Switching parameter optimization method based on utility function

Abstract

The present application relates to a kind of switching parameter optimization method based on utility function, belong to communication technical field.The present application dynamically adjusts HCPs setting according to user state and network environment to adapt to different network environment and user demand, mainly includes the following steps: first, using the bounded function of SINR, cell load and user-to-base station distance to assess user state and network environment;Second, the output of three bounded functions is used as the input of automatic weight function to calculate the relative importance of each network factor on switching decision;Finally, the utility function value is calculated and the HCPs setting of user is optimized.In order to verify the effectiveness of the method, the proposed method is simulated, and compared with the switching parameter optimization method in other literatures.The performance difference between the proposed method and the existing method is evaluated from the aspects of handover probability, ping-pong handover probability and radio link failure probability.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of communication, and relates to a switching parameter optimization method based on a utility function. BACKGROUND

[0002] In a user-dense distribution scenario, improper setting of switching control parameters affects the stability of cell switching. Using fixed switching control parameters or setting switching parameters for all users in a cell at the same time will lead to reduced communication quality of part of the users, which is mainly caused by different positions and states of the users. Traditional switching parameter optimization dynamically adjusts switching control parameters according to a single index, which leads to low effectiveness of the method. Distance, channel conditions, noise, interference, resource availability and terminal speed and other factors will affect the stability of user communication connection, so selecting appropriate switching indexes is the key to improving the switching control parameter optimization method. SUMMARY

[0003] Therefore, the purpose of the application is to provide a switching parameter optimization method based on a utility function.

[0004] To achieve the above purpose, the application provides the following technical scheme.

[0005] The switching parameter optimization method based on the utility function comprises the following steps.

[0006] S1: Defining a problem

[0007] Supposing that a switching trigger event is an A3 event, according to the definition method of the third generation partnership project (3GPP) organization, the definition method is shown in the following formula:

[0008] Mn+Ofn+Ocn-Hys>Mp+Ofp+Ocp+Off (1)

[0009] Wherein, Mn and Mp are measurement results of adjacent cells and a service cell respectively, and the measurement objects are reference signal received power or quality; Ofn and Ofp are specific bias parameters of a frequency band in which a cell system bandwidth is located; Ocn and Ocp are cell individual bias parameters of adjacent cells and the service cell respectively, and are used for load balancing; Off is an A3 event bias value of the service cell and the adjacent cell; and Hys is an A3 event hysteresis parameter, that is, a switching threshold in a switching trigger condition.

[0010] The switching trigger event is simplified, and the mathematical expression after the simplification is shown in the following formula:

[0011] R T >R S +HOM A (2)

[0012] Wherein, RS and R T denote the RSRP of serving cell and target cell, respectively; HOM A is a dynamically adjusted handover threshold, whose purpose is to reduce handover failure and ping-pong handover when the UE continuously switches between two cells; cell handover is triggered when the RSRP of target cell satisfies the formula within the triggering time;

[0013] The cell handover control parameter optimization problem is expressed by the formula:

[0014]

[0015]

[0016] C2: T min ≤ TTT A ≤ T max (3-3b)

[0017] C3: HOM min ≤ HOM A ≤ HOM max (3-3c)

[0018]

[0019] wherein: P denotes the handover probability, ping-pong handover probability and radio link failure probability, which is optimized by selecting appropriate TTT A and HOM A ; TTT A and HOM A denote the dynamically adjusted triggering time and handover threshold, respectively; constraint (3-3a) ensures that each UE is associated with only one cell, b ij is the user association indicator; constraints (3-3b) and (3-3c) ensure that the selected TTT A and HOM A do not exceed the range, whose range is 0-5.12s and 0-10dB, respectively; and (3-3d) is a binary constraint on the user association indicator, b ij =1 indicates that the corresponding user is associated with the cell; the UE periodically measures the RSRP of all serving cells and reports the measurement results;

[0020] S2: Establish system model

[0021] Consider a two-tier HetNet consisting of N m macro cells and N s micro cells; the set of macro cells is represented by M = {M1, M2,..., M Nm} and the set of micro cells is represented by S = {S1, S2,..., SNs} represents; define the set of all cells as E = M U S, use e i represents the cells in E, N all = N m + N s is the total number of cells; 3 micro base stations are deployed within the coverage of a macro base station, users are randomly generated in each macro cell and micro cell, and define the UE set as U = {u1, u2,..., u K}, where K is the total number of UEs in all cells;

[0022] S3: Dynamic parameter optimization

[0023] The input includes the switching index and the generation of the bounded function, the calculation of the weight corresponding to the bounded function, the calculation of the utility function value, and the output of the switching control parameter; the state of the UE and the network environment are evaluated through the utility function, the higher the utility function value, the better the communication environment of the UE, and vice versa; when the utility function value rises, the necessity of cell switching decreases, and the switching control parameter is adjusted upward to avoid ping-pong switching; when the utility function value decreases, the necessity of cell switching is higher, and the switching control parameter needs to be adjusted downward to avoid radio link failure;

[0024] S31: Establish a bounded utility function

[0025] The bounded functions of SINR, cell load, and the distance between the UE and the base station are represented by the SINR function f(γ), the load function f(L), and the distance function f(d) of the UE, and their weights are represented by ω γ , ω L , and ω d ; the representation method of the utility function is as shown in the formula:

[0026] f WF (γ, L, d) = ω γ f(γ) + ω L f(L) + ω d f(d) (4)

[0027] In the formula, f WF (γ, L, d) is divided into two parts; the first part is the bounded sub-function f(γ), f(L), and f(d); the second part is the weight ω γ , ω L , ω d of each sub-function;

[0028] S32: Calculate the weight of the bounded function

[0029] A weight value ω nand their sum equals 1; the weight function uses the formula shown below:

[0030]

[0031] wherein ω n represents the weight of the function f(x n ), ω n is ω γ , ω L or ω d ; f(x n ) is the corresponding nth bounded function, whose weight needs to be calculated; F represents the total number of indicators used to adjust the value of HCPs, considering three indicators γ, L and d, F is set to 3; f(x i ) is the ith bounded function, i = 1,..., F;

