Sensing and communication integrated resource allocation method based on multi-device cooperation

Abstract

The invention particularly relates to a sensing and communication integrated resource allocation method based on multi-device cooperation. The method comprises the following steps: defining bandwidth constraint conditions of ISAC devices at a time slot m according to fixed bandwidths of cellular users, the number of the cellular users at the time slot m and a total bandwidth; energy consumption constraint conditions of the ISAC equipment in M continuous time slots are defined according to the time when the ISAC equipment executes a sensing mode and a communication mode in each time slot, the transmitting power of the ISAC equipment in a single time slot and the total energy budget of the ISAC equipment in m time slots; defining a maximized discrimination gain problem based on each constraint condition, wherein the maximized discrimination gain problem comprises joint optimization of sensing power, transmitting power, quantization gain and communication bandwidth; optimizing the sensing power, the transmitting power and the quantization gain in a first stage based on a predefined bandwidth condition; and allocating the communication bandwidth based on the optimization result of the first stage in the second stage. According to the method, dynamic resource allocation can be realized, and sensing and communication integrated resource allocation adapting to multiple devices and multiple time slots can be realized.

Description

Technical Field

[0001] The present invention relates to the field of communication technology, and in particular to a method for allocating sensory integrated resources based on multi-device collaboration, a storage medium, a computer program product, and an electronic device. Background Art

[0002] In related technologies, with the rise of Integrated Sensing and Communication (ISAC) technology, the field of wireless communications has gradually evolved from the traditional single communication function to the multi-function of perception and communication integration. By sharing hardware modules and spectrum resources, ISAC technology can achieve the dual goals of environmental perception and information transmission without increasing additional spectrum consumption. Therefore, ISAC technology shows great application potential in fields such as intelligent transportation, environmental monitoring and unmanned driving. However, although ISAC technology can theoretically achieve higher spectrum efficiency and system performance, it still faces the challenge of perception and communication resource allocation in practical applications. In the design of resource allocation, existing studies often regard perception and communication as two independent processes for allocation and optimization. This strategy ignores the coupling relationship between the two in time and space, resulting in limited improvement in system performance.

[0003] In addition, most existing ISAC technologies adopt resource allocation strategies based on single devices or static scenarios, and fail to fully consider the resource competition and dynamic change characteristics of multiple devices working together in complex environments.

[0004] It should be noted that the information disclosed in the above background technology section is only used to enhance the understanding of the background of the present invention, and therefore may include information that does not constitute the prior art known to ordinary technicians in the field. Summary of the invention

[0005] The present invention provides a sensory integrated resource allocation method based on multi-device collaboration, a storage medium, a computer program product, and an electronic device, thereby overcoming the defects existing in the prior art to a certain extent.

[0006] Other features and advantages of the present invention will become apparent from the following detailed description, or may be learned in part by practice of the present invention.

[0007] According to a first aspect of the present invention, a method for allocating sensory integrated resources based on multi-device collaboration is provided, the method comprising:

[0008] According to the fixed bandwidth B of cellular users 0 , the number of cellular users u(m) at time slot m, and the total bandwidth B define the bandwidth constraint of the ISAC device at time slot m;

[0009] According to the time when the ISAC device executes the sensing mode and the communication mode in each time slot, the transmission power of the ISAC device in a single time slot, and the total energy budget of the ISAC device in m time slots, the energy consumption constraint of the ISAC device in M ​​consecutive time slots is defined;

[0010] Based on various constraints, the problem of maximizing the discriminant gain is defined, including the joint optimization of the sensing power, the transmission power, the quantization gain and the communication bandwidth;

[0011] Based on the predefined bandwidth conditions, the sensing power, the transmission power and the quantization gain are optimized in the first stage; and in the second stage, the communication bandwidth is allocated based on the optimization result of the first stage.

[0012] In some exemplary embodiments, optimizing perceived power in a first stage includes:

[0013] According to the number of action categories of the target to be identified and the posterior mean of the features under each action category, the duration for which the ISAC device executes the sensing mode in time slot m is combined with the Lagrange multiplier to optimize the sensing power.

[0014] In some exemplary embodiments, optimizing the transmit power in the first stage includes:

[0015] According to the channel gain of the ISAC device at time slot m, the bandwidth of the ISAC device at time slot m, and the Lagrange multiplier, the transmission power is optimized.

[0016] In some exemplary embodiments, optimizing the quantization gain in the first stage includes:

[0017] Based on the posterior distribution of the features obtained at time slot m, the number of action categories and performance requirements corresponding to the target to be identified, the quantization gain is optimized in combination with the Lagrange multiplier.

[0018] In some exemplary embodiments, the method further comprises:

[0019] The Lagrange multiplier is updated based on the currently obtained optimization results of the perceived power, quantization gain and transmit power, and the perceived power, quantization gain and transmit power are iterated based on the updated Lagrange multiplier until convergence.

