A method for position information assisted D2D communication resource allocation

Abstract

This invention relates to a location-information-assisted D2D communication resource allocation method, belonging to the field of communication technology. The invention includes the following steps: S1: Calculate the interference of each D2D pair to each CU in the licensed spectrum, obtaining an interference matrix. S2: Allocate the matched D2D pairs and CUs to the same sub-channel according to the interference matrix; S3: If there are more D2D pairs than CUs, offload the excess D2D pairs and those with significant interference to the CUs to the unlicensed spectrum; S4: Obtain the optimization problem that maximizes the total system throughput; S5: Divide the optimization problem into two sub-problems; S6: Offload the D2D pairs with significant interference to the CUs sequentially to the unlicensed spectrum, obtaining the D2D spectrum selection matrix and the optimal duty cycle; S7: Repeat S6 until convergence. This invention utilizes the information inherent in the D2D pairs, and optimizes the spectrum selection and duty cycle of each D2D pair to maximize the total system throughput based on the interference of each D2D pair to the CUs in the licensed spectrum.

Description

Technical Field

[0001] This invention belongs to the field of wireless communication technology and relates to a location information-assisted D2D communication resource allocation method. Background Technology

[0002] In recent years, with the advent of the 5G era, global data traffic has grown exponentially. Moreover, as the number of users increases, the mobile traffic generated by users also varies, such as short video services and ultra-high-definition video services. This requires not only extremely high data transmission rates but also enormous network capacity to meet users' demands for network latency.

[0003] To alleviate the growth of data traffic to some extent, the 3GPP (3rd Generation Partnership Project) organization proposed Device-to-device (D2D) technology, which can improve spectrum efficiency and expand network capacity. In 5G communication systems, D2D communication, as a component, has great potential in several aspects: increasing system throughput, reducing latency, alleviating network load, and improving energy efficiency.

[0004] While introducing D2D links into cellular systems offers numerous benefits, it also causes significant interference, a major problem associated with their introduction. This interference is exacerbated when D2D links share radio resources with cellular links. In traditional LTE systems, the absence of D2D links ensures orthogonal resource allocation for all cellular user equipment, resulting in negligible intra-cell interference. However, when D2D users share resources with cellular users, the orthogonality of spectrum resources within the cell is lost, making intra-cell interference non-negligible. D2D communication not only increases intra-cell interference but also leads to higher inter-cell interference, as D2D devices at cell edges may interfere with cellular users in adjacent cells using the same radio spectrum resources. Given the severe interference caused by D2D communication to cellular systems, proper resource allocation and interference management are crucial.

[0005] However, most existing technologies are implemented on licensed spectrum, to the point that the optimization potential on licensed spectrum is nearing saturation. Therefore, researchers have turned their attention to unlicensed spectrum, proposing D2D-Unlicensed (D2D-U) technology to utilize D2D in unlicensed bands alongside WiFi users. Since unlicensed spectrum is only used by WiFi users, and WiFi traffic is real-time and unpredictable, there are some unused idle resources on unlicensed spectrum. Therefore, the problem we aim to solve is to rationally utilize these idle resources on unlicensed spectrum without affecting the normal transmission of WiFi users, thereby improving the overall system throughput.

[0006] However, most of the existing D2D-U literature focuses on system optimization in terms of power allocation and mode selection for D2D, and rarely considers which D2D is suitable for communication in licensed spectrum, or which D2D will maximize the total system throughput when communicating in unlicensed spectrum. Summary of the Invention

[0007] In view of this, the present invention provides a location information-assisted D2D communication resource allocation method to maximize the total system throughput on licensed and unlicensed spectrum.

[0008] To achieve the above objectives, the present invention provides the following technical solution:

[0009] A location-based D2D communication resource allocation method includes the following steps:

[0010] S1: Assume there are M D2D pairs and N cellular users (CUs). Divide the licensed spectrum into N orthogonal sub-channels, and allow only one CU to use each orthogonal sub-channel. Calculate the interference I of each D2D pair in the licensed spectrum to each cellular user CU. Then an N*M interference matrix can be obtained.

[0011] S2: Using a matching algorithm, the matched D2D and CU are assigned to the same sub-channel according to the interference matrix. The matching principle here is to minimize the interference of D2D to CU.

[0012] S3: If N < M, then offload the excess D2D to the unlicensed spectrum, and also offload the D2D that causes significant interference to the CU to the unlicensed spectrum, coexisting with the WiFi using the Duty Cycle (DC) mechanism.

[0013] S4: The optimization problem is to maximize the total system throughput while satisfying various constraints. These constraints are generally to ensure normal WiFi transmission.

[0014] S5: Decompose the optimization problem into two sub-problems: 1. Which D2Ds and how many D2Ds are offloaded to the unlicensed spectrum? 2. The duty cycle problem of using the DC mechanism on the unlicensed spectrum;

[0015] S6: After using the matching algorithm, the D2D that causes significant interference to the CU is sequentially offloaded to the unlicensed spectrum, resulting in the D2D spectrum selection matrix Φ, which solves problem 1. Simultaneously, the Lagrange dual problem is used to solve the duty cycle problem of the DC mechanism, obtaining the optimal duty cycle.

