A high-speed communication transmission rate control method based on Kalman filter prediction

Abstract

The invention relates to a high-speed communication transmission rate control method based on Kalman filter prediction, and belongs to the technical field of network resource allocation. Based on the number of data packets in the real-time service queue at time nT, the number of data packets in the real-time service queue at time (n+1)T is calculated; based on the number of data packets in the real-time service queue at time (n+1)T and the length of the real-time service queue, the occupancy ratio of the real-time service queue at time (n+1)T is calculated, and the service rate corresponding to the interval where the occupancy ratio of the real-time service queue at time (n+1)T is located is used as the service rate of the real-time service queue at time (n+1)T; based on the service rate of the real-time service queue at time (n+1)T, the size of a single data packet in the real-time service queue, the size of a single data packet in the non-real-time service queue, and the size of the available data block, the service rate of the non-real-time service queue at time (n+1)T is calculated, and the transmission rate of each data packet in the real-time service queue and the non-real-time service queue is adjusted.

Description

Technical Field

[0001] The present invention relates to the technical field of network resource allocation, and in particular to a high-speed communication transmission rate control method, device and computer-readable storage medium based on Kalman filter prediction. Background Art

[0002] With the opening and application of high-speed railways and highways, mobile communications in high-speed environments have attracted people's attention. Passenger-centered broadband data services have gradually become the main services of high-speed railway broadband wireless access. When riding on high-speed trains, passengers need to be online in real time. High-speed trains cannot become information islands. Therefore, passengers must be able to keep in touch with the outside world in real time through wireless terminals such as mobile phones and notebooks. This requires that under high-speed mobile conditions, the communication system can still provide passengers with reliable and effective data services.

[0003] In a high-speed operation environment, passenger services are usually divided into real-time services and non-real-time services. Real-time services refer to services that are sensitive to end-to-end delay and delay jitter. Such services usually require real-time or near real-time response, such as video calls, audio transmission and other services. Such services require the network to be able to transmit data quickly and accurately to ensure the smoothness and real-time experience of the user; while non-real-time services refer to services that have a high tolerance for delay. Such services allow larger transmission delays, such as file transfers, emails and other services. The data transmission speed of such services can be slower, but the amount of data transmitted is usually larger, and the network needs to have sufficient bandwidth and stability to support such services.

[0004] In the prior art, a C-eNodeB queue management method based on ToS value priority is proposed to allocate resources for passenger services. Specifically, since the ToS value of real-time services is different from that of non-real-time services, the queue management method based on ToS value priority distinguishes whether the data packet is real-time service or non-real-time service based on the ToS value of the data packet, and sorts each data packet, putting the data packet belonging to the real-time service at the front end of the buffer queue, and putting the data packet belonging to the non-real-time service at the end of the buffer queue, so as to ensure that the real-time service can obtain network spectrum resources first, thereby reducing the reception delay of the real-time service and improving the user experience. However, in a high-speed operation environment, data packets belonging to real-time services arrive at every moment. Therefore, when using this method to allocate spectrum resources, the spectrum resources will be continuously and preferentially allocated to the real-time services that arrive one after another, thereby ignoring the service quality of non-real-time services.

[0005] In summary, the existing high-speed communication spectrum resource allocation method cannot take into account the service quality of both real-time services and non-real-time services at the same time. Summary of the invention

[0006] Therefore, the technical problem to be solved by the present invention is to overcome the problem that the high-speed communication transmission rate control method in the prior art cannot take into account the service quality of both real-time services and non-real-time services at the same time.