[0020] In some exemplary embodiments, allocating the communication bandwidth in the second stage based on the optimization result of the first stage includes:

[0021] Based on bandwidth constraints and energy consumption constraints, combined with the communication bandwidth and communication time of each ISAC device at time slot m, the average bandwidth allocation result of the ISAC device in the time slot is calculated;

[0022] The communication bandwidth allocation result is optimized by combining the number of cellular users, the fixed bandwidth of cellular users, the average bandwidth allocation result, the difference between the transmission overhead and the maximum communication volume.

[0023] In some exemplary embodiments, the method further comprises:

[0024] The constraint condition for defining the transmission overhead of the ISAC device in the time slot m includes: the transmission overhead is less than the maximum communication volume in the communication duration in one time slot.

[0025] In some exemplary embodiments, the method further comprises:

[0026] Each ISAC device responds to the sensing request of the edge server, executes the sensing mode, and transmits a frequency modulated continuous wave FMCW signal in a sensing time interval; the sensing time interval includes M consecutive time slots;

[0027] The ISAC device receives the echo signal, extracts features from the echo signal, quantizes the extracted action feature data corresponding to the target to be identified, and generates a corresponding quantized feature set; wherein the quantized feature set includes: the quantized gain and quantized distortion of the ISAC device at time slot m;

[0028] The ISAC device switches to a communication mode and sends the quantitative feature set to the edge server.

[0029] In some exemplary embodiments, the method further comprises:

[0030] The edge server receives the quantitative feature set fed back by each ISAC device; wherein each ISAC device shares a channel with the current cellular communication user through orthogonal frequency division multiple access;

[0031] Feature data integration processing is performed according to the quantitative feature sets corresponding to multiple consecutive time slots to obtain the action recognition result corresponding to the target to be recognized.

[0032] According to a second aspect of the present invention, there is provided a computer program product having a computer program stored thereon, which implements the above-mentioned method for allocating sensory integrated resources based on multi-device collaboration when the computer program is executed by a processor.

[0033] According to a third aspect of the present invention, there is provided a storage medium having a computer program stored thereon, wherein when the computer program is executed by a processor, the above-mentioned sensory integrated resource allocation method based on multi-device collaboration is implemented.

[0034] According to a fourth aspect of the present invention, there is provided an electronic device, comprising:

[0035] Processor; and

[0036] A memory, configured to store executable instructions of the processor;

[0037] Among them, the processor is configured to implement the above-mentioned sensory integrated resource allocation method based on multi-device collaboration by executing the executable instructions.

[0038] The method for allocating integrated sensing resources based on multi-device collaboration provided by the embodiment of the present invention defines the bandwidth constraint of the ISAC device at time slot m according to the fixed bandwidth of the cellular user, the number of cellular users at time slot m, and the total bandwidth; defines the energy consumption constraint of the ISAC device in M ​​consecutive time slots according to the time when the ISAC device executes the sensing mode and the communication mode in each time slot, the transmission power of the ISAC device in a single time slot, and the total energy budget of the ISAC device in m time slots; defines the maximum discrimination gain problem based on each constraint, including the joint optimization of the sensing power, the transmission power, the quantization gain and the communication bandwidth; optimizes the sensing power, the transmission power and the quantization gain in the first stage based on the predefined bandwidth conditions; and allocates the communication bandwidth in the second stage based on the optimization result of the first stage. The method optimizes the energy allocation, the quantization gain and the bandwidth allocation through a two-stage resource allocation mode, realizes dynamic resource allocation, can realize the sensing integrated resource allocation adapting to multiple devices and multiple time slots, and can obtain the optimal resource allocation scheme for each ISAC device in each time slot, thereby improving the reasoning accuracy of downstream tasks.

[0039] It is to be understood that the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] The accompanying drawings herein are incorporated into and constitute a part of the specification, illustrate embodiments consistent with the present invention, and together with the specification are used to explain the principles of the present invention. Obviously, the accompanying drawings described below are only some embodiments of the present invention, and for those of ordinary skill in the art, other accompanying drawings can be obtained based on these accompanying drawings without creative work.

[0041] Figure 1 A schematic diagram schematically illustrates an exemplary embodiment of the present invention, a method for allocating sensory integrated resources based on multi-device collaboration;

[0042] Figure 2 A schematic diagram schematically illustrates a method for an ISAC device to process a perception request of an edge server according to an exemplary embodiment of the present invention;

[0043] Figure 3 A schematic diagram schematically illustrates a method for an edge server to process a quantitative feature set according to an exemplary embodiment of the present invention;

[0044] Figure 4 A schematic diagram schematically illustrates a multi-device collaborative human motion recognition system in an exemplary embodiment of the present invention;

[0045] Figure 5 Schematically showing a perceptual communication FMCW waveform diagram of an ISAC device in an exemplary embodiment of the present invention;

[0046] Figure 6 A schematic diagram schematically illustrates a variation trend of human motion recognition accuracy with data collection time in an exemplary embodiment of the present invention;

[0047] Figure 7 A schematic diagram schematically illustrates a variation trend of human motion recognition accuracy with network scale in an exemplary embodiment of the present invention;