[0016] S7: Use the obtained Φ to optimize the duty cycle. The problem, and at the same time the result Conversely, optimize the D2D spectrum selection problem by repeating S6-S7 until the result converges.

[0017] Furthermore, in step S1, the interference matrix I can be expressed as:

[0018]

[0019] Each element represents the interference of the m-th D2D to the n-th CU, and its interference expression can be expressed as:

[0020] I n,m =P m G n,m

[0021] Where P m G represents the transmit power of the m-th D2D transmitter. n,m The channel gain from the m-th D2D transmitter to the n-th CU user is expressed as:

[0022]

[0023] Where |h n,m | represents the power gain from the m-th D2D transmitter to the n-th CU user, l n,m Let α represent the distance from the m-th D2D transmitter to the n-th CU user, and let α represent the path loss factor in the subchannel.

[0024] Furthermore, in step S2, based on the obtained interference matrix I, each CU user and its best matching D2D can be assigned to the same sub-channel according to the matching algorithm, that is, the D2D with the least interference to the CU user can be assigned to the same sub-channel.

[0025] The goal of the matching algorithm is to minimize interference to the cellular network and achieve optimal system performance by rationally allocating the licensed subchannel occupancy of D2D users. This invention assumes that the D2D user set, CU user set, and licensed subchannel set are independent and uncorrelated.m (m∈M) represents the D2D user, CU n (n∈N) are CU users, S n (n∈N) represents an unlicensed subchannel. It is assumed that exchanging information between users requires no additional signaling cost, and each user can obtain all state information from other users. The matching process assigns a suitable subchannel to each CU user, which is equivalent to generating a mapping relationship between two disjoint sets: the CU user set and the D2D user set. During the matching process, if the CU... n Matched D m , denoted as (CU n D m To ensure the basic traffic requirements of CU users, each CU user is matched with at least one licensed sub-channel, and the specific D2D pair matched by a CU user will be determined by the matching algorithm.

[0026] Furthermore, in step S3, after matching is completed, if N < M, meaning the number of D2D users is greater than the number of CU users, since each sub-channel allows at most one CU and one D2D user to be used simultaneously, the excess D2D users after matching are all offloaded to the unlicensed spectrum to coexist with WiFi users. Simultaneously, D2D users and WiFi users coexist on the unlicensed spectrum using a DC mechanism. Therefore, how to reasonably allocate their respective duty cycles to maximize the total system throughput without affecting the normal transmission of WiFi users is a problem that needs to be solved.

[0027] Furthermore, in step 4, the optimization problem of this invention can be derived, maximizing the total system throughput while satisfying various constraints. Here, throughput includes the throughput of licensed spectrum and the throughput of unlicensed spectrum. The throughput on licensed spectrum comprises the throughput of CU users and the throughput of D2D communication on selected licensed spectrum, while the throughput on unlicensed spectrum comprises the throughput of D2D communication on selected unlicensed spectrum and the throughput of WiFi users.

[0028] On the licensed spectrum: the throughput of CU users can be expressed as:

[0029]

[0030] One of the summation functions sums the number of CU users over a given number, and the other sums the number of sub-channels. Additionally, B... n P represents the sub-channel bandwidth selected by the CU user. n and P m These are the transmit powers for CU users and D2D users, respectively. G n and G n,m These are the channel gain from the CU user to the base station and the channel gain from the CU user to the D2D transmitter, respectively. n (S n=0,1) represents the sub-channel selected by the CU user, if S n =1, then the CU user selects this sub-channel for communication. If S n If φ = 0, then the CU user will not select this sub-channel for communication. m (φ m =0,1) indicates whether the D2D user matched with this CU is communicating in this sub-channel. Similarly, if φ m =1, then the D2D user selects the cellular user to communicate within the sub-channel, if φ m =0, then the D2D user selects unlicensed spectrum for communication. This represents the Gaussian noise of the channel.

[0031] Meanwhile, the throughput of D2D users on the licensed frequency band can be expressed as:

[0032]

[0033] Among them G m This represents the channel gain between the D2D transmitter and receiver. All other parameters are the same as described above. The D2D spectrum selection matrix Φ = [φ1, φ2...φ] can also be obtained. m ].

[0034] Therefore, the system throughput R on the licensed spectrum can be expressed as:

[0035] R = R C +R D

[0036] Furthermore, in unlicensed spectrum: D2D throughput It can be represented as:

[0037]

[0038] Where B represents the channel bandwidth of the unlicensed spectrum, I D This indicates interference from other D2D sources on unlicensed frequency bands, I D It can be represented as:

[0039]

[0040] Where φ i Indicates whether the D2D communication is on unlicensed spectrum, P i G represents the transmit power of the D2D transmitter. i Let represent the channel gain between the i-th D2D transmitter and the m-th D2D receiver.