[0007] In order to solve the above technical problems, the present invention provides a high-speed communication transmission rate control method based on Kalman filter prediction, comprising:

[0008] Using the Kalman filter algorithm, the number of data packets in the real-time service queue at time nT is calculated based on the number of data packets in the real-time service queue at time (n+1)T;

[0009] Calculate the occupancy ratio of the real-time service queue at time (n+1)T based on the number of data packets in the real-time service queue at time (n+1)T and the length of the real-time service queue, and use the service rate corresponding to the interval where the occupancy ratio of the real-time service queue at time (n+1)T is located as the service rate of the real-time service queue at time (n+1)T;

[0010] Calculate the service rate of the non-real-time service queue at time (n+1)T based on the service rate of the real-time service queue at time (n+1)T, the size of a single data packet in the real-time service queue, the size of a single data packet in the non-real-time service queue, and the available data block size;

[0011] Based on the service rate of the real-time service queue at time (n+1)T and the service rate of the non-real-time service queue at time (n+1)T, the transmission rate of each data packet in the real-time service queue and the non-real-time service queue is adjusted.

[0012] Preferably, the real-time service queue includes a telephone service queue and a video service queue.

[0013] Preferably, based on the number of data packets in the real-time service queue at (n+1)T and the length of the real-time service queue, calculating the occupancy ratio of the real-time service queue at (n+1)T, and taking the service rate corresponding to the interval where the occupancy ratio of the real-time service queue at (n+1)T is located as the service rate of the real-time service queue at (n+1)T, comprises:

[0014] Calculate the telephone service queue occupancy ratio at time (n+1)T based on the number of data packets in the telephone service queue at time (n+1)T and the length of the telephone service queue, and use the service rate corresponding to the interval where the telephone service queue occupancy ratio at time (n+1)T is located as the service rate of the telephone service queue at time (n+1)T;

[0015] Based on the number of data packets in the video service queue at time (n+1)T and the length of the video service queue, the occupancy ratio of the video service queue at time (n+1)T is calculated, and the service rate corresponding to the interval in which the occupancy ratio of the video service queue at time (n+1)T is located is used as the service rate of the video service queue at time (n+1)T.

[0016] Preferably, the calculation formula of the telephone service queue occupancy ratio at time (n+1)T is:

[0017] θ voip (n+1)=Q voip (n+1) / Len voip ,

[0018] Among them, θ voip (n) represents the telephone service queue occupancy ratio at time (n+1)T, Q voip (n+1) represents the number of packets in the telephone service queue at time (n+1)T, Len voip Indicates the length of the telephone service queue;

[0019] The service rate of the telephone service queue at time (n+1)T is:

[0020]

[0021] Among them, μ voip (n+1) represents the service rate of the telephone service queue at time (n+1)T;

[0022] The calculation formula for the video service queue occupancy ratio at time (n+1)T is:

[0023] θ video (n+1)=Q video (n+1) / Len video ,

[0024] Among them, θ video (n+1) represents the video service queue occupancy ratio at time (n+1)T, Q video (n+1) represents the number of packets in the video service queue at time (n+1)T, Len video Indicates the length of the video service queue;

[0025] The service rate of the video service queue at time (n+1)T is:

[0026]

[0027] Among them, μ video (n+1) represents the service rate of the video service queue at time (n+1)T.

[0028] Preferably, the calculation formula for the service rate of the non-real-time service queue at time (n+1)T is:

[0029]

[0030] Among them, μ BE (n+1) represents the service rate of the non-real-time service queue at time (n+1)T, TBs represents the available data block size, S voip Indicates the size of a single data packet in the telephone service queue, S video Indicates the size of a single data packet in the video service queue, S BE Indicates the size of a single data packet in the non-real-time service queue.

[0031] Preferably, the Kalman filter algorithm is used to place each data packet into a real-time service queue or a non-real-time service queue based on the number of data packets in the real-time service queue at time nT, including the ToS value of each data packet arriving at time nT, which specifically includes:

[0032] If the ToS value of the data packet is 0xff, the data packet is placed in the telephone service queue;

[0033] If the ToS value of the data packet is 0xb0, the data packet is placed in the video service queue;

[0034] If the ToS value of the data packet is 0x00, the data packet is placed in a non-real-time service queue.