[0048] Figure 8 A schematic diagram schematically illustrates a human motion recognition accuracy rate under different bandwidth budgets in an exemplary embodiment of the present invention;

[0049] Fig. 9 A schematic diagram schematically illustrates a human motion recognition accuracy rate under different energy budgets in an exemplary embodiment of the present invention;

[0050] Fig.10 The figure schematically shows the composition of an electronic device in an exemplary embodiment of the present invention. DETAILED DESCRIPTION

[0051] Example embodiments will now be described more fully with reference to the accompanying drawings. However, example embodiments can be implemented in many forms and should not be construed as limited to the examples set forth herein; rather, these embodiments are provided so that the present invention will be more comprehensive and complete and fully convey the concepts of the example embodiments to those skilled in the art. The described features, structures, or characteristics may be combined in any suitable manner in one or more embodiments.

[0052] In addition, the accompanying drawings are only schematic illustrations of the present invention and are not necessarily drawn to scale. The same reference numerals in the figures represent the same or similar parts, and their repeated description will be omitted. Some of the block diagrams shown in the accompanying drawings are functional entities and do not necessarily correspond to physically or logically independent entities. These functional entities can be implemented in software form, or implemented in one or more hardware modules or integrated circuits, or implemented in different networks and / or processor devices and / or microcontroller devices.

[0053] In the related technologies, for the problem of integrated sensory resource allocation in multi-device collaborative scenarios, the existing technologies still have significant deficiencies in the following aspects: First, the importance and communication requirements of the perception data between multiple devices are different, and how to effectively allocate them under limited resources to improve the overall performance of the system is still a problem. Because the position and perspective of different devices have obvious differences in the perception accuracy of the target, if the perception contribution of each device is not considered when allocating resources, it will lead to resource waste of some devices or resource shortage of some devices. In addition, since the resource allocation results of different time slots will affect the resource availability and overall reasoning performance of subsequent time slots, the existing technologies lack systematic modeling and analysis methods in the joint optimization across time slots, resulting in the resource allocation strategy of each time slot being carried out independently and unable to achieve the global optimum. It can be seen that designing a sensory integrated resource allocation method that can adapt to multiple devices and multiple time slots is the key to improving the performance of the ISAC system.

[0054] In view of the shortcomings and deficiencies of the prior art, a method for allocating sensory integrated resources based on multi-device collaboration is provided in this example implementation. Below, each step of the method for allocating sensory integrated resources based on multi-device collaboration in this example implementation will be described in more detail in conjunction with the accompanying drawings and embodiments.

[0055] In this example implementation, the integrated sensing and communication process of multiple ISAC devices can be modeled, and the problem can be expressed as a joint sensing and communication resource allocation problem, which includes the resources that each ISAC device needs to allocate in each time slot, namely, sensing power, transmission power, quantization gain, and communication bandwidth. Specifically, for ISAC devices, time slots can be divided into three stages: sensing, computing, and communication. Correspondingly, the corresponding models for the sensing stage and the communication stage can be defined respectively.

[0056] Specifically, the ISAC device may be in a sensing stage, in which the ISAC device transmits a radio signal and receives an echo from a target. The transmitted radio signal may be an fmcw wave. The features are extracted and quantified from the echo for transmission. The sensing signal and its processing model are described as follows.

[0057] (1) Sensing signal: During the sensing time interval of each time slot, the ISAC device transmits a chirp sequence with a linear frequency increase as the sensing signal, denoted as s k (t). In time slot m, the echo signal is:

[0058]

[0059] in, is the expected echo from the target, is the clutter signal through the jth path, J is the total number of reflected paths, and z(t) is the Gaussian noise in the receiver.

[0060] (2) Perception signal processing: echo With sampling rate f s Sampling and rearranging into a two-dimensional data matrix Where T chirp is the signal duration, T chirp f s is the length of the fast time dimension, and L is the length of the slow time dimension. In the slow time dimension, principal component analysis (PCA) is used to Extract independent N k Features.

[0061] Since all processing steps are linear, based on formula (1), the nth eigencomponent can be expressed as:

[0062]

[0063] in, and are the characteristic elements extracted from the direct echo and the clutter signal of the jth path, respectively. is the noise in the characteristic element.

[0064] By normalizing the nth feature component to the perceptual power get:

[0065]

[0066] in, represents the normalized true features, represents the normalized clutter signal.

[0067] The true features after normalization Assuming it obeys a mixed Gaussian distribution, it can be expressed as Where D is the total number of categories in the task, and are respectively the mean and variance of the d-th class distribution in the feature subspace at time slot m; this distribution is the prior distribution of the feature.

[0068] Since the interference object is assumed to have the same action category as the target, the normalized clutter It also obeys the mixed Gaussian distribution, expressed as The normalized noise has a mean of zero and a variance of Gaussian distribution.

