorithm for communication, sensing, and computing based on the sensing platform, it includes:

[0067] S101: The user reports the current status information to the sensing platform;

[0068] S102: The sensing platform is used to obtain a sensing task with a time limit.

[0069] Among them, the sensing platform periodically updates and maintains the status information of all mobile crowd sensing users within its current scope, as well as the sensing task information that has currently arrived. For each arrived sensing task, the sensing platform needs to execute the joint communication-sensing-computation optimization algorithm of this solution based on the currently available network resources and user status information, and send relevant policies to the corresponding users and base station terminals.

[0070] Furthermore, in the status information reported by the user to the sensing platform, the status information includes channel status information and user capability information, where:

[0071] The terminal sends a sounding reference signal to help the base station obtain the channel status information of each user, and the base station reports the channel status information to the sensing platform; the sensing platform periodically interrogates the terminal through the base station to obtain the user capability information reported by the terminal.

[0072] In this embodiment, the base station / sensing platform triggers the sensing users to periodically measure and report status information (mainly channel gain, remaining energy of the user, etc.). Among them, the sounding reference signal (SRS): as a UL signal, the terminal directly reports the channel information to the base station, and the terminal helps the base station obtain the channel state information (CSI) of each user by sending SRS. The MCS needs to periodically obtain the capabilities of the terminal (such as information such as the available power of the user), set a timer on the sensing platform side, and periodically interrogate the terminal through the RRC of the base station to obtain the capability information reported by the terminal. Among them, the base station obtaining the terminal capability information mainly involves two processes: terminal capability inquiry and terminal capability reporting. Specifically, please refer to Figure 5 : Step 1, the MSC sets a terminal capability detection timer for all sensing terminals. If the timer times out, it sends a terminal capability request message to the base station; Step 2, when the MCS triggers the need to obtain power information, the base station issues a terminal capability inquiry command; Step 3, when receiving the terminal capability inquiry command, the terminal reports the corresponding capability information according to the command, that is, UE Capability Information. The RRC needs to configure Optional features without UE radio access capability parameters when the terminal accesses the base station; Step 4, the base station reports all the received terminal capabilities to the MCS, and the terminal capability detection ends.

[0073] Furthermore, the joint communication-sensing-computation optimization algorithm includes:

[0074] The first algorithm is used to solve the data awareness strategy, data transmission strategy, and data computing strategy in the partial offloading computing mode based on the given user selection strategy and bandwidth allocation strategy.

[0075] The second algorithm is used to solve the optimal number of bandwidth units allocated to each user based on the dynamic programming method.

[0076] The third algorithm is used to solve the data awareness strategy, data transmission strategy, and data computing strategy in the binary offloading computing mode based on the given user selection strategy and bandwidth allocation strategy.

[0077] The JSCC mechanism of this embodiment aims to jointly optimize the design of user selection, bandwidth allocation, collection, transmission, and processing strategies of sensed data for the sensing task. This problem can be transformed into a joint optimization problem, that is, under the conditions of time constraint, energy constraint, and bandwidth constraint, jointly optimize the design of user selection variable l i , bandwidth allocation variable k i , sensing time t i,s , local processing time t of sensed data i,c , time t for transmitting sensed data to the server i,r , and transmit power P i (i = 1,..., N), so as to maximize the amount of sensed data processed in the sensing task.