[0032] The calculation method of the SINR bounded function weight is shown in the formula:

[0033]

[0034] In the formula, ω γ represents the weight of the SINR bounded function; f(γ) is the bounded function value; F is the number of all bounded functions, which is 3; the calculation methods of ω L and ω d are similar to the formula;

[0035] S33: Dynamic estimation of handover control parameters

[0036] After obtaining the bounded functions of different handover indicators and their corresponding weights, and calculating the utility function value, the output of the utility function will be used for dynamic estimation of the HCPs setting of each user; the following is the specific adjustment of the handover threshold and trigger time:

[0037] (1) Handover threshold adjustment

[0038] The dynamically adjusted handover threshold value HOM A is represented by the formula:

[0039] HOM A = H + ΔM (13)

[0040] Wherein H is the handover threshold value at the previous time, and its initial value H0 is defined as the average of the maximum handover threshold and the minimum handover threshold; ΔM is the dynamically adjusted part; the calculation method of H0 is shown in the formula:

[0041]

[0042] In the formula, M max and M minHmax and Hmin are the maximum and minimum values of the handover threshold; suppose they are 10dB and 0dB respectively;

[0043] The dynamic adjustment of the handover threshold setting is shown in the formula:

[0044]

[0045] The calculation of the initial value of the handover threshold H0 is shown in the formula; the utility function f WF (γ, L, d) is shown in the formula; when the target cell SINR is less than the threshold and the serving cell SINR is greater than the threshold, f WF (γ, L, d) is used to replace f WF (γ, L, d) in the formula; when the serving cell SINR is less than the threshold and the target cell SINR is greater than the threshold, f γ (γ, L, d) is used to replace f WF (γ, L, d) in the formula; when the serving cell SINR is less than the threshold and the target cell SINR is greater than the threshold, f γ (γ, L, d) is used to replace f

[0046] (2) Trigger time adjustment

[0047] The range of the trigger time interval has been specified by the 3GPP organization, and its setting range is 0-5.12 seconds;

[0048] The adjustment of the trigger time is shown in the formula:

[0049]

[0050] Wherein: AT is the trigger time of the instant adjustment; T is the trigger time threshold level initially defined as a fixed appropriate interval, and then T will be equal to AT in the next adjustment; p is the reference adjustment value of the specified trigger time; Q is the trigger threshold value of the utility function change value, when the utility function change value is higher than Q, the adjustment of the handover parameter is triggered; Δf WF is the change value of the utility function f WF (γ, L, d) relative to the last time;

[0051] The adaptive trigger time value TTT A is represented by the formula:

[0052]

[0053] In the formula, the constants p and Q are used to adjust the sensitivity when updating the trigger time, and the values of p and Q are set to 0.04s and 0.1 respectively; T max is set to 5.12s, and T minare set to 0, which are the maximum and minimum of the trigger time specified by 3GPP respectively; the initial value of the trigger time of each handover parameter self-optimization method is assumed to be 100 ms.

[0054] Optionally, in the S31, the bounded function corresponding to different handover indexes is specifically described as follows:

[0055] (1) Bounded SINR function f(γ)

[0056] In the cell handover process, the trigger method based on received signal strength is the most practical, that is, the difference between the RSS of the serving cell and the selected target cell and the specific handover threshold are used for decision-making; the RSS is measured at the user side and then reported to the serving cell, and finally the serving cell decides whether to trigger handover; the best parameter representing RSS is SINR because noise and interference are considered at the same time; therefore, the bounded SINR function is used as the input of the utility function; the user u k from the cell e i receives the downlink signal-to-noise ratio The calculation method is shown in the formula:

[0057]

[0058] Wherein: is the downlink RSRP received by the user u k from the cell e i ; is the sum of the downlink power of all interference cells except the serving cell e i ; is the channel gain between the user u k and the cell e i , considering the path loss and fading effect; δ 2 is the noise power, that is, the measured value of the total noise in a given channel bandwidth;

[0059] The difference between the SINRs of the serving cell and the target cell is represented by the bounded SINR function; when the bounded SINR function is used in the utility function, the output needs to be between [-1, 1]; in order to ensure that the output value of the bounded SINR function is between [-1, 1], the difference of the SINRs is divided by the maximum SINR level; the bounded SINR function f(γ) in the utility function is represented by the formula:

[0060]

[0061] Wherein: γ S and γ T represent the SINRs of the user in the serving cell and the target cell respectively, and γ maxdenotes the maximum SINR received at the UE side; the acceptable SINR range is -10~30dB, and the maximum SINR value is assumed to be 30dB, i.e. γ max = 30;

[0062] (2) Bounded load function f(L)

[0063] The bounded load function is taken as the input of the utility function; the load of the cell e i The representation method of the load of the cell e

[0064]

[0065] In the formula, B W denotes the system bandwidth, β i is the ratio of the resources allocated to all active UEs of the cell e i to the total resources of the cell, and is defined as:

[0066]

[0067] Wherein: is the resource allocated to the user u i by the cell e k ; is the total resource allocated to all active users in the cell;

[0068] The loads of the serving cell and the target cell are evaluated by the load boundary function; the bounded input function in the utility function value needs to have a unified boundary, so the load difference of the cell is divided by the maximum load capacity of the cell, so that the output of f(L) is limited between [-1, 1]; the bounded load function f(L) in the utility function is represented by the formula:

[0069]

[0070] Wherein: L T and L S denote the loads of the target cell and the serving cell respectively; L max is the maximum load capacity of the cell;

[0071] (3) Bounded distance function f(d)

[0072] ​​When the UE is close to the cell edge, the handover condition is more likely to be triggered, but when the serving cell signal quality is good enough, the probability of ping-pong handover is increased; by considering the bounded distance function to adjust the utility function, when the UE is close to the cell edge, the value of the bounded distance function will increase, so that the value of the utility function increases and triggers the up-regulation of the handover control parameter, to a certain extent, to alleviate the ping-pong effect; when the user is at the edge of the cell, the weight of the bounded distance function will be lower, at this time the SINR and the load have a greater influence on the value of the utility function; the output range of the bounded distance function f(d) is [-1, 1], and its expression is as shown in the formula:

[0073]

[0074] Wherein: d max represents the maximum acceptable distance between the UE and the base station, assuming that d max is the cell radius.