[0069] After normalization, the features are further processed by a linear quantizer to generate a quantized version, thereby reducing communication overhead. Specifically, when using high quantization bits, the quantization result can be expressed as:

[0070]

[0071] in, and are the quantization gain and quantization distortion of ISAC device k at time slot m, and the quantization distortion Approximately has a mean of zero and a variance of Gaussian variable.

[0072] Exemplarily, a communication model corresponding to the communication mode executed by the ISAC device in the time slot may also be defined.

[0073] In the communication mode, each ISAC device transmits its quantized feature set to the edge server. In this process, the ISAC device shares the channel with existing cellular communication users through orthogonal frequency division multiple access (OFDMA). Assume that the number of subcarriers is large enough, so the bandwidth allocation can be approximated as continuous. Let B be the total bandwidth, B 0 is the fixed bandwidth allocated to each cellular user. Since the number of cellular users u(m) varies in different time slots, the total bandwidth required by all cellular users will also vary, which in turn causes the available bandwidth for ISAC devices to vary. is the bandwidth allocated to ISAC device k in time slot m. Assume that the relevant channel is a frequency-flat channel with zero mean and variance Rayleigh distribution. is the channel gain from ISAC device k to the edge server at time slot m, then the communication rate is:

[0074]

[0075] in, is the transmit power of ISAC device k in time slot m, N 0 is the power of additive white Gaussian noise (AWGN).

[0076] On the edge server side, the features of ISAC device k are recovered from the quantized features as:

[0077]

[0078] Then, the transmission overhead of ISAC device k in time slot m can be calculated by the recovered feature set With normalized feature set It is characterized by the mutual information of:

[0079]

[0080] From formula (6) and formula (7), we can see that the quantization gain Determines the distortion and transmission overhead of the restored features. Specifically, a higher quantization gain leads to a higher transmission overhead and a lower quantization distortion. Therefore, the quantization gain is used to ensure that the transmission overhead is less than the maximum communication volume. is optimized.

[0081] For example, for the edge server, an action recognition model can be predefined. By integrating the features recovered in M ​​consecutive time slots, the edge server predicts the action classification label of the target, thereby realizing action recognition. We use the discriminant gain to evaluate the action recognition accuracy of the target, which is defined as the Kullback–Leibler (KL) divergence between any two classes under given features.

[0082] After substituting formula (3) into formula (6), the recovered features can be rewritten as:

[0083]

[0084] Then, restore the features The probability density function can be expressed as:

[0085]

[0086] in and are the posterior mean and posterior variance of the features respectively.

[0087] For the restored features The discriminant gain between category d and category d′ is defined as:

[0088]

[0089] Here, KL[·||·] represents the KL divergence of one probability distribution with respect to another probability distribution.

[0090] This discriminant gain is extended to the feature sets of K ISAC devices and M time slots, recovering the feature The overall discriminant gain of It is expressed as:

[0091]

[0092] It can be seen from formula (10) and formula (11) that the discriminant gain depends on the probability distribution of the recovered features, which is determined by the perception process and the communication process. In particular, the perception power affects the quality of the extracted features, see formula (3). At the same time, the quantization gain is affected by the communication constraints and affects the quality of the recovered features, see formula (6). Therefore, the device needs to jointly optimize the resources in the perception process and the communication process to maximize the action recognition accuracy, that is, the discriminant gain.

[0093] In this example implementation, reference Figure 1 As shown, the sensory integrated resource allocation method based on multi-device collaboration specifically includes:

[0094] Step S11, defining a bandwidth constraint condition of the ISAC device at time slot m according to the fixed bandwidth of the cellular user, the number of cellular users at time slot m, and the total bandwidth;

[0095] Step S12, defining energy consumption constraints of the ISAC device in M ​​consecutive time slots according to the time when the ISAC device executes the sensing mode and the communication mode in each time slot, the transmission power of the ISAC device in a single time slot, and the total energy budget of the ISAC device in m time slots;

[0096] Step S13, defining a problem of maximizing the discrimination gain based on various constraints, including joint optimization of perception power, transmission power, quantization gain and communication bandwidth;

[0097] Step S14, based on the predefined bandwidth conditions, optimizing the perceived power, the transmit power and the quantization gain in the first stage; and allocating the communication bandwidth in the second stage based on the optimization result of the first stage.

[0098] Specifically, the task-oriented ISAC design goal is to maximize the target's motion recognition accuracy while meeting the communication needs of cellular users. Since the ISAC device shares the entire bandwidth with the cellular user, the bandwidth constraint of the ISAC device at time slot m is:

[0099]

[0100] Among them, B 0 is the fixed bandwidth allocated to each cellular user to meet its communication needs, u(m) is the number of cellular users at time slot m, and B is the total bandwidth. Due to the change in the number of existing cellular users, the available bandwidth of the ISAC device in different time slots also changes.

[0101] For each ISAC device, its energy consumption constraint in M ​​consecutive time slots is:

[0102]

[0103] Among them, T s and T c are the sensing and communication time in each time slot, which is a constant; E k is the total energy budget of ISAC device k in m time slots.