[0078] Note that when user i is not allocated bandwidth, it proves that this user is not selected to execute this sensing task, that is, l i = 0; on the contrary, then l i = 1. That is to say, the user selection variable l i can be completely determined by the bandwidth allocation strategy, that is, there is Therefore, the above optimization problem can remove the variable l i . On the other hand, the amount of data processed in the sensing task cannot exceed the amount of sensed data collected by all selected users. In particular, let the sensing rate of user i, that is, the amount of sensed data collected per second, be o i , then the amount of sensed data collected by the user in t i,s time is t i,s o i . Let c i be the number of CPU cycles required to calculate one bit of the sensed data collected by user i, then the amount of data locally processed by user i in t i,c is where f i is the local computing power (or CPU frequency) of user i. At the same time, the amount of data transmitted by user i to the server for processing in t i,r is t i, r r i , where is the uplink transmission rate of user i, k i B min is the allocated uplink bandwidth of user i, B min is the minimum bandwidth allocation unit. To maximize the amount of data processed in the sensing task, there must be a constraint that the amount of data processed is equal to the amount of data collected, that is, in the partial offloading computing mode, there is And in the binary offloading computing mode, there is where a i is a binary variable used to represent whether the sensing data of user i is processed locally. That is to say, the variable t i,s can be determined by other parameters, so the above optimization problem can also be simplified by removing the variable t i,s for simplification.

[0079] This scheme considers two computing modes, namely partial offloading and binary offloading modes, and different modes correspond to different constraints and optimization models. Therefore, these two modes need to be analyzed separately. First, consider the partial offloading computing mode. Based on the data sensing time the following optimization problem can be obtained:

[0080]

[0081]

[0082]

[0083]

[0084]

[0085]

[0086] The objective function in Equation (4) is to maximize the amount of data processed in the sensing task, or the weighted sum of computation bits (WSCB), where b i is the weighting coefficient, which is related to the type of sensing data. Equations (4a) and (4b) are the sensing task time constraints, that is, it is required that t i,s +t i,r ≤T and t i,s +t i,c ≤T. Equation (4c) is the energy constraint of the sensing device, where e i is the energy consumed by user i to collect a unit bit, κt i,c f i3 is the energy consumed for local processing of the sensed data by user i, while P i t i,r is the energy consumed for user i to transmit the sensed data. Equation (4d) is the bandwidth constraint, which requires that the uplink bandwidth allocated to user i is an integer multiple of the minimum allocated bandwidth B min , that is, k i is an integer. Equation (4e) is the transmit power constraint of the user. When the optimal solution of the above optimization problem is obtained, the best user selection strategy is determined by , and the best sensing time strategy is determined by . Similarly, in the binary offloading calculation mode, there is an optimization problem P2 as follows:

[0087]

[0088]

[0089]

[0090]

[0091] Among them, equations (5), (5a)-(5c) are the objective function, the sensing task time constraint, and the energy constraint of the sensing device in the binary offloading calculation mode. In particular, an additional binary variable a i is introduced in the binary offloading calculation mode to represent whether the sensed data of user i is processed locally. In this mode, t i,c and t i,r cannot be greater than 0 at the same time.

[0092] In an alternative embodiment of the present application, in the communication-sensing-computation joint optimization algorithm for obtaining the communication-sensing-computation strategy of the sensing task, it includes:

[0093] S1211: Obtain the best bandwidth allocation strategy in the partial offloading calculation mode by using the second algorithm;

[0094] S1212: Determine the user selection strategy according to the bandwidth allocation strategy; wherein, when a user is allocated a certain amount of bandwidth, it means that this user is selected to execute this sensing task; otherwise, it is not selected;

[0095] S1213: According to the bandwidth allocation and user selection strategies, use the first algorithm to obtain the data sensing strategy, the data transmission strategy, and the data processing strategy;

[0096] Such as Figure 2As shown, steps S1211 - S1213 are the application of the joint optimization algorithm for communication, sensing, and computing in the partial offloading computing mode. Considering the joint optimization algorithm for communication, sensing, and computing in the partial offloading computing mode, that is, how to solve problem P1. Since the variable k i is an integer, and there is a product relationship between k i and t i,r in the objective function, P1 is a non - convex optimization problem. To solve P1, it is first transformed into an equivalent inner - outer sub - problem, as Figure 6 shown:

[0097] where the inner sub - problem P3 jointly optimizes the data sensing, transmission, and computing strategies of each selected user under the condition of a given user selection and bandwidth allocation strategy (i.e., given k i ), while the outer sub - problem P4 further optimizes the user selection and bandwidth allocation strategy.