[0041] Since WiFi user transmission uses the CSMA / CA protocol, the throughput of a WiFi system can be modeled as a discrete-time Markov chain mathematical model, with the steady-state probability being:

[0042]

[0043] CW min Here, n is the minimum competition window size, n is the maximum backoff stage, and p represents the conditional conflict probability, which can be expressed as:

[0044]

[0045] Assume p tr To determine the probability that at least one WiFi device is working in a given time slot, we obtain...

[0046]

[0047] Thus, the normalized throughput is obtained:

[0048]

[0049] Where E[T packet The average time interval (E[T]) represents the average time taken for a WiFi AP to successfully transmit data packets. slot It can be divided into three phases: the idle phase, the successful transmission phase, and the collision phase. Each phase is represented by P. i P s and P c The probability of each stage can be expressed as:

[0050]

[0051]

[0052] P c =P tr (1-p s )

[0053] Where N W P represents the number of WiFi networks, and the average value of the payload information successfully transmitted in a time slot is P. s P tr E[p], E[p] is a probability P r P tr The average payload for successful transmission; conversely, if a collision occurs in the considered time slot, the probability is: P tr (1-P s Finally, the WiFi throughput expression is obtained as follows:

[0054]

[0055] Where T S T is the average time that the channel is busy due to successful transmissions. c σ is the average time the channel is busy due to collisions. σ is the average time the channel is empty.

[0056] T S =H+E[P]+SIFS+δ+ACK+DIFS+δ

[0057] T C =H+E[P]+DIFS+δ

[0058] Where E[P] represents the average size of the packets that collide, H is the duration of the Media Access Control (MAC) and Physical Layer (PHY) headers of the WiFi network, δ represents the propagation delay, and DIFS, ACK, and SIFS represent the DCF inter-frame interval, transmission slot, and shorter inter-frame interval, respectively.

[0059] Therefore, the system throughput R on unlicensed spectrum U It can be represented as:

[0060]

[0061] Meanwhile, the total system throughput R total It can be represented as

[0062]

[0063] Furthermore, since D2D offloading to unlicensed spectrum will cause some interference to WiFi, in order to ensure the normal transmission of the WiFi system, the throughput of the WiFi system needs to be no less than its minimum throughput threshold, i.e.: R w ≥R wmin , where R wmin This represents the minimum throughput required for normal WiFi transmission.

[0064] Therefore, the optimization problem of this invention can be obtained:

[0065]

[0066] C1:I n,m ≤I max

[0067]

[0068] C3:R w ≥R wmin

[0069] Where C1 indicates that the interference of D2D to CU users on the licensed spectrum cannot exceed its maximum interference threshold, C2 indicates the range of D2D duty cycle values ​​on the unlicensed spectrum, and C3 indicates that the throughput of WiFi on the unlicensed spectrum cannot be less than its minimum threshold.

[0070] Furthermore, in step S5, since the objective function is complex, it is decoupled into two sub-problems for joint solution.

[0071] 1. Fixed To solve the Φ problem, the main issue is to determine which D2D operations are offloaded to unlicensed spectrum to maximize the total system throughput under constraints.

[0072] 2. Fix Φ to solve The main problem is how to allocate the duty cycle of D2D on unlicensed spectrum to maximize the total system throughput under constraints.

[0073] Furthermore, in step S6, the two sub-problems in step S5 are mainly addressed.

[0074] The interference matrix of each D2D to each CU user has been obtained, and the optimal match between D2D and CU has been obtained according to the matching algorithm. Here, we only need to unload the D2Ds that have the greatest interference to CU after matching to the unlicensed spectrum in sequence to obtain the D2D spectrum selection matrix Φ, calculate its throughput, and then output the Φ that maximizes the throughput under the constraints to solve problem 1.

[0075] Secondly, the main issue addressed in the unlicensed spectrum is the duty cycle problem in D2D. This invention primarily uses Lagrange duality combined with its KKT conditions to solve this problem. The mathematical model for the unlicensed spectrum, specifically the Lagrange function, can be expressed as:

[0076]

[0077] Where α represents the variable vector of D2D duty cycle on the unlicensed spectrum. Then, the Lagrange dual function can be obtained as:

[0078]

[0079] C1:R w ≥R wmin

[0080] C2:α≥0

[0081] Therefore, the objective function on the unauthorized spectrum can be rewritten as:

[0082]

[0083] Furthermore, the dual problem of the above equation can be expressed as:

[0084]

[0085] C1:R w ≥R wmin

[0086] C2:α≥0

[0087] Therefore, the problem can be expressed by the following formula:

[0088]

[0089] Thus, the optimal vector is obtained. Therefore:

[0090]

[0091] According to the above formula, the optimal D2D duty cycle is:

[0092]

[0093] Therefore, the optimal vector After solving the duty cycle problem, in order to obtain a given θ * The α of (α) can be obtained using the gradient method. Here, the dual vector α can be updated in each iteration, and finally the optimal dual vector α for solving the problem can be obtained. * The gradient of α can be derived from the following partial derivatives:

[0094]

[0095] Therefore, α can be updated using the following formula:

[0096]

[0097] Here, s( i) Defined as the step size of the i-th iteration, it can be expressed as:

[0098]

[0099] Therefore, by combining the Lagrange duality of D2D with its KKT conditions on the unlicensed spectrum, the optimal duty cycle allocation within a time slot T can be obtained, thereby solving problem 2.