[0104] In addition, to ensure the successful transmission of the feature, the transmission overhead needs to be less than the communication duration T in one time slot. c The maximum amount of communication in is expressed as:

[0105]

[0106] Under the above constraints, the problem of maximizing the discriminant gain is stated as follows:

[0107]

[0108] Among them, the perceived power Transmit power Quantization gain and communication bandwidth Joint optimization is required. Due to the non-convexity of the objective function, solving this problem is challenging. To solve this problem, we propose a new TSO (Two-Stage Offline Optimization Algorithm) algorithm to solve it.

[0109] Exemplarily, the perception power is optimized in the first stage, including: optimizing the perception power in combination with Lagrange multipliers according to the number of action categories of the target to be identified, the posterior mean of the features under each action category, and the duration for which the ISAC device executes the perception mode in time slot m.

[0110] Exemplarily, optimizing the transmit power in the first stage includes optimizing the transmit power according to the channel gain of the ISAC device at time slot m, the bandwidth of the ISAC device at time slot m, and the Lagrange multiplier.

[0111] Exemplarily, optimizing the quantization gain in the first stage includes: optimizing the quantization gain based on the posterior distribution of the features acquired at time slot m, the number of action categories corresponding to the target to be identified and the performance requirements, combined with Lagrange multipliers.

[0112] Exemplarily, the method further includes: updating the Lagrange multiplier based on currently obtained optimization results of the perceived power, quantization gain and transmit power, and iterating the perceived power, quantization gain and transmit power based on the updated Lagrange multiplier until convergence.

[0113] Specifically, first, problem P1 is transformed into a convex optimization problem, and a two-stage optimization algorithm (TSO) is proposed to solve it. Since the objective function of P1 is a ratio in the form of summation, it can be transformed into an equivalent subtraction form. The converted problem is:

[0114]

[0115] in, is a variable transformation to make the constraints of P1 convex. Here, is an auxiliary variable.

[0116] function is defined as:

[0117]

[0118] To solve problem P1, P2 can be solved with given auxiliary variables, and the auxiliary variables can be updated according to the solution of P2, and alternating iterations can be performed until convergence. For the solution of problem P2, the first stage optimizes the perceived power, transmission power and quantization gain under a feasible bandwidth; in the second stage, the bandwidth is optimized based on the above optimization results.

[0119] Specifically, in the first stage, the power and quantization gain allocation is realized. Specifically, in problem P2, given a feasible communication bandwidth, the joint optimization problem of power allocation and quantization gain is:

[0120]

[0121] This is a convex optimization problem and the solution is as follows:

[0122] Theorem 1: The channel gain when ISAC device k is in time slot m is The optimal sensing power and transmission power for problem P2.1 are:

[0123]

[0124] Where D is the number of action categories, is the posterior mean of the feature under this category, N 0 is the AWGN power, α k and β k，m is the Lagrange multiplier, α k ≥0 and β k,m ≥0,

[0125] Proof: Since problem P2.1 is convex, it can be solved by using the Karush-Kuhn-Tucker (KKT) condition.

[0126] From Theorem 1, it can be seen that the optimal perception power depends on the posterior distribution of the features obtained in the time slot and the characteristics of the human action recognition task, that is, the number of action categories and the required performance. In particular, when the difference in the posterior mean of the features between the two categories is large, it means that the influence of the interfering object is small and it is easy to distinguish the two categories from the features. At this time, the required perception power can be reduced. In addition, the optimal transmission power depends on the wireless channel state of the time slot. Since the total energy constraint of M consecutive time slots is given, the power allocation between time slots is interrelated, so it is necessary to jointly allocate the perception power and transmission power among all time slots.

[0127] Theorem 2: The optimal quantization gain of ISAC device k in time slot m is:

[0128]

[0129] in,

[0130] Proof: Please refer to Theorem 1 for proof.

[0131] From Theorem 2, we can see that the optimal quantization gain depends on the posterior distribution of the features obtained in the time slot and the requirements of the overall human action recognition task, that is, the number of action categories and performance requirements. Similarly, when the difference in the posterior mean of the features between the two categories is large, the required quantization gain is small, and vice versa.

[0132] After obtaining the optimal sensing power, quantization gain, and transmission power, the Lagrange multiplier is updated to complete an iteration. This iterative process continues until convergence, thus solving problem P2.1.

[0133] For example, the allocating of communication bandwidth in the second stage based on the optimization result of the first stage includes:

[0134] Based on bandwidth constraints and energy consumption constraints, combined with the communication bandwidth and communication time of each ISAC device at time slot m, the average bandwidth allocation result of the ISAC device in the time slot is calculated;

[0135] The communication bandwidth allocation result is optimized by combining the number of cellular users, the fixed bandwidth of cellular users, the average bandwidth allocation result, the difference between the transmission overhead and the maximum communication volume.