[0098] First, consider how to solve the inner sub - problem P3. Problem P3 jointly optimizes the local processing time t i of the sensed data, the time t i,c to transmit the sensed data to the server, and the transmit power P i,r given the variable k i . Since when k i = 0, that is, when user i is not selected to participate in the sensing task, the variables t i,c , t i,r and P i are all 0. Therefore, consider the case where k i > 0.

[0099] Due to the non - convexity of the log function in constraints (4a) - (4c), P3 is still a non - convex optimization problem. Based on this, this scheme considers solving P3 based on the idea of iterative optimization. Specifically, consider iteratively optimizing two sub - problems of P3, P3 - A and P3 - B, to iteratively optimize the parameters {P i} and {t i,c , t i,r}. That is to say, P3 - A optimizes the transmit power P i,c given t i,r and t i , while P3 - B optimizes the parameters t i and t i,c given P i,r . The purpose is to ensure obtaining the optimal {P i} or {t i,c , t i,r} under the condition of given parameters.

[0100] Based on the above discussion, given t i,c and ti,r Under the condition of, the optimal solution of P3 - A can be obtained from Theorem 1.

[0101] Theorem 1: When given k i > 0, t i,c and t i,r , the optimal transmit power of user i is:

[0102]

[0103] where and P i,3 satisfies F(P i,3 ) = E i .

[0104] Proof: Obviously, when t i,r = 0, that is, when there is no sensed data transmitted to the server for processing, the transmit power P i = 0. When t i,r > 0, there must be 0 < P i ≤ P max . According to (4a)-(4c), we can obtain the following constraints on P i :

[0105]

[0106]

[0107] F(P i ) ≤ E i (9)

[0108] 0 < P i ≤ P max (10)

[0109] Let Then according to equations (7) and (8), we get P i ≤ P i,1 and P i ≤ P i,2 . Since is a monotonically increasing function of P i , so from equation (9) we can also get P i ≤ P i,3 , where P i,3 satisfies F(P i,3 ) = E i . Since the objective function increases as P i increases, so according to equations (7)-(10), the optimal transmit power can be obtained as Thus, Theorem 1 is proved.

[0110] While given P i under the condition, the optimal solution of P3 - B can be obtained by Theorem 2.

[0111] Theorem 2: When given k i > 0 and P i > 0, the optimal local processing time t i,c and the optimal transmission time t i,r of user i are:

[0112]

[0113] Where

[0114]

[0115]

[0116]

[0117] When P i is given, P3 - B is a linear programming problem with respect to variables t i,c and t i,r . Therefore, Theorem 2 can be easily obtained based on the graphical method. Note that Theorem 2 only considers the case of P i > 0. When P i = 0, there are and It can be seen from Theorem 2 that when the user's energy (i.e., E i ) is large enough, the optimal local processing time t i,c is equal to the optimal transmission time t i,r to make full use of the communication and computing resources allocated to the user. When the user's energy is insufficient, especially when or , the user may choose to locally process the sensed data (i.e., t i,r = 0) or completely offload the sensed data to the server for processing (i.e., t i,c = 0).

[0118] According to Theorem 1 and 2, this scheme solves P3 based on the iterative optimization algorithm. The specific process is shown in the first algorithm, see Figure 7 . The first algorithm contains multiple iterations, and each iteration process contains two steps. The first step will output the best transmit power P i,c ,t i,r found so far based on (t i ) obtained in the previous iteration process and Theorem 1. The second step will output the best t i found so far based on the obtained Pi,c and t i,r 。Note that since there is no closed - form expression for P i,3 in Theorem 1, we adopt a binary search method in the first algorithm to obtain the value of P i,3 . The above process is repeated until the first algorithm converges, and the convergence of the first algorithm is guaranteed by Theorem 3.

[0119] Theorem 3: The first algorithm is guaranteed to converge to a local optimal solution of problem P3.