[0075] The beneficial effects of the present application are that the present application proposes a parameter optimization method based on the utility function in the face of the challenges of cell handover. The method takes SINR, cell load and the distance between UE and base station as the input of the utility function, monitors the change of the utility function and dynamically adjusts the handover control parameter.

[0076] Other advantages, objects and features of the present application will be in part apparent and in part pointed out hereinafter in the specification, and will be observed by those skilled in the art upon examination of the specification, or can be learned by practice of the present application. The objects and other advantages of the present application can be realized and attained by the following description. BRIEF DESCRIPTION OF DRAWINGS

[0077] In order to make the objects, technical solutions and advantages of the present application clearer, the preferred detailed description of the present application will be combined with the drawings, in which:

[0078] Figure 1 The handover control parameter is;

[0079] Figure 2 The network deployment diagram is;

[0080] Figure 3 The dynamic parameter optimization method flow is;

[0081] Figure 4 The simulation flowchart is;

[0082] Figure 5 The influence of the moving speed on the average handover probability of the user is;

[0083] Figure 6 The average ping-pong handover probability of different optimization methods is;

[0084] Figure 7 The influence of the moving speed on the average radio link failure probability. DETAILED DESCRIPTION

[0085] The present application is described and explained with additional specificity and detail through the use of the accompanying drawings in which:

[0086] The accompanying drawings, which are included to provide a further understanding of the application and are incorporated in and constitute a part of this specification, illustrate embodiments of the application and together with the description serve to explain the principles of the application. In the drawings:

[0087] The same or similar components in the drawings of the embodiments of the present application correspond to the same or similar components; in the description of the present application, it should be understood that if the terms "upper", "lower", "left", "right", "front", "back" and the like indicate the orientation or positional relationship based on the orientation or positional relationship shown in the drawings, they are only for the convenience of describing the present application and simplifying the description, and do not indicate or imply that the devices or elements referred to must have a particular orientation, be constructed and operated in a particular orientation, therefore the terms describing the positional relationship in the drawings are only for illustrative purposes, and cannot be understood as a limitation of the present application, for ordinary skilled in the art can understand the specific meaning of the above terms according to the specific circumstances.

[0088] 1. Definition problem

[0089] The handover trigger event considered by the present application is A3 event, according to the definition method shown in the formula as specified by the 3rd Generation Partnership Project (3GPP) organization:

[0090] Mn+Ofn+Ocn-Hys>Mp+Ofp+Ocp+Off (1)

[0091] Where: Mn and Mp are the measurement results of neighboring cells and serving cells, respectively, and the measurement object is the reference signal received power or quality; Ofn and Ofp are specific offset parameters of the frequency band where the cell system bandwidth is located; Ocn and Ocp are the cell-specific offset parameters of neighboring cells and serving cells, respectively, used for load balancing; Off is the A3 event offset value of serving cell and neighboring cells; Hys is the A3 event hysteresis parameter, which is the handover threshold in the handover triggering condition.

[0092] The simplified mathematical representation of the above switching trigger event is shown in the formula:

[0093] R T >R S +HOM A (2)

[0094] Where: R S and R T Representing the RSRP of the serving cell and the target cell respectively; HOM A It is a dynamically adjusted handover threshold, the purpose of which is to reduce handover failures and ping-pong handovers when the UE is continuously handovering between two cells; when the target cell RSRP is met within the trigger time, cell handover is triggered.

[0095] like Figure 1 As shown, when a UE moves from one cell to another, the RSRP of the serving cell decreases, while the RSRP of the target cell it is approaching increases. Cell handover is triggered when the difference in RSRP between the target and serving cells exceeds the handover threshold and lasts for more than the trigger time. The handover threshold and trigger time are the two important cell handover parameters used in this paper, and their configuration determines the impact of the handover method on network performance. The handover threshold and trigger time can be fixed or dynamic values, with units of dB and ms, respectively. Fixed values ​​indicate that the HCPs value remains constant during UE movement, while dynamic values ​​indicate that the HCPs value is automatically adjusted based on different inputs.

[0096] The problem of optimizing cell handover control parameters can be expressed by a formula:

[0097]

[0098]

[0099] C2:T min ≤TTT A ≤T max (3-3b)

[0100] C3:HOM min ≤HOM A ≤HOMmax (3-3c)

[0101]

[0102] Where: P represents the handover probability, ping-pong handover probability, and radio link failure probability, which are determined by selecting an appropriate TTT. A and HOM A To optimize, TTT A and HOM A These represent the trigger time and handover threshold for dynamic adjustment, respectively; constraint (3-3a) ensures that each UE is associated with only one cell, b ij It is a user association indicator; constraints (3-3b) and (3-3c) ensure that the selected TTT A and HOM A Not exceeding the range, which is 0–5.12s and 0–10dB respectively; and (3–3d) is the binary constraint on the user-associated indicator, b ij A value of 1 indicates that the corresponding user is associated with a cell; the UE periodically measures the RSRP of all serving cells and reports the measurement results.

[0103] 2. System Model

[0104] Consider two layers of HetNets, consisting of N m One macro cell and N s It consists of microcells. The set of macrocells is denoted by M = {M1, M2, ..., M}. Nm The set of microregions is denoted by S = {S1, S2, ..., S}. Ns} represents the set of all cells. The set of all cells is defined as E = M∪S, using e i Indicates E as the middle cell, N all =N m +N s This represents the total number of cells. The deployment methods for macrocells and microcells are as follows: Figure 2 As shown. Three micro base stations are deployed within the coverage area of ​​a macro base station. Users are randomly generated within each macro cell and micro cell. The UE set is defined as U = {u1, u2, ..., u...} K}, where K is the total number of UEs in all cells.