[0100] Furthermore, in step S7, the Φ obtained in problem 1 is used to optimize the duty cycle. The question, and also the result of question 2 Conversely, optimize the D2D spectrum selection problem by repeating S6-S7 until the result converges.

[0101] The beneficial effects of this invention are as follows: by rationally offloading D2D users from licensed spectrum to unlicensed spectrum, the overall system throughput is greatly improved, while idle resources on unlicensed spectrum are rationally utilized, thereby alleviating the problem of congestion and shortage of licensed spectrum resources, and providing a reference for the harmonious coexistence of other networks.

[0102] Advantages and benefits of the present invention

[0103] With the rapid development of mobile terminals and wireless communication, global mobile traffic has exploded, leading to an extreme shortage of existing licensed spectrum resources. Given the short communication distance and low transmission power of D2D, it is advisable to consider deploying some D2D communication in unlicensed frequency bands with abundant spectrum resources to further alleviate spectrum demand pressure.

[0104] Most existing research focuses on maximizing system throughput by controlling D2D transmit power in licensed spectrum or enabling harmonious coexistence between D2D and WiFi by controlling D2D duty cycles in unlicensed spectrum. Few studies combine licensed and unlicensed spectrum for system optimization, and most researchers tend to focus on D2D power control while neglecting D2D information such as location and which D2D pairs are best suited for transmission in which mode. Therefore, this invention fully utilizes relevant D2D information, such as location and quantity, and rationally allocates spectrum for each D2D based on its interference to cellular users in licensed spectrum, thereby maximizing overall system throughput.

[0105] The design objective of this invention is to maximize the throughput of the entire system while satisfying all constraints. Attached Figure Description

[0106] To make the objectives, technical solutions, and beneficial effects of this invention clearer, the following figures are provided for illustration:

[0107] Figure 1 Algorithm flowchart

[0108] Figure 2 A network model diagram for system coexistence;

[0109] Figure 3 Unlicensed spectrum channel resource allocation diagram Detailed Implementation

[0110] The present invention will now be described in detail with reference to the accompanying drawings.

[0111] This invention primarily focuses on the resource allocation and management of licensed and unlicensed spectrum, proposing a location-information-assisted D2D communication resource allocation method. Referring to numerous D2D resource allocation schemes in various studies, this invention optimizes system throughput by making reasonable spectrum selection for D2D while fully utilizing the inherent information of D2D components.

[0112] like Figure 1 As shown, the algorithm includes the following steps:

[0113] S1: Assume there are M D2D pairs and N cellular users (CUs). Divide the licensed spectrum into N orthogonal sub-channels, and allow only one CU to use each orthogonal sub-channel. Calculate the interference I of each D2D pair in the licensed spectrum to each cellular user CU. Then an N*M interference matrix can be obtained.

[0114] S2: Using a matching algorithm, the matched D2D and CU are assigned to the same sub-channel according to the interference matrix. The matching principle here is to minimize the interference of D2D to CU.

[0115] S3: If N < M, then offload the excess D2D to the unlicensed spectrum, and also offload the D2D that causes significant interference to the CU to the unlicensed spectrum, coexisting with the WiFi using the Duty Cycle (DC) mechanism.

[0116] S4: The optimization problem is to maximize the total system throughput while satisfying various constraints. These constraints are generally to ensure normal WiFi transmission.

[0117] S5: Decompose the optimization problem into two sub-problems: 1. Which D2Ds and how many D2Ds are offloaded to the unlicensed spectrum? 2. The duty cycle problem of using the DC mechanism on the unlicensed spectrum;

[0118] S6: After using the matching algorithm, the D2D that causes significant interference to the CU is sequentially offloaded to the unlicensed spectrum, resulting in the D2D spectrum selection matrix Φ, which solves problem 1. Simultaneously, the Lagrange dual problem is used to solve the duty cycle problem of the DC mechanism, obtaining the optimal duty cycle.

[0119] S7: Use the obtained Φ to optimize the duty cycle. The problem, and at the same time the result Conversely, optimize the D2D spectrum selection problem by repeating S6-S7 until the result converges.

[0120] In a coexistence scenario, there are base stations, D2D users, cellular users (CU), and WiFi users. The network model is as follows: Figure 2As shown, here we assume there are N cellular users and M D2D pairs. Some D2D pairs coexist with cellular users on the licensed spectrum, dividing the channel into N orthogonal sub-channels. Each sub-channel allows at most one cellular user and one D2D pair to use it. The other D2D pairs coexist with WiFi users on the unlicensed spectrum using the DC mechanism.

[0121] Access frameworks for D2D users and WiFi users in coexisting scenarios, such as Figure 3 As shown, a time axis is divided into many time periods T. Within a time period T, the duty cycle is... This time period is allocated partly to D2D users and partly to WiFi users. Because WiFi user traffic is random and unpredictable, the duty cycle is updated using this algorithm after each time period.