[0120] Proof: Let be the solution obtained after the j - th iteration of the first algorithm, and be the objective function value obtained based on this solution. Then, in the (j + 1)-th iteration, the variable P i will first be optimized based on to obtain a new value of P i , that is, After that, based on the variables (t i,c , t i,r ) are optimized to obtain the new local processing time and transmission time, that is, Since each step in each iteration is based on Theorem 1 or Theorem 2 to obtain the optimal solution of P3 - A or P3 - B, it can be guaranteed that the objective function values obtained in each step show an increasing trend, that is, there is Also, because the maximum value of the objective function of problem P3 does not exceed the maximum amount of sensing that all users can collect within the sensing task time, which is Therefore, the first algorithm is guaranteed to converge and reach a local optimum.

[0121] Note that the first algorithm requires a feasible initial solution as input. Therefore, in this scheme, at initialization, t i,c = 0, P i = P max . According to the first algorithm, we have obtained the data sensing, transmission, and computing strategies under a given user selection and bandwidth allocation (determined by the variable k i ). Next, we need to further optimize the user selection and bandwidth allocation strategies, that is, to solve the outer sub - problem P4 of P1. Specifically, based on the first algorithm, the optimization problem P1 can be transformed into the following problem:

[0122]

[0123]

[0124] where is the objective function of P1, that is, to calculate the weighted sum of computation bits (WSCB). Specifically, when k i > 0, where Pi*, ti, c*, ti, r* are obtained according to the first algorithm; when ki = 0, Riki, Pi*, ti, c*, ti, r* = 0. It can be seen that the optimization problem P1 has been transformed into how to allocate the optimal number of bandwidth units for each user, that is, k i , so as to maximize the WSCB, where the total number of bandwidth units available for allocation in the network is To solve the above problem, the dynamic programming method can be used. Specifically, consider numbering the users in the network as 1, 2,..., N. Let s i represent the number of bandwidth units allocated to the first i users (1 ≤ i ≤ N), and k i represent the number of bandwidth units allocated to user i, then there is s i-1 = s i - k i . Use f i (s i ) to represent the maximum WSCB value obtained when we allocate s i bandwidth units to the first i users. Based on this, we obtain the following state transition equation:

[0125]

[0126] According to Equation (15), it can be calculated that when there are N users in the network, bandwidth units, the maximum WSCB value (that is, ), and the optimal number of bandwidth units allocated to each user. The specific solution process is shown in the second algorithm. Please refer to Figure 8 . Specifically, the second algorithm will first calculate, based on the first algorithm, the i parameters corresponding when the user is allocated k i (0 ≤ k ≤ M) bandwidth units, so as to obtain the value of WSCB under the current conditions (step 3); then traverse all possible states according to the state transition equation of Equation (15) to output the optimal solution (steps 4 and 5). Note that although the second algorithm can ensure that the obtained bandwidth allocation variable k i is the optimal solution of the external sub-problem P4 of P1, we cannot guarantee that k i is the optimal solution of P1. This is because the obtained based on the first algorithm can only guarantee to be the local optimal solution of the internal sub-problem P3. Therefore, finally, the It can only guarantee a local optimal solution of P1.

[0127] Furthermore, in the determination of the data sensing strategy, data transmission strategy, and data processing strategy according to the bandwidth allocation and user selection strategy, it further includes:

[0128] S1221: Obtain the optimal bandwidth allocation strategy in the binary offloading calculation mode using the second algorithm;

[0129] S1222: Determine the user selection strategy according to the bandwidth allocation strategy; where, when a user is allocated a certain amount of bandwidth, it means this user is selected to execute this sensing task; otherwise, it is not selected;

[0130] S1223: Obtain the data sensing strategy, data transmission strategy, and data processing strategy using the third algorithm according to the bandwidth allocation and user selection strategies.