[0136] Specifically, based on the optimization of the perceived power, transmission power, and quantization gain in the previous stage, the second stage can obtain the optimal communication bandwidth allocation between ISAC devices. From the constraints of problem P2, it can be seen that the allocated bandwidth needs to meet the successful transmission of the characteristics. In order to ensure that there is a feasible solution to problem P2, problem P2.2 can be expressed as follows:

[0137]

[0138] This is a convex optimization problem, which can be solved by Based on the bandwidth allocation results obtained in P2.2, the bandwidth will be updated according to Theorem 3.

[0139] Theorem 3: When there are u(m) cellular users sharing bandwidth B with the ISAC device in time slot m, the optimal bandwidth allocated to ISAC device k is:

[0140]

[0141] Among them, B 0 is the fixed bandwidth allocated to each cellular user, is the total bandwidth obtained through Problem P2.2, Represents the difference between the required transmission overhead and the maximum communication volume.

[0142] From Theorem 3, we can see that when When , the solution to problem P2.2 is the optimal bandwidth allocation, that is, when When , more remaining bandwidth will be allocated to the ISAC device to improve the discrimination gain, that is, Otherwise, the ISAC device's allocated bandwidth will be reduced to meet the bandwidth constraint, i.e. In addition, when the difference The larger the value, the larger the bandwidth allocation update. In other words, when the maximum traffic volume is much smaller than the required transmission overhead, more bandwidth will be allocated to the ISAC device. Figure 8 As shown in Figure 2, the accuracy of human action recognition under different bandwidth budgets varies.

[0143] Since the TSO algorithm is static, it needs to obtain information of all time slots in advance. The TSO algorithm can maximize the recognition accuracy in a dynamic environment by jointly optimizing resources in all time slots. Fig. 9 As shown, the accuracy of human action recognition under different energy budgets varies.

[0144] Exemplary, reference Figure 2 As shown, the method also includes:

[0145] Step S21, each ISAC device executes a sensing mode in response to a sensing request from the edge server, and transmits a frequency modulated continuous wave FMCW signal within a sensing time interval; the sensing time interval includes M consecutive time slots;

[0146] Step S22, the ISAC device receives the echo signal, extracts features from the echo signal, quantizes the extracted motion feature data corresponding to the target to be identified, and generates a corresponding quantized feature set; wherein the quantized feature set includes: the quantized gain and quantized distortion of the ISAC device at time slot m;

[0147] Step S23: The ISAC device switches to a communication mode and sends the quantitative feature set to an edge server.

[0148] Exemplary, reference Figure 3 As shown, the method also includes:

[0149] Step S31, the edge server receives the quantitative feature set fed back by each ISAC device; wherein each ISAC device shares a channel with the current cellular communication user through orthogonal frequency division multiple access;

[0150] Step S32, performing feature data integration processing according to the quantitative feature sets corresponding to the multiple consecutive time slots to obtain the action recognition result corresponding to the target to be recognized.

[0151] Exemplary, reference Figure 4 As shown in Figure 2, consider a multi-device collaborative human action recognition system. In this system, K A single-antenna device performs radar perception to obtain different views of the target and transmits its observed features to an edge server. By fusing the received features, the edge server recognizes the target's motion. Assume that some unknown interfering objects move around the target, resulting in different perception qualities between devices and on different time slots. Each device is equipped with a dual-function radar and communication (DFRC) transceiver, called an integrated sensing and communication (ISAC) device. The ISAC device can switch between perception mode and communication mode in a time-division manner. This means that a time slot is divided into a perception time interval and a communication time interval. When a perception request is received from the edge server, the data collection process starts and continues M In each time slot, the ISAC device first performs a sensing mode. Specifically, it transmits a frequency modulated continuous wave (FMCW) signal during the sensing time interval. Figure 6 As shown in the figure, the accuracy of human action recognition varies with the length of data collection. By processing the echo, the ISAC device extracts and quantifies the features that capture the target's action information. The ISAC device then switches to communication mode and transmits its quantified features to the edge server. After data collection is completed, the edge server summarizes all received features to identify the target's action.

[0152] The DFRC transceiver uses a shared beamforming front-end circuit to switch between sensing mode and communication mode in a flexible manner, thereby achieving the integrated fusion of sensing and communication. Figure 5 As shown in the figure, the dual-function radar-communication (DFRC) transceiver flexibly switches between the perception mode and the communication mode in a time-division manner by sharing the RF front-end circuit, thereby realizing the integration of perception and communication (ISAC). In the perception mode, the DFRC system transmits a frequency modulated continuous wave (FMCW) signal composed of multiple uplink ramp chirps. Subsequently, by processing the received radar echo signal, the ISAC device can obtain perception data containing the motion information of the perceived target. In the communication mode, the DFRC system adopts a digital modulation scheme (for example, Quadrature Amplitude Modulation (QAM)) for signal transmission. Through flexible mode switching, the DFRC transceiver can achieve high-precision detection of the perceived target while meeting the communication requirements, fully reflecting the versatility and resource utilization efficiency of the ISAC system. Reference Figure 7 As shown, the accuracy of human action recognition varies with the size of the network.