[0122] The proposed algorithm mainly addresses two problems: the spectrum selection problem for each D2D operation and the duty cycle problem on unlicensed spectrum.

[0123] First, based on the unknown information regarding cellular users and D2D users, the interference matrix I can be expressed as:

[0124]

[0125] Each element represents the interference of the m-th D2D to the n-th CU, and its interference expression can be expressed as:

[0126] I n,m =P m G n,m

[0127] Where P m G represents the transmit power of the m-th D2D transmitter. n,m The channel gain from the m-th D2D transmitter to the n-th CU user is expressed as:

[0128]

[0129] Where |h n,m | represents the power gain from the m-th D2D transmitter to the n-th CU user, l n,m Let α represent the distance from the m-th D2D transmitter to the n-th CU user, and let α represent the path loss factor in the subchannel.

[0130] Then, based on the obtained interference matrix I, each CU user and its best matching D2D can be assigned to the same sub-channel according to the matching algorithm, that is, the D2D with the least interference to the CU user can be assigned to the same sub-channel.

[0131] The goal of the matching algorithm is to minimize interference to the cellular network and achieve optimal system performance by rationally allocating the licensed subchannel occupancy of D2D users. This invention assumes that the D2D user set, CU user set, and licensed subchannel set are independent and uncorrelated. m (m∈M) represents the D2D user, CU n (n∈N) are CU users, S n (n∈N) represents an unlicensed subchannel. It is assumed that exchanging information between users requires no additional signaling cost, and each user can obtain all state information from other users. The matching process assigns a suitable subchannel to each CU user, which is equivalent to generating a mapping relationship between two disjoint sets: the CU user set and the D2D user set. During the matching process, if the CU... n Matched D m , denoted as (CU n D m To ensure the basic traffic requirements of CU users, each CU user is matched with at least one licensed sub-channel, and the specific D2D pair matched by a CU user will be determined by the matching algorithm.

[0132] Furthermore, after matching is complete, if N < M, meaning the number of D2D users is greater than the number of CU users, since each sub-channel allows at most one CU and one D2D user to be used simultaneously, the excess D2D users after matching are all offloaded to the unlicensed spectrum to coexist with WiFi users. Simultaneously, D2D users and WiFi users coexist on the unlicensed spectrum using a DC mechanism. Therefore, how to reasonably allocate their respective duty cycles to maximize the total system throughput without affecting the normal transmission of WiFi users is a problem that needs to be solved.

[0133] The optimization problem of this invention can then be derived, maximizing the total system throughput while satisfying various constraints. Here, throughput includes both licensed spectrum throughput and unlicensed spectrum throughput. The throughput on the licensed spectrum comprises the throughput of CU users and the throughput of D2D communication on the selected licensed spectrum, while the throughput on the unlicensed spectrum comprises the throughput of D2D communication on the selected unlicensed spectrum and the throughput of WiFi users.

[0134] On the licensed spectrum: the throughput of CU users can be expressed as:

[0135]

[0136] One of the summation functions sums the number of CU users over a given number, and the other sums the number of sub-channels. Additionally, B... n P represents the sub-channel bandwidth selected by the CU user. n and P mThese are the transmit powers for CU users and D2D users, respectively. G n and G n,m These are the channel gain from the CU user to the base station and the channel gain from the CU user to the D2D transmitter, respectively. n (S n =0,1) represents the sub-channel selected by the CU user, if S n =1, then the CU user selects this sub-channel for communication. If S n If φ = 0, then the CU user will not select this sub-channel for communication. m (φ m =0,1) indicates whether the D2D user matched with this CU is communicating in this sub-channel. Similarly, if φ m =1, then the D2D user selects the CU user to communicate within the sub-channel, if φ m =0, then the D2D user selects unlicensed spectrum for communication. This represents the Gaussian noise of the channel.

[0137] Meanwhile, the throughput of D2D users on the licensed frequency band can be expressed as:

[0138]

[0139] Among them G m This represents the channel gain between the D2D transmitter and receiver. All other parameters are the same as described above. The D2D spectrum selection matrix Φ = [φ1, φ2...φ] can also be obtained. m ].

[0140] Therefore, the system throughput R on the licensed spectrum can be expressed as:

[0141] R = R C +R D

[0142] Furthermore, in unlicensed spectrum: D2D throughput It can be represented as:

[0143]

[0144] Where B represents the channel bandwidth of the unlicensed spectrum, I D This indicates interference from other D2D sources on unlicensed frequency bands, I D It can be represented as:

[0145]

[0146] Where φ i Indicates whether the D2D communication is on unlicensed spectrum, P i G represents the transmit power of the D2D transmitter. iLet represent the channel gain between the i-th D2D transmitter and the m-th D2D receiver.