[0131] Please refer to Figure 3 , Steps S1221 - S1223 are the application of the joint optimization algorithm for communication and sensing calculation in the binary offloading calculation mode. Please refer to Figure 9 , Similar to the idea of solving P1, consider first solving the internal sub - problem P5, and then further solving the external sub - problem P6 based on the solution of P5. Since P6 is similar to the external sub - problem P4 of P1, P6 can also be solved based on the second algorithm. Different from P3, P5 introduces a new binary variable a i , and when a i = 0, t i,c = 0; when a i = 1, t i,r = 0. Based on this discovery, P5 can be solved respectively under the conditions of a i = 0 and a i = 1, and then compare the objective function values corresponding to the solutions under these two conditions, so as to output the solution with a larger objective function value. In particular, P5 under the conditions of a i = 0 and a i = 1 can be solved based on Theorem 4 and Theorem 5.

[0132] Theorem 4: When given k i >0, t i,c , t i,r and a i , the optimal transmit power of user i is:

[0133]

[0134] where and P i,3 satisfies F(P i,3 ) = E i .

[0135] Theorem 5: When given k i > 0, P i and a i , the optimal local processing time t i,c and transmission time t i,r of user i are as follows:

[0136]

[0137] where

[0138] The proof of Theorem 4 is similar to that of Theorem 1. The proof process of Theorem 5 is briefly given below. Note that when a i = 0, we have t i,c = 0; while when a i = 1, we have t i,r = 0. According to a i = 0 and t i,c = 0, equations (5a)-(5c) can be transformed into:

[0139]

[0140]

[0141] e i t i,r r i + P i t i,r ≤ E i (20)

[0142] Since the objective function increases with the increase of t i,r , it can be obtained that under the conditions of a i = 0 and t i,c = 0 Similarly, according to a i = 1 and t i,r = 0, equations (5a)-(5c) can be transformed into:

[0143]

[0144]

[0145]

[0146] Similarly, since the objective function increases with the increase of t i,c , it can be obtained that under the conditions of a i = 1 and t i,r = 0

[0147] Based on Theorem 4 and Theorem 5, the solution process of P5 is as shown in the third algorithm, specifically referring to Figure 10 : First, consider solving P5 under the condition of a i = 0. Consider using the iterative optimization algorithm to solve P5 under the condition of a i = 0, as shown in Step 3. Specifically, in each iteration process, according to Theorem 4, obtain a i = 0 under the given t i,r of the optimal transmit power P i , and then through the obtained P i , according to Theorem 5, obtain the optimal transmission time t i,r . The above process is iterated until it converges to the local optimal solution, which is similar to the first algorithm. Then consider solving P5 under the condition of a i = 1. According to Theorem 5, it can be obtained that a i = 1 under the condition of and as shown in Step 4. Finally, by comparing the objective function values corresponding to the solutions obtained in Step 3 and Step 4, output the solution that makes the output objective function value larger, as shown in Step 5.

[0148] Further, in the process of the terminal participating in the perception task collecting, transmitting, and calculating perception data based on the above-mentioned communication-sensing computing strategy, as Figure 3 shown, it includes:

[0149] S131: The sensing platform issues the above-mentioned communication-sensing computing strategy to the base station;

[0150] Among them, the base station allocates a known random access preamble sequence for the terminal to prepare for the subsequent upload of perception data and the processing results of perception data by the terminal. When performing the optimal bandwidth allocation (broadband allocation strategy) of the second algorithm, it is necessary to enable the terminal performing the perception task to obtain uplink authorization.

[0151] S132: The base station accesses and authorizes the terminals participating in the perception task;

[0152] Among them, in the non-competitive mode (Contention Free Random Access, CFRA), the base station specifies a random access preamble sequence for the UE through the DCI of PDCCH or the RRC (Radio Resource Control) signaling and accesses the base station using this preamble sequence. The terminal bears the signaling of the random access process through RRCMessage1 (MSG1). Further, the base station allocates corresponding time-frequency resources for it to transmit the perception data and the processing results of perception data, and adjusts the transmit power of the terminal.

