[0043] The specific embodiments of the present invention will be described below in conjunction with the drawings in the specification:
[0044] See figure 1 Shown is a conceptual diagram of the LTE-A self-organizing cellular network adopted by the present invention. Any base station only communicates with its neighboring base stations for user information and traffic information. Although there is no such information exchange between non-adjacent base stations, they can influence each other through common adjacent base stations.
[0045] See figure 2 Shown is a schematic diagram of co-channel interference. This figure shows that only in the same time, a cell that uses the same resource block as the source base station will cause co-frequency interference to users served by the source base station.
[0046] See image 3 As shown, the flow chart is jointly considered for the number of users and traffic information. This figure explains the cooperation process of user number information and traffic information.
[0047] See Figure 4 As shown, it is a conceptual diagram of algorithm complexity reduction based on user location. This figure illustrates the idea of determining the shrinking range of the receiving set based on the location of the user in the cell.
[0048] See Figure 5 Shown is a schematic diagram of the random distribution of users. In the picture Indicates the location of the base station, Indicates user location
[0049] Step (1), for the 19-cell self-organizing network model, set the station spacing as Meter. And based on the consideration of frequency resource management, each communication cell is divided into three sectors, and directional antennas with an argument of 120° are used. Each sector is allocated 25 different resource blocks, and the bandwidth of each resource block is 180 kHz. All users are randomly distributed in 19 communication cells, and it is assumed to keep the position unchanged, and always exchange data with the base station to which they belong. In order to save frequency resources, it is assumed that each user is allowed to occupy only one resource block, so the transmission power of the base station for the user will be adjusted according to the data transmission rate selected by the user. The maximum transmission power of the base station is set to 43dBm, and it is set to allocate to each downlink data transmission channel, that is, the minimum transmission power and maximum transmission power of each user are 15dBm and 30dBm, respectively. Set the default allocated power for each user as the minimum transmit power.
[0050] Step (2), for a single user UE i , Its received SINR is:
[0051] γ i = P i , recieved I i , intra + I i , inter + P N
[0052] Where P i,recieved Is the received power of the useful signal received by the user end, I i,intra Is the intra-cell co-channel interference, I i,inter Co-channel interference between cells, P N It is additive white Gaussian noise with a noise density of -174dBm/Hz. Since this method uses OFDM technology, there is no co-channel interference in the communication cell, that is, I i,intra =0, and the inter-cell co-channel interference is the same time as the UE i All users that use the same resource block and affect it ∑UE i'≠i Of the base station for the UE i The power superposition, namely ∑P i'≠i,recieved. To get P i,recieved , The present invention adopts WINNER+'s non-visual link (Non-Line-Of-Sight, NLOS) urban macro cell (Urban Macrocell, UMa) channel model:
[0053] PL=(44.9-6.55log 10 (h BS ))log 10 (d)+5.83log 10 (h BS )+14.78+34.97log 10 (f c )
[0054] Where base station height h BS Set to 25 meters, carrier frequency f c Set to 2GHz, and d is the geometric distance between the base station and the user. Combined with the above formula and other related parameters, then P i,recieved Can be expressed as:
[0055] dBm(P i,recieved )=P i +AG-PL-PeL-NF
[0056] Where P i For the UE i Where AG is the antenna gain, PeL is the penetration loss, NF is the noise figure of the user end, and dBm(x) represents the dBm form of x.
[0057] Step (3), according to the known SINR, the UE i The spectral efficiency SE can be expressed as:
[0058] ψ i = 0 if γ i γ min ξ log 2 ( 1 + γ i ) if γ min ≤ γ i γ max ψ max if γ i ≥ γ max
[0059] Where 0≤ξ≤1 is the attenuation factor including the actual loss, γ min Is the minimum SINR that can be used to calculate the spectral efficiency, γ max Is the maximum SINR that can be used to calculate the spectral efficiency, ψ max Is the maximum spectrum efficiency available;
[0060] Step (4), according to the known SE, the UE i The occupied resource block RB can be expressed as:
[0061] β i = [ R W RB ψ i ]
[0062] Where R i Is UE i The requested data transfer rate; W RB Is the bandwidth occupied by a resource block; the operator [x] represents the rounding operation of x; according to the setting in step (1), β i Must be equal to 1, so in β i When ≠1, it needs to be P i Adjust within the allowable range, re-obtain SINR and SE, and calculate the number of resource blocks again.