[0105] The UE receives its requested services via micro or macro base stations. When the UE moves from the serving cell to a neighboring cell, a handover is performed. Communication between cells during the handover process uses the X2 interface. The X2 interface essentially supports this process, allowing the exchange of operation reports, parameter configurations, and RLF status. After the UE periodically sends measurement reports to the serving cell, the serving cell decides to hand over the UE to the target cell. At each micro and macro base station, a distributed ad hoc network collects handover information to optimize HCPs (Handover Processing Centers).

[0106] 3. Dynamic parameter optimization method and process

[0107] The process of the dynamic parameter switching optimization method based on utility function is as follows: Figure 3 As shown. The main process includes inputting handover indicators and generating bounded functions, calculating the weights corresponding to the bounded functions, calculating the utility function value, and outputting handover control parameters. The utility function evaluates the UE's state and network environment; a higher utility function value indicates a better communication environment for the UE, and vice versa. When the utility function value increases, it indicates a lower necessity for cell handover, and the handover control parameters are adjusted upwards to avoid ping-pong handover; when the utility function value decreases, it indicates a higher necessity for cell handover, and the handover control parameters need to be adjusted downwards to avoid radio link failure.

[0108] 3.1 Bounded utility function

[0109] A utility function-based handover parameter optimization method is used to dynamically estimate the appropriate HCPs (Host Computational Values) for each UE. To avoid the influence of a single or inappropriate metric on the method's effectiveness, the utility function in this method consists of bounded functions related to SINR, cell load, and the distance between the UE and the base station. These three bounded functions of handover metrics can be represented by the SINR function f(γ), the load function f(L), and the UE distance function f(d), respectively, and their weights are denoted by ω. γ ω L and ω d The utility function is represented as shown in the formula:

[0110] f WF (γ,L,d)=ω γ f(γ)+ω L f(L)+ω d f(d) (4)

[0111] In the formula, f WF (γ,L,d) can be divided into two parts. The first part is the bounded sub-functions f(γ), f(L), and f(d); the second part is the weight ω of each sub-function. γ ω L ω d .

[0112] The bounded function corresponding to different switching indicators is specifically described as follows:

[0113] (1) Bounded SINR function f(γ)

[0114] In the cell switching process, the triggering method based on received signal strength is the most practical, that is, the difference between the RSS of the serving cell and the selected target cell and the specific switching threshold are used for decision-making. The user measures the RSS and then reports it to the serving cell, and finally the serving cell decides whether to trigger the switching. The best parameter representing the RSS is SINR, because it takes both noise and interference into account. Therefore, the bounded SINR function is used as the input of the utility function. The user u k receives the downlink signal-to-noise ratio from the cell e i . The calculation method is shown in the formula:

[0115]

[0116] wherein: is the downlink RSRP received by the user u k from the cell e i ; is the sum of the downlink power of all interference cells except the serving cell e i ; is the channel gain between the user u k and the cell e i , considering the path loss and fading effect; δ 2 is the noise power, that is, the measurement value of the total noise in a given channel bandwidth.

[0117] The difference between the SINRs of the serving cell and the target cell is represented by the bounded SINR function. When the bounded SINR function is used in the utility function, its output needs to be between [-1, 1]. To ensure that the output value of the bounded SINR function is between [-1, 1], the difference of the SINRs is divided by the maximum SINR level. Therefore, the bounded SINR function f(γ) in the utility function can be represented by the formula:

[0118]

[0119] wherein: γ S and γ T represent the SINRs of the user in the serving cell and the target cell, respectively, and γ max represents the maximum SINR received at the UE end. The acceptable SINR range is -10-30 dB, and it is assumed that the maximum SINR value is 30 dB, that is, γ max = 30.

[0120] (2) Bounded load function f(L)

[0121] There are two main reasons for inputting the cell load as the utility function. First, after making the handover decision and sending the handover request to the target cell, the target cell will decide to accept or reject the handover request according to the availability of its resources; second, balancing the load between cells is one of the reasons for handover, i.e. some users near the cell boundary can switch the connection to the adjacent cell to balance the load between cells. Therefore, the bounded load function is input as the utility function. The load of cell e i is represented as follows:

[0122]

[0123] In the formula, B W represents the system bandwidth, β i is the ratio of the resources allocated to all active UEs to the total resources of cell e i , which can be defined as:

[0124]

[0125] where: is the resource allocated to user u i by cell e k ; and is the total resource allocated to all active users in the cell.

[0126] The load of the serving cell and the target cell is evaluated by the load boundary function. The bounded input function in the utility function value needs to have a unified boundary, so the difference between the cell loads is divided by the maximum load capacity of the cell, so that the output of f(L) is limited between [-1, 1]. The bounded load function f(L) in the utility function is represented by the formula:

[0127]

[0128] where: L T and L S represent the load of the target cell and the serving cell, respectively; and L max is the maximum load capacity of the cell.

[0129] (3) Bounded distance function f(d)

[0130] ​​When the UE is close to the cell edge, the handover condition is more likely to be triggered, but when the serving cell signal quality is good enough, it will cause the probability of ping-pong handover to increase. By considering the bounded distance function to adjust the utility function, when the UE is close to the cell edge, the value of the bounded distance function will increase, so as to increase the value of the utility function and trigger the up-regulation of the handover control parameter, to a certain extent, to alleviate the ping-pong effect. When the user is at the edge of the cell, the weight of the bounded distance function will be lower, at this time the SINR and the load have a greater impact on the value of the utility function. The output range of the bounded distance function f(d) should be [-1, 1], and its expression is as shown in the formula:

[0131]

[0132] wherein d max represents the maximum acceptable distance between the UE and the base station, assuming d max is the cell radius.

[0133] 3.2 Bounded function weight calculation

[0134] The weight function can adaptively estimate the weight of different bounded functions according to the user's handover index input, thereby avoiding the uncertainty brought by static weight setting. Through the weight function, a weight value ω n is generated for each bounded function, and the sum of them is equal to 1. Therefore, the weight function uses the formula shown:

[0135]

[0136] wherein ω n represents the weight of the function f(x n ), ω n may be ω γ , ω L or ω d ; f(x n ) is the corresponding nth bounded function, and its weight needs to be calculated; F represents the total number of indicators used to adjust the value of HCPs, and three indicators γ, L and d are considered in this method, and F is set to 3; f(x i ) is the ith bounded function, i = 1,..., F.