[0147] Since WiFi user transmission uses the CSMA / CA protocol, the throughput of a WiFi system can be modeled as a discrete-time Markov chain mathematical model, with the steady-state probability being:

[0148]

[0149] CW min Here, n is the minimum competition window size, n is the maximum backoff stage, and p represents the conditional conflict probability, which can be expressed as:

[0150]

[0151] Assume p tr To determine the probability that at least one WiFi device is working in a given time slot, we obtain...

[0152]

[0153] Thus, the normalized throughput is obtained:

[0154]

[0155] Where E[T packet The average time interval (E[T]) represents the average time taken for a WiFi AP to successfully transmit data packets. slot It can be divided into three phases: the idle phase, the successful transmission phase, and the collision phase. Each phase is represented by P. i P s and P c The probability of each stage can be expressed as:

[0156]

[0157]

[0158] P c =P tr (1-p s )

[0159] Where N W P represents the number of WiFi networks, and the average value of the payload information successfully transmitted in a time slot is P. s P tr E[p], E[p] is a probability P r P tr The average payload for successful transmission; conversely, if a collision occurs in the considered time slot, the probability is: P tr (1-P sFinally, the WiFi throughput expression is obtained as follows:

[0160]

[0161] Where T S T is the average time that the channel is busy due to successful transmissions. c σ is the average time the channel is busy due to collisions. σ is the average time the channel is empty.

[0162] T S =H+E[P]+SIFS+δ+ACK+DIFS+δ

[0163] T C =H+E[P]+DIFS+δ

[0164] Where E[P] represents the average size of the packets that collide, H is the duration of the Media Access Control (MAC) and Physical Layer (PHY) headers of the WiFi network, δ represents the propagation delay, and DIFS, ACK, and SIFS represent the DCF inter-frame interval, transmission slot, and shorter inter-frame interval, respectively.

[0165] Therefore, the system throughput R on unlicensed spectrum U It can be represented as:

[0166]

[0167] Meanwhile, the total system throughput R total It can be represented as

[0168]

[0169] Furthermore, since D2D offloading to unlicensed spectrum will cause some interference to WiFi, in order to ensure the normal transmission of the WiFi system, the throughput of the WiFi system needs to be no less than its minimum throughput threshold, i.e.: R w ≥R wmin , where R wmin This represents the minimum throughput required for normal WiFi transmission.

[0170] Therefore, the optimization problem of this invention can be obtained:

[0171]

[0172] C1:I n,m ≤I max

[0173]

[0174] C3:R w ≥Rwmin

[0175] Where C1 indicates that the interference of D2D to CU users on the licensed spectrum cannot exceed its maximum interference threshold, C2 indicates the range of D2D duty cycle values ​​on the unlicensed spectrum, and C3 indicates that the throughput of WiFi on the unlicensed spectrum cannot be less than its minimum threshold.

[0176] However, since the objective function is quite complex, it is decoupled into two sub-problems for joint solution.

[0177] (1) Fixed To solve the Φ problem, the main issue is to determine which D2D operations are offloaded to unlicensed spectrum to maximize the total system throughput under constraints.

[0178] (2) Fix Φ to solve The main problem is how to allocate the duty cycle for D2D on unlicensed spectrum.

[0179] This maximizes the total system throughput under constraints.

[0180] Therefore, the main problem to solve is the two sub-problems decomposed above.

[0181] The interference matrix of each D2D to each CU user has been obtained. The optimal match between D2D and CU has been obtained according to the matching algorithm. Here, we only need to unload the D2Ds that have the greatest interference to CU after matching to the unlicensed spectrum in sequence to obtain the D2D spectrum selection matrix Φ, calculate its throughput, and then output the Φ that maximizes the throughput under the constraints to solve the problem (1).

[0182] Secondly, the main issue addressed in the unlicensed spectrum is the duty cycle problem in D2D. This invention primarily uses Lagrange duality combined with its KKT conditions to solve this problem. The mathematical model for the unlicensed spectrum, specifically the Lagrange function, can be expressed as:

[0183]

[0184] Where α represents the variable vector of D2D duty cycle on the unlicensed spectrum. Then, the Lagrange dual function can be obtained as:

[0185]

[0186] C1:R w ≥R wmin

[0187] C2:α≥0

[0188] Therefore, the objective function on the unauthorized spectrum can be rewritten as:

[0189]

[0190] Furthermore, the dual problem of the above equation can be expressed as:

[0191]

[0192] C1:R w ≥R wmin

[0193] C2:α≥0

[0194] Therefore, the problem can be expressed by the following formula:

[0195]

[0196] Thus, the optimal vector is obtained. Therefore:

[0197]

[0198] According to the above formula, the optimal D2D duty cycle is:

[0199]

[0200] Therefore, the optimal vector After solving the duty cycle problem, in order to obtain a given θ * The α of (α) can be obtained using the gradient method. Here, the dual vector α can be updated in each iteration, and finally the optimal dual vector α for solving the problem can be obtained. * The gradient of α can be derived from the following partial derivatives:

[0201]

[0202] Therefore, α can be updated using the following formula:

[0203]

[0204] Here, s (i) Defined as the step size of the i-th iteration, it can be expressed as:

[0205]

[0206] Therefore, by combining the Lagrange duality of D2D with its KKT conditions on the unlicensed spectrum, the optimal duty cycle allocation within a time slot T can be obtained, thereby solving the problem (2).