[0153] S133: The terminals participating in the sensing task determine information such as sensing time, sensing data transmission time, transmit power, local processing time, and local processing result transmission time, and perform the collection, transmission, and calculation of sensing data.

[0154] Further, after the terminals participating in the sensing task perform the collection, transmission, and calculation of sensing data, the following steps are also included:

[0155] The base station then transmits the sensing data or the sensing data processing result to the sensing platform to complete this sensing task.

[0156] For the MCS system with limited multi-dimensional network resources, this embodiment designs a joint sensing, communication, and computing optimization (JSCC) framework. By jointly optimizing and designing user selection, bandwidth allocation, data sensing, transmission, and computing strategies during the implementation of the sensing task, this solution aims to maximize the amount of data processed within the sensing task time under the condition of limited network resources. Our joint sensing, communication, and computing optimization algorithm is far superior to other existing algorithms. The modeling process of this embodiment can provide support for the relationship between sensing-computing-transmission strategies and resource consumption. This embodiment can also be further extended to other related scenarios, such as the cooperation of multiple MCS systems, resource management under the joint sensing, communication, and computing optimization framework, etc.

[0157] In another specific embodiment of the present invention, the present invention provides a computer device, including:

[0158] A memory for storing computer programs;

[0159] A processor for implementing the steps of the above-mentioned joint sensing, communication, and computing optimization method for crowd-sourced sensing when executing the computer program.

[0160] Among them, the computer system includes a processor (CPU), which can perform various appropriate actions and processes according to the program stored in the read-only memory (ROM) or the program loaded from the storage part into the random access memory (RAM). In the RAM, various programs and data required for system operation are also stored.

[0161] The CPU, ROM, and RAM are connected to each other through a bus. The input / output (I / O) interface is also connected to the bus.

[0162] The following components are connected to the I / O interface: an input part including a keyboard, a mouse, etc.; an output part including a cathode ray tube (CRT), a liquid crystal display (LCD), etc. and a speaker, etc.; a storage part including a hard disk, etc.; and a communication part including a network interface card such as a LAN card, a modem, etc. The communication part performs communication processing via a network such as the Internet. A drive is also connected to the I / O interface as needed. A removable medium such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc. is mounted on the drive as needed so that a computer program read from it can be installed into the storage part as needed.

[0163] Specifically, according to an embodiment of the present disclosure, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, an embodiment of the present disclosure includes a computer program product that includes a computer program carried on a computer-readable medium, and the computer program includes program codes for performing the methods shown in the flowcharts. In such an embodiment, the computer program can be downloaded and installed from a network through the communication part, and / or installed from a removable medium. When the computer program is executed by a processor (CPU), the above functions defined in the method of the present application are performed. It should be noted that the computer-readable medium described in the present application can be a computer-readable signal medium or a computer-readable medium or any combination of the two. The computer-readable medium can be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination of the above. More specific examples of the computer-readable medium can include, but are not limited to: an electrical connection having one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present application, the computer-readable medium can be any tangible medium that contains or stores a program that can be used by or in conjunction with an instruction execution system, apparatus, or device. And in the present application, the computer-readable signal medium can include a data signal propagated in a baseband or as part of a carrier wave, in which the computer-readable program code is carried. Such a propagated data signal can take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. The computer-readable signal medium can also be any computer-readable medium other than the computer-readable medium, which can send, propagate, or transmit a program for use by or in conjunction with an instruction execution system, apparatus, or device. The program code contained on the computer-readable medium can be transmitted by any appropriate medium, including but not limited to: wireless, wire, optical cable, RF, etc., or any suitable combination of the above.

[0164] Computer program code for performing the operations of the present application can be written in one or more programming languages or combinations thereof. The programming languages include object-oriented programming languages such as Java, Smalltalk, C++, and also include conventional procedural programming languages such as the "C" language or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, executed as an independent software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server.

[0165] In the case of a remote computer, the remote computer can be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or can be connected to an external computer (e.g., by connecting through the Internet using an Internet service provider).