[0153] The method provided by the embodiment of the present invention can be applied to a multi-device collaborative human motion recognition system, which has a dynamic interference and time-varying network state, and is composed of multiple devices and edge servers with limited energy and bandwidth. These devices obtain perception data from the received signals in continuous time periods and transmit them to the edge server to complete the motion recognition task of the mobile target. The solution of the present invention proposes a novel resource allocation problem, which optimizes the perception and communication resource allocation of multiple devices in multiple time periods, aiming to maximize the accuracy of human motion recognition. Specifically, the accuracy of human motion recognition is measured by the discriminant gain theory. A two-stage offline (TSO) algorithm is proposed to solve the resource allocation problem proposed by us, optimize energy allocation, quantization gain and bandwidth allocation, and dynamically allocate resources, thereby maximizing the accuracy of human motion recognition. Extensive numerical results show that the proposed algorithm is superior to the comparison algorithm in the overall performance of human motion recognition.

[0154] It should be noted that the above figures are only schematic illustrations of the processes included in the method according to an exemplary embodiment of the present invention, and are not intended to be limiting. It is easy to understand that the processes shown in the above figures do not indicate or limit the time sequence of these processes. In addition, it is also easy to understand that these processes can be performed synchronously or asynchronously, for example, in multiple modules.

[0155] It should be noted that, although several modules or units of the equipment for action execution are mentioned in the above detailed description, this division is not mandatory. In fact, according to an embodiment of the present invention, the features and functions of two or more modules or units described above can be embodied in one module or unit. On the contrary, the features and functions of one module or unit described above can be further divided into being embodied by multiple modules or units.

[0156] Fig.10 A schematic diagram of an electronic device suitable for implementing an embodiment of the present invention is shown.

[0157] It should be noted that Fig.10 The electronic device 1000 shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present invention.

[0158] like Fig.10 As shown, the electronic device 1000 includes a central processing unit (CPU) 1001, which can perform various appropriate actions and processes according to the program stored in the read-only memory (ROM) 1002 or the program loaded from the storage part 1008 to the random access memory (RAM) 1003. In the RAM 1003, various programs and data required for system operation are also stored. The CPU 1001, ROM 1002 and RAM 1003 are connected to each other through a bus 1004. An input / output (I / O) interface 1005 is also connected to the bus 1004.

[0159] The following components are connected to the I / O interface 1005: an input section 1006 including a keyboard, a mouse, etc.; an output section 1007 including a cathode ray tube (CRT), a liquid crystal display (LCD), etc., and a speaker, etc.; a storage section 1008 including a hard disk, etc.; and a communication section 1009 including a network interface card such as a LAN (Local Area Network) card, a modem, etc. The communication section 1009 performs communication processing via a network such as the Internet. A drive 1010 is also connected to the I / O interface 1005 as needed. A removable medium 1011, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc., is installed on the drive 1010 as needed, so that a computer program read therefrom is installed into the storage section 1008 as needed.

[0160] In particular, according to an embodiment of the present invention, the process described below with reference to the flowchart can be implemented as a computer software program. For example, an embodiment of the present invention includes a computer program product, which includes a computer program carried on a storage medium, and the computer program includes a program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from a network through a communication part 1009, and / or installed from a removable medium 1011. When the computer program is executed by a central processing unit (CPU) 1001, various functions defined in the system of the present application are executed.

[0161] It should be noted that the storage medium shown in the embodiment of the present invention may be a computer-readable signal medium or a computer-readable storage medium or any combination of the above two. The computer-readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or device, or any combination of the above. More specific examples of computer-readable storage media may include, but are not limited to: an electrical connection with 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), a 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 invention, a computer-readable storage medium may be any tangible medium containing or storing a program, which may be used by or in combination with an instruction execution system, device or device. In the present invention, a computer-readable signal medium may include a data signal propagated in a baseband or as part of a carrier wave, which carries a computer-readable program code. Such propagated data signals may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. Computer-readable signal media may also be any storage medium other than computer-readable storage media, which may send, propagate, or transmit programs for use by or in conjunction with an instruction execution system, apparatus, or device. The program code contained on the storage medium may be transmitted using any appropriate medium, including but not limited to: wireless, wired, etc., or any suitable combination of the above.

[0162] The flow chart and block diagram in the accompanying drawings illustrate the possible architecture, function and operation of the system, method and computer program product according to various embodiments of the present invention. In this regard, each box in the flow chart or block diagram can represent a module, a program segment, or a part of a code, and the above-mentioned module, program segment, or a part of a code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in a different order from the order marked in the accompanying drawings. For example, two boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram or flow chart, and the combination of the boxes in the block diagram or flow chart can be implemented with a dedicated hardware-based system that performs a specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0163] The units involved in the embodiments of the present invention may be implemented by software or hardware, and the units described may also be arranged in a processor. The names of these units do not, in some cases, limit the units themselves.

[0164] It should be noted that, as another aspect, the present application also provides a storage medium, which can be included in an electronic device; or it can exist independently without being installed in the electronic device. The above storage medium carries one or more programs, and when the above one or more programs are executed by an electronic device, the electronic device implements the method described in the following embodiments. For example, the electronic device can implement the following Figure 1 The individual steps of the method are shown.