[0207] Finally, the Φ obtained from problem (1) is used to optimize the duty cycle. The problem, and at the same time, the result obtained from problem (2) Conversely, optimize the D2D spectrum selection problem by repeating S6-S7 until the result converges.

[0208] Finally, it should be noted that the above preferred embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail through the above preferred embodiments, those skilled in the art should understand that various changes can be made to it in form and detail without departing from the scope defined by the claims of the present invention.

Claims

1. A location-information-assisted D2D communication resource allocation method, characterized in that: The specific steps are as follows: S1: Assume there are M D2D pairs and N cellular users (CUs). Divide the licensed spectrum into N orthogonal sub-channels, and allow only one CU to use each orthogonal sub-channel. Calculate the interference of each D2D pair in the licensed spectrum on each cellular user CU. Then we can arrive at a conclusion. The interference matrix; S2: Using a matching algorithm, the matched D2D and CU are assigned to the same sub-channel according to the interference matrix. The matching principle here is to minimize the interference of D2D to CU. S3: After the matching is completed, if When the number of D2D users is greater than the number of CU users, since each sub-channel allows at most one CU and one D2D user to be used simultaneously, after matching, all the excess D2D users are offloaded to the unlicensed spectrum to coexist with WiFi users. S4: The optimization problem is to maximize the total system throughput while satisfying all constraints; S5: Decompose the optimization problem into two sub-problems: (1) which D2Ds to unload and how many D2Ds to unlicensed spectrum; (2) the duty cycle problem of using the DC mechanism on unlicensed spectrum; S6: Based on the interference matrix of each D2D to each CU user obtained previously, and the optimal match between D2D and CU determined by the matching algorithm, we only need to offload the D2Ds that cause the most interference to the CU after matching to the unlicensed spectrum to obtain the D2D spectrum selection matrix. This is used to solve problem 1, and the duty cycle problem of the DC mechanism is solved using the Lagrange dual problem to obtain the optimal duty cycle. ; S7: The result will be Used to optimize duty cycle The problem, and at the same time the result Conversely, optimize the D2D spectrum selection problem by repeating S6-S7 until the result converges.

2. The location-information-assisted D2D communication resource allocation method according to claim 1, characterized in that: In step S1, for the interference matrix It can be represented as: ； Each element represents the first The D2D pair of the first The interference of a CU can be expressed as: ； in Indicates the first The transmit power of a single D2D transmitter. Indicates the first The D2D transmitter to the first The channel gain for each CU user is expressed as: ； in Indicates the first The D2D transmitter to the first Power gain of each CU user Indicates the first The D2D transmitter to the first The distance between CU users This represents the path loss factor in the sub-channel.

3. The location-information-assisted D2D communication resource allocation method according to claim 1, characterized in that: In step S2, based on the obtained interference matrix The matching algorithm can be used to assign each CU user and its optimal D2D matching to the same subchannel, that is, to assign the D2D with the least interference to that CU user to the same subchannel. The goal of the matching algorithm is to minimize the interference to the cellular network and achieve optimal system performance by rationally allocating the licensed subchannel occupancy of D2D users. To represent D2D users, For CU users, This represents unlicensed sub-channels that are independent and unrelated to each other; it is assumed that there is no additional signaling cost for users to exchange information, and each user can obtain all the state information of other users; the matching process assigns a suitable sub-channel to each CU user, which is equivalent to generating a mapping relationship between two disjoint sets, the CU user set and the D2D user set; During the matching process, if Matched , recorded as To ensure the basic traffic needs of CU users, each CU user is matched with at least one authorized sub-channel, and the specific D2D pair matched by a CU user will be determined by the matching algorithm.