[0063] Step (5), in β i And P i After no more adjustments, BS j The load factor LF can be expressed as:
[0064] ρ j = X β i , j R B total
[0065] Where β i,j Is subordinate to BS j UE i The number of RBs requested; RB total Is BS j The number of all RBs owned.
[0066] Step (6): After generating the flow information of all base stations, each base station is judged and considered in turn according to the predetermined shutdown sequence:
[0067] Step (6.1), judge BS j In the current judgment, whether it is the first time to enter the judgment to consider.
[0068] Step (6.2), if it is the first time to enter the decision consideration, first count the number of users in the surrounding cells, and average them, if the BS j The number of users is less than the average, and BS j If the traffic load is less than the set closing threshold, it is ready to start searching for a collection of receiving base stations that can accept all the transferred traffic without any QoS loss. If BS j The number of users is not less than the average of users in surrounding cells, and the BS j If the traffic load is less than the set shutdown threshold, consider shutting down other base stations first. If BS j If the traffic load is not less than the set shutdown threshold, the base station will no longer be considered for shutdown.
[0069] Step (6.3), if it is not the first time to enter the judgment consideration, only according to the BS j Consider whether it is suitable for closing, that is, if the BS j If the traffic load is less than the set closing threshold, it is ready to start searching for a collection of receiving base stations that can accept all the transferred traffic without any QoS loss.
[0070] Step (7), calculate BS j The number of users in the corresponding area of the neighboring cells. If the number of users in a certain area is zero, the receiving set containing the cell corresponding to the area is excluded from the search range. The reduced number of trying receiving sets can be expressed as
[0071] T dec = T tran - T loca = X n = 1 n max X l = 1 l total RS n , l - X n = 1 n max - n exc X l = 1 l total = X n = 1 n max C n max n - X n = 1 n max - n exc C n max - n exc n RS n , l
[0072] Where T tran To use the full search method to get the number of receiving sets, T loca In order to exclude the number of receiving sets obtained after excluding the cell with zero users in the corresponding area, RS n,l Is the lth receiving set containing n cells, n max Is the largest number of cells in all receiving sets, l total Is the total number of receiving sets containing n cells, n exc For BS j The number of excluded neighborhoods, C n m = n ! ( n - m ) ! m ! .
[0073] Step (8), from the receiving base station range according to the minimization algorithm, without losing the user communication quality, select a set of optimal sets to receive the transferred users and close the BS j , If there is no receiving set available, BS j Then keep it on:
[0074] Step (8.1), calculate RS n,l The estimated load factor of each base station in the Estimated Load Factor, ELF:
[0075] ρ j',elf =ρ j' +ρ j→j' ≤ρ max (j'≠j)
[0076] Where ρ j' For BS j Neighboring base station BS j' Own load, ρ j→j' For BS j Transfer to BS j' Load, ρ max Is the maximum acceptable load threshold of the base station. If ρ j ' +ρ j→j' ρ max , It is determined that the receiving set is not available.
[0077] If the receiving set is available, sum the estimated loads of all base stations in the set to obtain the total traffic load of the set after receiving all transferred users:
[0078] S n , l = X j A RS n , l ρ j , elf
[0079] Step (8.2), take the maximum load flow of all receiving sets with the same number of base stations, and record the receiving set:
[0080] S n , max = max ( X l = 1 l max S n , l )
[0081] Where l max Is the total number of receiving sets containing n cells. The purpose of taking the maximum load flow is to make full use of the receiving capacity of the receiving base station as much as possible, and to increase the possibility of shutting down other base stations with lower traffic load, because after receiving the transferred user, the transferring user may cause the receiving base station to be unable to shut down.