[0166] In yet another specific embodiment of the present invention, the present invention provides a computer-readable storage medium, characterized in that a computer program is stored on the computer-readable storage medium, and when the computer program is executed by a processor, the steps of the above-described joint optimization method of communication, sensing, and computing for crowdsensing are implemented.

[0167] The computer-readable medium can be included in the computer or terminal device described in the above embodiments; or can exist separately without being assembled into the computer device. The above computer-readable medium carries one or more programs, and when the one or more programs are executed by the computer device, the computer device: establishes a joint optimization algorithm for communication, sensing, and computing based on a sensing platform; obtains a communication, sensing, and computing strategy for a sensing task by using the joint optimization algorithm based on the current user status information and network resources; and based on the communication, sensing, and computing strategy, the terminals participating in the sensing task collect, transmit, and calculate sensing data.

[0168] The foregoing storage medium includes: various media such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disc that can store program code.

[0169] The specific embodiments described above further elaborate on the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above description is only the specific embodiments of the present invention and is not used to limit the protection scope of the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A joint optimization method for communication, sensing, and computing oriented to crowd-sourced sensing, characterized in that , including the following steps: Establish a joint communication-sensing-computation optimization algorithm based on the sensing platform; Based on the current user status information and network resources, use the joint communication-sensing-computation optimization algorithm to obtain the communication-sensing-computation strategy for the sensing task; Based on the communication-sensing-computation strategy, the terminals participating in the sensing task collect, transmit, and calculate the sensing data; The joint communication-sensing-computation optimization algorithm includes: a first algorithm for solving the data sensing strategy, data transmission strategy, and data calculation strategy in the partial offloading computation mode based on a given user selection strategy and bandwidth allocation strategy; a second algorithm for solving the optimal number of bandwidth units allocated to each user based on the dynamic programming method; a third algorithm for solving the data sensing strategy, data transmission strategy, and data calculation strategy in the binary offloading computation mode based on a given user selection strategy and bandwidth allocation strategy; In obtaining the communication-sensing-computation strategy for the sensing task using the joint communication-sensing-computation optimization algorithm, it includes: obtaining the optimal bandwidth allocation strategy in the partial offloading computation mode using the second algorithm; determining the user selection strategy according to the bandwidth allocation strategy; where when a user is allocated a certain amount of bandwidth, it means this user is selected to execute this sensing task; otherwise, it is not selected; according to the bandwidth allocation and user selection strategies, obtaining the data sensing strategy, data transmission strategy, and data processing strategy using the first algorithm; In determining the data sensing strategy, data transmission strategy, and data processing strategy according to the bandwidth allocation and user selection strategies, it further includes: obtaining the optimal bandwidth allocation strategy in the binary offloading computation mode using the second algorithm; determining the user selection strategy according to the bandwidth allocation strategy; where when a user is allocated a certain amount of bandwidth, it means this user is selected to execute this sensing task; otherwise, it is not selected; according to the bandwidth allocation and user selection strategies, obtaining the data sensing strategy, data transmission strategy, and data processing strategy using the third algorithm; Among them, the algorithm solution is transformed into a joint optimization problem, that is, under the conditions of time constraint, energy constraint and bandwidth constraint, jointly optimize and design the user selection variable , bandwidth allocation variable , sensing time , local processing time of sensed data , time to transmit sensed data to the server and transmit power to maximize the amount of sensed data processed in the sensing task; User selected variables Determined by the bandwidth allocation strategy, that is, , so remove the variable from the above optimization problem ; The amount of data processed in the perception task cannot exceed the amount of perception data collected by all selected users, so that users The perception rate, that is, the amount of perception data collected per second, is , then the user is The amount of perception data collected during the time is ;make For computing users The number of CPU cycles required to collect one bit of sensory data is the number of CPU cycles required to collect one bit of sensory data. exist The amount of data processed locally during this time is ,in For users Local computing power or CPU frequency of the user exist The amount of data transmitted to the server for processing within a certain period of time is ,in For users The uplink transmission rate is For users The allocated upstream bandwidth, is the minimum bandwidth allocation unit; in order to maximize the amount of data processed in the perception task, there must be a constraint that the amount of data processed is equal to the amount of data collected, that is, in the partial offloading computing mode, , while in the binary offloading computing mode, ,in is a binary variable used to characterize the user Whether the perception data is processed locally, that is, the variable can be determined by other parameters, so we remove the variable ; Consider two computation modes, namely the partial offloading and binary offloading modes; Partial unloading computing mode, based on data perception time , the following optimization problem can be obtained: P1： （4） s.t. （4a） （4b） （4c） （4d） （4e） Among them, Equation (4) is the objective function, that is, to maximize the weighted sum of the amount of data processed in the sensing task or the number of computed bits, where is the weighting coefficient, which is related to the type of sensing data; Equations (4a) and (4b) are the time constraints of the sensing task, that is, it is required that and ; Equation (4c) is the energy constraint of the sensing device, where is the energy consumed by the user to collect one bit of data, is the energy consumed by the user to locally process the sensing data, while is the energy consumed by the user to transmit the sensing data; Equation (4d) is the bandwidth constraint condition, which requires that the uplink bandwidth allocated to the user is an integer multiple of the minimum allocated bandwidth , that is, is an integer; Equation (4e) is the transmit power constraint of the user, is the maximum transmit power; when the optimal solution of the above optimization problem is obtained, the best user selection strategy is determined by , and the best sensing time strategy is determined by ; In the binary offloading computation mode, based on the data sensing time , there is an optimization problem P2 as shown below: P2： （5） s.t. （5a） （5b） （5c） Among them, formulas (5), (5a)-(5c) are the objective function, the perception task time constraint, and the energy constraint of the perception device in the binary offloading calculation mode; an additional binary variable is introduced in the binary offloading calculation mode used to represent whether the perception data of the user is processed locally. In this mode and cannot be greater than 0 at the same time.