4. The location-information-assisted D2D communication resource allocation method according to claim 1, characterized in that: In step S4, the optimization problem can be derived to maximize the total system throughput while satisfying various constraints. Here, the throughput includes the throughput of the licensed spectrum and the throughput of the unlicensed spectrum. The throughput on the licensed spectrum is the throughput of CU users and the throughput of D2D communication on the selected licensed spectrum. The throughput on the unlicensed spectrum is the throughput of D2D communication on the selected unlicensed spectrum and the throughput of WiFi users. On the licensed spectrum: the throughput of CU users can be expressed as: ； One of the two summation functions sums the number of CU users, and the other sums the number of sub-channels. This indicates the sub-channel bandwidth selected by the CU user. and These are the transmit powers for CU users and D2D users, respectively. and These are the channel gain from the CU user to the base station and the channel gain from the CU user to the D2D transmitter, respectively. Indicates the sub-channel selected by the CU user, if If the CU user selects this sub-channel for communication, then... If so, the CU user will not select this sub-channel for communication. This indicates whether the D2D user matched with this CU is communicating in this sub-channel. Similarly, if... If the D2D user selects the CU user to communicate within the sub-channel, then... In this case, D2D users choose unlicensed spectrum for communication. This represents Gaussian noise in the channel; Meanwhile, the throughput of D2D users on the licensed frequency band can be expressed as: ； in This represents the channel gain between the D2D transmitter and receiver; all other parameters are the same as described above. The D2D spectrum selection matrix can also be obtained. ; Therefore, the system's throughput on the licensed spectrum It can be represented as: ； Furthermore, in unlicensed spectrum: D2D throughput It can be represented as: ； in Indicates the channel bandwidth of unlicensed spectrum. This indicates interference from other D2D sources on unlicensed frequency bands. It can be represented as: ； in This indicates whether the D2D communication is on unlicensed spectrum. This indicates the transmit power of the D2D transmitter. Indicates the first The D2D transmitter to the first Channel gain between D2D receivers; Since WiFi user transmission uses the CSMA / CA protocol, the throughput of a WiFi system can be modeled as a discrete-time Markov chain mathematical model, with the steady-state probability being: ； in It is the minimum contention window size. It is the maximum retreat phase. The probability of a condition conflict can be expressed as: ； Assumption To determine the probability that at least one WiFi device is active in the time slot under consideration, we obtain: ； Thus, the normalized throughput is obtained: ； in This indicates the average time taken for a WiFi AP to successfully transmit data packets; Average time slot It can be divided into three phases: the idle phase, the successful transmission phase, and the collision phase. Each phase is as follows: , and The probability of each stage can be expressed as: ； ； ； in This represents the number of WiFi networks, and the average value of payload information successfully transmitted within a time slot is... , For probability The average payload for successful transmission; conversely, the probability of a collision occurring in the considered time slot is: Finally, the expression for WiFi throughput is obtained as follows: ； in This is the average time that the channel becomes busy due to successful transmissions, and σ is the average time the channel is busy due to collisions, and σ is the average time the channel is empty. , ,in This represents the average size of the packets that collide. This refers to the duration of the Media Access Control (MAC) and Physical Layer (PHY) headers in a WiFi network. Indicates the propagation delay. , and express Interframe interval, transmission time slot, and shorter interframe interval; Therefore, system throughput on unlicensed spectrum It can be represented as: ； Similarly, the total system throughput It can be represented as: ； Since D2D offloading to unlicensed spectrum will cause some interference to WiFi, in order to ensure the normal transmission of the WiFi system, the throughput of the WiFi system needs to be no less than its minimum throughput threshold, that is: ,in This represents the minimum throughput required for normal WiFi transmission. Therefore, the optimization problem can be expressed as: ； ； in This means that D2D interference to CU users on the licensed spectrum cannot exceed its maximum interference threshold. This indicates the range of values ​​for the D2D duty cycle in the unlicensed spectrum. This indicates that the throughput of WiFi on unlicensed spectrum cannot be less than its minimum threshold.

5. A location-information-assisted D2D communication resource allocation method according to claim 4, characterized in that: In step S5, since the objective function is complex, it is decoupled into two sub-problems for joint solution: (1) Fixed to solve The main problem is to determine which D2D operations should be offloaded to unlicensed spectrum to maximize the total system throughput under constraints. (2) Fixed to solve The main problem is how to allocate the duty cycle of D2D on unlicensed spectrum to maximize the total system throughput under constraints. Furthermore, in step S6, the two sub-problems from step S5 are mainly addressed: Based on the previously derived D2D spectrum selection matrix Calculate its throughput, and then output the maximum throughput that satisfies the constraints. In order to solve the problem (1); Secondly, the main issue addressed in the unlicensed spectrum is the duty cycle problem in D2D, which is solved using Lagrange duality combined with its KKT conditions. Firstly, the mathematical model of the Lagrange function on the unauthorized spectrum can be expressed as: ； in The variable vector represents the D2D duty cycle on the unauthorized spectrum; then the Lagrange dual function can be obtained as: ； Therefore, the objective function on the unauthorized spectrum can be rewritten as: ； Furthermore, the dual problem of the above equation can be expressed as: ； Therefore, the problem can be expressed by the following formula: ； Thus, the optimal vector is obtained Therefore: ； According to the above formula, the optimal D2D duty cycle is: ； Therefore, the optimal vector ; After solving the duty cycle problem, in order to obtain a given of The gradient method can be used here, where the dual vector... Through each iteration, the optimal dual vector for solving the problem can be obtained. , The gradient can be derived from the following partial derivatives: ； Therefore, it can be updated using the following formula. : ； here, Defined as the first The step size of the next iteration can be expressed as: ； Therefore, by combining the Lagrangian duality of D2D with its KKT conditions on the unlicensed spectrum as described above, it is possible to obtain a time slot The optimal duty cycle allocation within the time limit is used to solve the problem (2).

6. The location-information-assisted D2D communication resource allocation method according to claim 1, characterized in that: The result obtained from problem (1) Used to optimize duty cycle The problem, and at the same time, the result obtained from problem (2) Conversely, optimize the D2D spectrum selection problem by repeating S6-S7 until the result converges.

Images