[0165] In one embodiment, the present application provides a computer program product, including a computer program, which implements the steps in the above-mentioned method embodiments when executed by a processor.

[0166] In addition, the above-mentioned figures are only schematic illustrations of the processes included in the method according to an exemplary embodiment of the present invention, and are not intended to be limiting. It is easy to understand that the processes shown in the above-mentioned figures do not indicate or limit the time sequence of these processes. In addition, it is also easy to understand that these processes can be performed synchronously or asynchronously, for example, in multiple modules.

[0167] Other embodiments of the invention will readily occur to those skilled in the art after considering the specification and practicing the invention herein. This application is intended to cover any variations, uses, or adaptations of the invention that follow the general principles of the invention and include common knowledge or customary techniques in the art that are not disclosed by the present invention. The specification and examples are to be considered exemplary only, and the true scope and spirit of the invention are indicated by the claims.

[0168] It should be understood that the present invention is not limited to the exact construction that has been described above and shown in the drawings and that various modifications and changes may be made without departing from the scope thereof. The scope of the present invention is limited only by the appended claims.

Claims

1. A sensory integrated resource allocation method based on multi-device collaboration, characterized in that: The method comprises: Define the bandwidth constraint of the ISAC device at time slot m according to the fixed bandwidth of the cellular users, the number of cellular users at time slot m, and the total bandwidth; According to the time when the ISAC device executes the sensing mode and the communication mode in each time slot, the transmission power of the ISAC device in a single time slot, and the total energy budget of the ISAC device in m time slots, the energy consumption constraint of the ISAC device in M ​​consecutive time slots is defined; Based on various constraints, the problem of maximizing the discriminant gain is defined, including the joint optimization of the sensing power, the transmission power, the quantization gain and the communication bandwidth; Based on the predefined bandwidth conditions, the sensing power, the transmission power and the quantization gain are optimized in the first stage; and in the second stage, the communication bandwidth is allocated based on the optimization result of the first stage.

2. The method according to claim 1, characterized in that: In the first stage, the perceived power is optimized, including: According to the number of action categories of the target to be identified and the posterior mean of the features under each action category, the duration for which the ISAC device executes the sensing mode in time slot m is combined with the Lagrange multiplier to optimize the sensing power.

3. The method according to claim 1, characterized in that In the first stage, the transmit power is optimized, including: According to the channel gain of the ISAC device at time slot m, the bandwidth of the ISAC device at time slot m, and the Lagrange multiplier, the transmission power is optimized.

4. The method according to claim 1, characterized in that: In the first stage, the quantization gain is optimized, including: Based on the posterior distribution of the features obtained at time slot m, the number of action categories and performance requirements corresponding to the target to be identified, the quantization gain is optimized in combination with the Lagrange multiplier.

5. The method according to any one of claims 1 to 4, characterized in that: The method further comprises: The Lagrange multiplier is updated based on the currently obtained optimization results of the perceived power, quantization gain and transmit power, and the perceived power, quantization gain and transmit power are iterated based on the updated Lagrange multiplier until convergence.

6. The method according to claim 1, characterized in that The allocation of the communication bandwidth in the second stage based on the optimization result of the first stage includes: Based on bandwidth constraints and energy consumption constraints, combined with the communication bandwidth and communication time of each ISAC device at time slot m, the average bandwidth allocation result of the ISAC device in the time slot is calculated; The communication bandwidth allocation result is optimized by combining the number of cellular users, the fixed bandwidth of cellular users, the average bandwidth allocation result, the difference between the transmission overhead and the maximum communication volume.

7. The method according to claim 6, characterized in that The method further comprises: The constraint condition for defining the transmission overhead of the ISAC device in the time slot m includes: the transmission overhead is less than the maximum communication volume in the communication duration in one time slot.

8. The method according to claim 1, characterized in that The method further comprises: Each ISAC device responds to the sensing request of the edge server, executes the sensing mode, and transmits a frequency modulated continuous wave FMCW signal in a sensing time interval; the sensing time interval includes M consecutive time slots; The ISAC device receives the echo signal, extracts features from the echo signal, quantizes the extracted action feature data corresponding to the target to be identified, and generates a corresponding quantized feature set; wherein the quantized feature set includes: the quantized gain and quantized distortion of the ISAC device at time slot m; The ISAC device switches to a communication mode and sends the quantitative feature set to the edge server.

9. The method according to claim 8, characterized in that The method further comprises: The edge server receives the quantitative feature set fed back by each ISAC device; wherein each ISAC device shares a channel with the current cellular communication user through orthogonal frequency division multiple access; Feature data integration processing is performed according to the quantitative feature sets corresponding to multiple consecutive time slots to obtain the action recognition result corresponding to the target to be recognized.

10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the sensory integration resource allocation method based on multi-device collaboration described in any one of claims 1 to 9 is implemented.

Images