2. The joint optimization method of communication, sensing, and computing for crowd-sensing according to claim 1, wherein Before establishing the joint communication-sensing-computation optimization algorithm based on the sensing platform, it includes: The user reports the current status information to the sensing platform; Use the sensing platform to obtain the sensing task with time constraints.

3. The joint optimization method of communication, sensing and computing for crowd-sensing according to claim 2, wherein In the user reporting the current status information to the sensing platform, the status information includes channel status information and user ability information, where: The terminal sends a sounding reference signal to help the base station obtain the channel status information of each user, and the base station reports the channel status information to the sensing platform; The sensing platform periodically queries the terminals through the base station to obtain the user ability information reported by the terminals.

4. The joint optimization method of communication, sensing and computing for crowd-sensing according to claim 1, wherein In the terminals participating in the sensing task collecting, transmitting, and calculating the sensing data based on the communication-sensing-computation strategy, it includes: The sensing platform issues the communication-sensing-computation strategy to the base station; The base station accesses and authorizes the terminals participating in the sensing task; The terminals participating in the sensing task determine the sensing time, sensing data transmission time, transmit power, local processing time, and local processing result transmission time information, and collect, transmit, and calculate the sensing data.

5. The joint optimization method of communication, sensing and computing for crowd sensing according to claim 4, characterized in that After the terminal participating in the sensing task collects, transmits, and calculates sensing data, it further includes: The base station then transmits the sensing data or the processing result of the sensing data to the sensing platform to complete this sensing task.

6. A computer device, characterized in that, It includes: A memory for storing computer programs; A processor for implementing the steps of the joint optimization method of communication, sensing, and computing for crowdsensing as described in any one of claims 1 to 5 when executing the computer program.

7. A computer-readable storage medium, characterized in that, A computer program is stored on the computer-readable storage medium, and when the computer program is executed by the processor, the steps of the joint optimization method of communication, sensing, and computing for crowdsensing as described in any one of claims 1 to 5 are implemented.

Images