[0137] The calculation method of the SINR bounded function weight is as shown in the formula:

[0138]

[0139] In the formula, ω γ represents the weight of the SINR bounded function; f(γ) is the value of the bounded function; F is the number of all bounded functions, and takes the value of 3. ω L and ω dThe calculation method and formula are similar.

[0140] 3.3 Dynamic estimation of handover control parameters

[0141] Through the above steps, the bounded function of different handover indicators and its corresponding weight are obtained, and the utility function value is calculated. The output of the utility function will be used to dynamically estimate the HCPs setting of each user. The following is the specific adjustment of the handover threshold and trigger time:

[0142] (1) Handover threshold adjustment

[0143] The handover threshold is one of the important parameters for controlling the handover trigger time. Because low handover threshold and high handover threshold settings can lead to high ping-pong handover probability and high RLF probability, which will affect the user experience. With the development of 5G ultra-dense networks, due to the dense deployment of small base stations, the adjustment of the handover threshold becomes particularly important. In addition, adjusting the handover threshold for all users in the cell will also cause handover problems. Therefore, the bounded function of different handover indicators of the UE and its weight need to be considered comprehensively to dynamically adjust the handover threshold.

[0144] The dynamically adjusted handover threshold value HOM A Using the formula:

[0145] HOM A = H + AM (13)

[0146] Where: H is the handover threshold value at the previous time, and its initial value H0 is defined as the average of the maximum handover threshold and the minimum handover threshold; AM is the dynamically adjusted part; The calculation method of H0 is shown in the formula:

[0147]

[0148] In the formula, M max and M min are the maximum and minimum values of the handover threshold. Assume that these values are 10 dB and 0 dB, respectively.

[0149] The amplitude of the dynamically adjusted part AM should not be too large, because rapid or large changes in the handover threshold setting can lead to handover too early or handover too late, depending on the direction of adjustment. Therefore, in order to avoid this problem, the adjustment amplitude needs to be controlled by the utility function considering the signal-to-noise ratio of the serving cell and the target cell, and the initial value of the handover threshold is taken as the reference value, and the dynamic adjustment is made around this reference value. The dynamically adjusted part of the handover threshold setting is shown in the formula:

[0150]

[0151] Where: the calculation method of the initial value of the handover threshold H0 is shown in the formula; the utility function fWF (γ, L, d) is calculated as shown in the formula; when the target cell SINR is less than the threshold and the serving cell SINR is greater than the threshold, f WF (γ, L, d) controls the amplitude of H0, which produces a very low handover threshold setting, resulting in early handover, handover to the wrong cell and ping-pong handover, where f WF (γ, L, d) ω γ f(γ); when the serving cell SINR is less than the threshold and the target cell SINR is greater than the threshold, f WF (γ, L, d) controls the amplitude of H0, which produces a very high handover threshold setting, which can result in late handover, thus increasing the probability of RLF, where f WF (γ, L, d) ω γ f(γ).

[0152] (2) Trigger time adjustment

[0153] The range of the trigger time interval has been specified by the 3GPP organization, and its setting range is 0-5.12 seconds. When the trigger time setting value is large, the handover condition needs to last for a longer time to trigger cell handover, and a too high trigger time value is easy to cause radio link failure; when the trigger time value is small, the time required for the handover condition to last is shortened, and the UE is more likely to occur cell handover, which will cause ping-pong handover when the wireless environment changes quickly. Adjusting the trigger time for all users in the entire cell will cause some users to have handover problems because their environments are different. Some users at the edge of the cell have poor signal quality, while some users can have good signal quality. Therefore, a more reasonable solution is to adjust the trigger time for each user based on the user state and the surrounding network environment.

[0154] Based on the change of the utility function value, the trigger time is adjusted up and down by increasing or decreasing a specific interval p. In order to avoid unnecessary trigger time adjustment, the trigger threshold of the utility function change is defined as Q. When the utility function value increases by more than the threshold Q, the trigger time value is adjusted up; otherwise, when the utility function value decreases by more than the threshold Q, the trigger time value is adjusted down. The instantaneous utility function value is usually compared with the recorded value at the previous time, that is, when the absolute value of the difference between the instantaneous utility function value and the utility function at the previous time is greater than Q, the system can be adjusted. The adjustment method of the trigger time is shown in the formula:

[0155]

[0156] Where: AT is the instant adjustment trigger time; T is the trigger time threshold level initially defined as a fixed appropriate interval, then T will equal AT in the next adjustment; p is the reference adjustment value of the trigger time; Q is the trigger threshold of the utility function change value, when the utility function change value is higher than Q, the adjustment of the handover parameter is triggered; Af WF is the change value of the utility function f WF (γ, L, d) relative to the last time.

[0157] The equation cannot be achieved in two cases. The first case is when the update value is less than 0, the trigger time value cannot be negative, so it cannot be lowered, at this time the trigger time should be set to zero; the second case is when the update value is greater than T max , the trigger time cannot exceed the maximum value, so it cannot be raised, at this time the trigger time should be set to T max . The adaptive value of the trigger time can be dynamically adjusted according to the change of the weight function in the formula. Accordingly, the adaptive trigger time value TTT A can be represented by the formula:

[0158]

[0159] In the formula, the constants p and Q are used to adjust the sensitivity when updating the trigger time. If they take too low values, they will bring higher computational complexity and latency to the system; therefore, in order to avoid too much complexity in calculation, the values of p and Q are set to 0.04s and 0.1 respectively in the entire simulation process; T max is set to 5.12s, and T min is set to 0, which are respectively the maximum and minimum values of the trigger time specified by 3GPP. It is assumed that the initial value of the trigger time of each handover parameter self-optimization method is 100ms.

[0160] 4. Simulation results and performance analysis of the method

[0161] 4.1 Simulation parameter setting

[0162] The dynamic parameter optimization method proposed in this chapter is simulated and verified, and the specific simulation process is shown in Figure 4 . When making handover decisions in this chapter, the traditional method is used.