[0082] In step (8.3), in order to make the transferred users access as few base stations as possible, the minimum value of all the maximum load flows obtained in step (8.2) is taken, and the receiving set with the minimum load flow is the optimal Receiving collection:
[0083] S min i max = min ( X n = 1 n max S n , max )
[0084] Step (9), repeat step (2) to step (8) until all base stations suitable for shutdown are shut down.
[0085] based on Image 6 The dynamic interference effect comparison chart can show that the dynamic interference considerations included in the present invention, that is, the method of updating the transmission power and interference information every time the base station is turned off, can significantly improve the energy saving effect brought by the cell shutdown method. Each simulation data in the figure is derived from 100 different user randomly distributed scenarios. In the picture Indicates a closing method that does not consider dynamic interference (non dynamic interference, non_di), Represents a close method that considers dynamic interference (dynamic interference, di).
[0086] based on Figure 7 The joint consideration effect comparison chart can show that the combined consideration of the number of users and traffic information included in the present invention can increase the number of closed cells to a certain extent. And it is worth noting that for a fixed threshold of the number of users, the energy-saving effects obtained vary, and there is a general trend that the lower the threshold of the number of users, the more base stations are shut down. However, a too low threshold setting will also cause instability of the energy saving effect, and a fixed threshold lacks adaptability to different distribution conditions. The method for limiting the number of dynamic average users proposed in the present invention can dynamically adjust the threshold according to the number of users in surrounding cells and obtain certain energy-saving benefits, that is, increase the number of closed base stations. In the picture Indicates the di close method, Represents the di close method (dynamic interference number 15, di_num_15) that considers the limit of the number of users and the threshold is 15, Indicates the di close method (dynamic interference number 17, di_num_17) that considers the limit of the number of users and the threshold is 17, Represents the di close method (dynamic interference number 19, di_num_19) that considers the limit of the number of users and the threshold is 19, Represents the di close method (dynamic interference number 21, di_num_21) that considers the limit of the number of users and the threshold is 21, Represents the di close method (dynamic interference number average, di_num_average) considering the limit of the number of dynamic average users.
[0087] based on Figure 8 The method complexity reduction effect comparison diagram of the present invention can show that the algorithm complexity reduction method based on the user position in the closed cell included in the present invention does not greatly affect the energy saving effect of the system. In the picture Indicates the di close method, Represents the di close method (dynamic interference predict, di_predict) that takes into account the location of the user to reduce the complexity of the algorithm.
[0088] based on Picture 9 The effect comparison diagram of different consideration schemes can show that the joint consideration of dynamic interference, number of users, user location and traffic information can reduce the complexity of the algorithm while increasing the number of closed cells in the system. It can be seen from the figure that in the di_num_average method and the di_pre_num_aver method, the number of closed base stations is greater than 5, reaching about 85%, while the di method and di_predict method remain at about 70%. Although the di_num_average method includes the consideration of reducing the complexity of the algorithm, it does not significantly affect the number of closed cells. In the picture Indicates the di close method, Indicates the di_predict closing method, Represents the di_num_average closing method, Represents the di close method (dynamic interference predict and number average, di_pre_num_aver) that takes into account the limit on the number of users and the location of users.
[0089] based on Picture 10 The connection diagram between the number of closed base stations and the total energy consumption of the system can show that as the number of closed base stations in a cell increases, the total energy consumption of the system also decreases, which further verifies the effectiveness of the cell shutdown method. In the picture Indicates the energy consumption of the system under different considerations, Indicates the number of base stations closed under different considerations. In the figure, di_predict is abbreviated as di_p, di_num_average is abbreviated as di_n, and di_pre_num_aver is abbreviated as di_p_n.