[0163] The simulation uses HetNets composed of 19 macro cells and 57 micro cells. UEs receive traffic demand through macro cell or micro cell frequency bands. In the initial stage, the UE starts from a random position and moves at a constant speed along a straight line in a random direction. The simulation parameters are referred to in the literature, and the specific simulation parameters are shown in Table 1:

[0164] Table 1 Simulation parameter table of dynamic parameter optimization method

[0165]

[0166] For user u k to cell e i , the path loss model of different frequencies in urban HetNets is:

[0167]

[0168] where r0represents the reference distance between user u k and cell e i ; r represents the distance between user u k and cell e i , and r0is used instead of r in the calculation process when r n represents the path loss exponent; λ is the wavelength at different carrier frequencies, which is calculated by the carrier frequency of the macro cell or micro cell.

[0169] 4.2 Handover performance indicators

[0170] This section evaluates the performance of the dynamic parameter optimization method, considering the handover probability, ping-pong handover probability and radio link failure probability of users. The definitions of these indicators are introduced as follows:

[0171] (1) Handover probability

[0172] The handover probability is the likelihood of a user moving from one cell to another, which is represented as follows:

[0173]

[0174] where R T represents the downlink RSRP of the target cell; R S represents the downlink RSRP of the serving cell; HOM A is the dynamically adjusted handover threshold.

[0175] The average value of the handover probability of all UEs in the cell is calculated as follows:

[0176]

[0177] where K is the total number of users in the simulation; U is the set of UEs.

[0178] (2) Ping-pong handover probability

[0179] The ping-pong handover probability is the probability of repeatedly performing handover operations between the serving cell and the adjacent cell in a short period of time during user movement. In order to count the number of ping-pong handovers, a ping-pong handover time interval threshold Tc , assuming T c = 2s. When the UE switches from the serving cell to the target cell and then switches back to the serving cell within T c , it is considered that a ping-pong handover occurs. The representation of the ping-pong handover probability is shown in the formula:

[0180] P(HOPP) = P r [T i ≤ T c ] (21)

[0181] In the formula, T c is the ping-pong handover time interval threshold; T i is the time spent by the UE switching from the target cell to the serving cell, and its representation is shown in the formula:

[0182] T i = T L - T hb (22)

[0183] Wherein: T L is the time spent by the UE switching from the serving cell to the target cell; T hb is the time spent by the UE reconnecting to the same cell; when T i ≤ T c , it is considered that a ping-pong handover occurs, that is, the user reconnects to the serving cell again in a short time.

[0184] The calculation method of the average ping-pong handover probability is shown in the formula:

[0185]

[0186] In the formula, N HOPP is the number of ping-pong handovers occurring in the total simulation time, and its calculation method is shown in the formula:

[0187] N RHO = N S + N F (24)

[0188] Wherein: N RHO is the total number of handover requests; N F and N S are the numbers of handover failures and handover successes, respectively.

[0189] (3) Radio link failure probability

[0190] There are two main reasons for radio link failure: handover failure and communication link interruption. Handover failure refers to the failure of a UE to establish a connection when switching to a new cell, while communication link interruption refers to the loss of connection when a UE is communicating with the current serving cell. The probability of radio link failure is represented as shown in the formula:

[0191]

[0192] wherein, is the SINR of the user; γ thr is the SINR threshold.

[0193] The average radio link failure probability is calculated as shown in the formula:

[0194]

[0195] wherein: K is the total number of users in the simulation; U is the set of UEs.

[0196] 4.3 Analysis of simulation results

[0197] The stability of the method proposed in this paper is evaluated and analyzed by comparing it with the Handover Failure Type (HFT) based handover parameter optimization method and the FLC based handover parameter optimization method. In the HFT based method, the handover problem is monitored and the cell handover failure is classified, and the handover threshold and trigger time are adjusted according to the handover failure type. The FLC based method takes the user moving speed, reference signal received power and reference signal received quality as the input of the fuzzy logic system, and obtains the adjustment value of the handover threshold by processing the input through the fuzzy logic system. This method only considers the size of the moving speed when considering the moving speed, ignoring the direction of the moving speed.

[0198] As the moving speed increases, more and more UEs will move across the cells, resulting in an increase in the handover probability of the user. In the case of good user signal quality, if the handover threshold and trigger time are set to low values, unnecessary handover of the UE will occur. The average handover probability of the UE at different moving speeds is as shown in the formula: Figure 5As shown, the average handover probability of each method increases with the increase of the moving speed, which is mainly due to the high mobility of the UE. The proposed method can better adjust the handover threshold and trigger time according to the distance from the UE to the base station and the SINR conditions of the serving cell and the target cell, so the handover probability is lower than the comparative methods under different moving speeds; the sensitivity of the HFT-based method is lower in this scenario, which cannot timely adjust the handover control parameters, resulting in a higher handover probability; the FLC-based method takes signal quality and moving speed as input to adjust the method, which slows down the increase of handover probability to a certain extent.

[0199] Ping-pong handover refers to the back and forth switching between the serving cell and the target cell in a short time, and a lower handover threshold is more likely to cause ping-pong handover. As the user handover probability increases, the probability of ping-pong handover will also increase, which is mainly due to the fact that both the handover probability and the ping-pong handover probability are closely related to the handover threshold. In the proposed method, the value of the utility function will increase when the SINR of the neighboring cell is higher or the UE is close to the cell edge, thereby triggering the adjustment of the handover threshold and the trigger time, effectively preventing ping-pong handover. In order to verify the stability of the method, the average ping-pong handover probability of each method is evaluated under the scenario of moving speed of 30 km / h. The simulation results are shown in Figure 6 As shown, the average ping-pong handover probability of the proposed method is lower than that of the comparative methods. The FLC-based method considers the UE's moving speed, reference signal received power, and received quality, and can timely adjust the handover threshold according to the signal quality, making it perform better than the HFT-based method. However, the FLC-based method only adjusts the handover threshold, ignoring the trigger time, resulting in a slightly higher ping-pong handover probability than the proposed method.

[0200] Wireless link failure is usually caused by suboptimal handover control parameter settings. Suboptimal handover control parameter settings can cause handover too late, handover too early, or handover to the wrong cell, thereby increasing the probability of wireless link failure, especially during high UE moving speed. In addition, as the UE moving speed increases, the Doppler effect also increases the probability of wireless link failure. The average wireless link failure probability of different optimization methods under each moving speed scenario is shown in Figure 7 As shown, the wireless link failure probability of each method increases with the increase of the speed. The proposed method considers the SINR and cell load of the serving cell and the target cell, and triggers the adjustment of the handover threshold and the trigger time when the load of the target cell is high to avoid wireless link failure, so that its wireless link failure rate is the lowest under different moving speed scenarios. The HFT-based method has a higher wireless link failure probability due to the lack of consideration of moving speed. Although the FLC-based method has a lower wireless link failure probability than the HFT-based method by taking the moving speed as input, the lack of trigger time update results in a slightly higher wireless link failure probability than the proposed method.

[0201] The present application aims to improve the stability of cell handover in UDN, and proposes a handover parameter optimization method based on utility function. This method dynamically adjusts the HCPs settings according to user state and network environment to adapt to different network environments and user needs, mainly including the following steps: first, using the bounded functions of SINR, cell load and user-to-base station distance to evaluate user state and network environment; second, taking the output of the three bounded functions as the input of the automatic weight function to calculate the relative importance of each network factor on the handover decision; finally, calculating the utility function value and optimizing the user's HCPs settings. In order to verify the effectiveness of the method, the proposed method is simulated and compared with other literature handover parameter optimization methods. The performance difference between the proposed method and the existing method is evaluated from three aspects of handover probability, ping-pong handover probability and wireless link failure probability. The simulation results show that the proposed method is superior to the comparison method in various mobile speed scenarios, and improves the stability of the handover process.

[0202] Finally, it should be pointed out that the above embodiments are only used to illustrate the technical solutions of the present application and are not limiting. Although the present application has been described in detail with reference to the preferred embodiments, it should be understood by those skilled in the art that the technical solutions of the present application can be modified or replaced equivalently without departing from the purpose and scope of the technical solutions, which should be covered in the scope of the claims of the present application.

Claims

1. A method for optimizing handover parameters based on utility function, characterized in that: The method comprises the following steps: S1: defining a problem Suppose the switching trigger event is an A3 event, which is defined according to the third generation partnership project (3GPP) organization, and the definition is shown in formula (1): Mn+Ofn+Ocn-Hys>Mp+Ofp+Ocp+Off (1) Wherein: Mn and Mp are measurement results of adjacent cells and serving cells, respectively, and the measurement object is the reference signal received power or quality; Ofn and Ofp are specific bias parameters of the frequency band where the cell system bandwidth is located; Ocn and Ocp are cell individual bias parameters of the adjacent cell and the serving cell, respectively, for load balancing; Off is the A3 event bias value of the serving cell and the adjacent cell; Hys is the A3 event hysteresis parameter, that is, the handover threshold in the handover trigger condition; Simplify the above switching trigger event, and the mathematical expression after simplification is shown in formula (2): R T > R S + HOM A (2) wherein: R S and R T respectively represent the RSRP of the serving cell and the target cell; HOM A is a dynamically adjusted handover threshold, the purpose of which is to reduce handover failures and ping-pong handovers when the UE continuously switches between two cells; cell handover is triggered when the RSRP of the target cell satisfies equation (2) within a triggering time; The cell handover control parameter optimization problem is expressed by formula (3): where P denotes the handover probability, ping-pong handover probability and radio link failure probability, which are optimized by selecting appropriate TTT A and HOM A TTT A and HOM A denote the dynamically adjusted triggering time and handover threshold, respectively; constraint (3-3a) ensures that each UE is associated with only one cell, b ij is the user association indicator; constraints (3-3b) and (3-3c) ensure that the selected TTT A and HOM A do not exceed the ranges, which are 0-5.12s and 0-10dB, respectively; and (3-3d) is a binary constraint on the user association indicator, b ij with 1 indicating that the corresponding user is associated with the cell; the UE periodically measures the RSRP of all serving cells and reports the measurement results; S2: establishing a system model Consider two layers of HetNets, consisting of N m One macro cell and N s It consists of microcells; a collection of macrocells is used The set of micro-regions is represented by... It is indicated that the set of all cells is defined as E = M∪S, using e i Indicates E as the middle cell, N all =N m +N s The total number of cells; three micro base stations are deployed within the coverage area of ​​one macro base station. Users are randomly generated in each macro cell and micro cell, and the UE set is defined as U = {u1, u2, ..., u...} K }, where K is the total number of UEs in all cells; S3: dynamic parameter optimization Including the input of the handover index and the generation of the bounded function, the calculation of the corresponding weight of the bounded function, the calculation of the utility function value, and the output of the handover control parameter; the state of the UE and the network environment are evaluated through the utility function, and the higher the utility function value is, the better the communication environment of the UE is, and vice versa. The communication environment of the UE is poor; when the utility function value rises, it indicates that the necessity of cell handover decreases, and the handover control parameter is adjusted upward to avoid ping-pong handover; when the utility function value decreases, it indicates that the necessity of cell handover is higher, and the handover control parameter needs to be adjusted downward to avoid radio link failure; S31: establishing a bounded utility function SINR, cell load and distance between UE and base station are represented by SINR function f(γ), load function f(L) and UE distance function f(d) whose weights are represented by ω γ , ω L and ω d respectively; the utility function is represented as shown in equation (4): f WF (γ, L, d) = ω γ f(γ) + ω L f(L) + ω d f(d) (4) In the formula, f WF (γ, L, d) is divided into two parts; the first part is the bounded sub-functions f(γ), f(L) and f(d); the second part is the weight ω of each sub-function γ , ω L , ω d ; S32: calculating the weight of the bounded function A weight value ω is generated for each bounded function by a weight function n and their sum equals 1; the weight function uses the formula (11) shown: wherein ω n represents the weight of the function f(x n ), ω n is ω γ , ω L or ω d ; f(x n ) is the corresponding nth bounded function, whose weight needs to be calculated; F represents the total number of indexes used to adjust the value of the cell handover control parameter HCPs, three indexes γ, L and d are considered, and F is set to 3; f(x i ) is the ith bounded function, i = 1,..., F; The calculation method of the SINR bounded function weight is shown in formula (12): In the formula, ω γ is the weight of the SINR bounded function; f(γ) is the bounded function value; F is the number of all bounded functions, taking 3; ω L and ω d are calculated in a similar way to the formula. S33: dynamic estimation of the handover control parameter After obtaining the bounded function of different handover indexes and the corresponding weight, and calculating the utility function value; the output of the utility function will be used to dynamically estimate the HCPs setting of each user; the following is the specific adjustment of the handover threshold and the trigger time: (1) Handover threshold adjustment The dynamically adjusted handover threshold value HOM A is expressed using equation (13): HOM A = H + AM (13) Wherein: H is the handover threshold value at the previous moment, and the initial value H0 is defined as the average value of the maximum handover threshold and the minimum handover threshold; ΔM is the dynamically adjusted part; the calculation method of H0 is shown in formula (14): where M max and M min are the maximum and minimum values of the switching threshold; let us assume that these values are 10 dB and 0 dB, respectively; The dynamic adjustment part of the handover threshold setting is shown in formula (15): wherein the calculation of the handover threshold initial value H0 is shown in equation (14); the utility function f WF (γ, L, d) is shown in equation (4); when the target cell SINR is less than the threshold and the serving cell SINR is greater than the threshold, f WF (γ, L, d) controls the magnitude of H0 results in a very low handover threshold setting, resulting in early handover, handover to the wrong cell and ping-pong handover, in which case f WF (γ, L, d) is replaced by 1 γ (γ, L, d) is replaced by -1 WF (γ, L, d) is replaced by 1 γ (γ, L, d) is replaced by -1 (2) Trigger time adjustment The range of the trigger time interval has been specified by the 3GPP organization, and the setting range is 0-5.12 seconds; The adjustment method of the trigger time is shown in formula (16): where: AT is the instant adjustment trigger time; T is the trigger time threshold level initially defined as a fixed appropriate interval, then T will equal AT in the next adjustment; p is the reference adjustment value of the trigger time specified; Q is the trigger threshold of the utility function change value, when the utility function change value is higher than Q, the adjustment of the switching parameter is triggered; Af WF is the change value of the utility function f WF (γ, L, d) relative to the last time Adaptive time-to-trigger value TTT A is expressed by formula (17): where the constants p and Q are used to adjust the sensitivity of the time-to-trigger update, with values of 0.04 s and 0.1 for p and Q, respectively; T max is set to 5.12 s, and T min are set to 0, which are the maximum and minimum values of the time-to-trigger, respectively, as specified by 3GPP; the initial value of the time-to-trigger for each handover parameter self-optimization method is assumed to be 100 ms; In the S31, the bounded function corresponding to different handover indexes is specifically expressed as follows: (1) Bounded SINR function f(γ) In the cell handover process, the trigger method based on the received signal strength, that is, the difference between the serving cell and the selected target cell RSS and the specific handover threshold are used for decision; the RSS is measured at the user side, then reported to the serving cell, and finally the serving cell decides whether to trigger the handover; the best parameter representing RSS is SINR, which takes into account noise and interference; the bounded SINR function is used as the input of the utility function; the user u k From the cell e i The received downlink signal-to-noise ratio The calculation method is shown in formula (5): wherein: is the received downlink RSRP from the serving cell e k ; i is the sum of the downlink power of all interfering cells except the serving cell e i ; is the sum of the downlink power of all interfering cells except the serving cell e i ; is the channel gain between the user u k and the cell e i , taking into account path loss and fading effects; δ 2 is the noise power, i.e. a measure of the total noise in a given channel bandwidth; The difference between the SINR of the serving cell and the target cell is represented by the bounded SINR function; when the bounded SINR function is used in the utility function, its output needs to be between [-1, 1]; in order to ensure that the output value of the bounded SINR function is between [-1, 1], the difference value of SINR is divided by the maximum SINR level; the bounded SINR function f(γ) in the utility function is represented by formula (6): wherein: γ S and γ T represent the SINR of the user at the serving cell and the target cell, respectively, γ max represents the maximum SINR received at the UE; the acceptable SINR range is -10~30dB, and it is assumed that the maximum SINR value is 30dB, i.e., γ max = 30; (2) Bounded load function f(L) The bounded load function is inputted as the utility function; the cell e i The load of the cell e The representation method is shown as formula (7): In the formula, B W represents the system bandwidth, β i is the ratio of the total resources of the cell e i allocated to all active UEs, defined as formula (8): ​ wherein: is the total resource allocated to all active users in the cell i is the resource allocated to user u k ; is the total resource allocated to all active users in the cell The load of the serving cell and the target cell is evaluated by the load boundary function; the bounded input function in the utility function value needs to have a unified boundary, and the difference between the cell loads is divided by the maximum load capacity of the cell, so that the output of f(L) is limited between [-1, 1]; the bounded load function f(L) in the utility function is expressed by formula (9): wherein: L T and L S denote the load of the target cell and the serving cell, respectively; L max is the maximum load capacity of the cell. (3) Bounded distance function f(d) When the UE is close to the cell edge, it is more likely to trigger the handover condition, but when the signal quality of the serving cell is good enough, it will cause the probability of ping-pong handover to increase; by considering the bounded distance function, the utility function is adjusted, and when the UE is close to the cell edge, the value of the bounded distance function will increase, so that the utility function value increases and triggers the up-regulation of the handover control parameter, to a certain extent, to alleviate the ping-pong effect; when the user is at the edge of the cell, the weight of the bounded distance function will be low, at this time, the influence of SINR and load on the utility function value is greater; the output range of the bounded distance function f(d) is [-1, 1], and its expression is shown in formula (10): where: d max represents the maximum acceptable distance between the UE and the base station, assuming d max is the cell radius.

Images