[0035] Preferably, calculating the number of data packets in the real-time service queue at time (n+1)T based on the number of data packets in the real-time service queue at time nT using a Kalman filter algorithm comprises:

[0036] Based on the number of data packets in the real-time service queue at time nT, a state variable a posteriori estimate of the number of data packets in the real-time service queue at time nT is obtained; based on the state variable a posteriori estimate of the number of data packets in the real-time service queue at time nT and the state transfer matrix, a state variable a priori estimate of the number of data packets in the real-time service queue at time (n+1)T is calculated;

[0037] Based on the a posteriori error covariance matrix and the state transition matrix of the number of data packets in the real-time service queue at time nT, the a priori error covariance matrix of the number of data packets in the real-time service queue at time (n+1)T is calculated;

[0038] Based on the covariance matrix of the observation noise of the number of data packets in the real-time service queue at time nT and the a priori error covariance matrix of the number of data packets in the real-time service queue at time (n+1)T, the Kalman gain of the number of data packets in the real-time service queue at time (n+1)T is calculated;

[0039] Calculate a posterior estimate of the state variable of the number of data packets in the real-time service queue at time (n+1)T based on the prior estimate of the state variable of the number of data packets in the real-time service queue at time (n+1)T, the Kalman gain of the number of data packets in the real-time service queue at time (n+1)T, and the observed value of the number of data packets in the real-time service queue at time (n+1)T;

[0040] Calculate the a posteriori error covariance matrix of the number of data packets in the real-time service queue at time (n+1)T based on the a priori error covariance matrix of the number of data packets in the real-time service queue at time (n+1)T and the Kalman gain of the number of data packets in the real-time service queue at time (n+1)T;

[0041] The number of data packets in the real-time service queue at time (n+1)T is obtained based on the state variable a posteriori estimate of the number of data packets in the real-time service queue at time (n+1)T.

[0042] Preferably, the calculation formula for the state variable a posteriori estimate of the number of data packets in the real-time service queue at time nT is:

[0043]

[0044] Where Q(n) represents the number of packets in the real-time service queue at time nT, Q 1 (n) represents the first-order differential of the number of packets in the real-time service queue at time nT;

[0045] The calculation formula for the a priori estimate of the state variable of the number of data packets in the real-time service queue at time (n+1)T is:

[0046]

[0047] in, represents the state variable prior estimate of the number of packets in the real-time service queue at time (n+1)T, F(n+1,n) represents the state transfer matrix, X(n) represents the state variable a posteriori estimate of the number of packets in the real-time service queue at time nT, and w(n) represents the system noise;

[0048] The calculation formula of the prior error covariance matrix of the number of data packets in the real-time service queue at time (n+1)T is:

[0049]

[0050] in, represents the a priori error covariance matrix of the number of packets in the real-time service queue at time (n+1)T, P(n) represents the a posteriori error covariance matrix of the number of packets in the real-time service queue at time nT, T represents the transpose, and V(n) represents the variance of the Gaussian white noise of the number of packets in the real-time service queue at time nT;

[0051] The calculation formula of the Kalman gain of the number of data packets in the real-time service queue at time (n+1)T is:

[0052]

[0053] Among them, G(n+1) represents the Kalman gain of the number of data packets in the real-time service queue at time (n+1)T, represents the a priori error covariance matrix of the number of data packets in the real-time service queue at time (n+1)T, C represents the observation matrix, and R(n) represents the covariance matrix of the observation noise of the number of data packets in the real-time service queue at time nT;

[0054] The calculation formula for the a posteriori estimate of the state variable of the number of packets in the real-time service queue at time (n+1)T is:

[0055]

[0056] Wherein, X(n+1) represents the a posteriori estimate of the state variable of the number of data packets in the real-time service queue at time (n+1)T, and y(n+1) represents the observed value of the number of data packets in the real-time service queue at time (n+1)T;

[0057] The calculation formula of the a posteriori error covariance matrix of the number of data packets in the real-time service queue at time (n+1)T is:

[0058]

[0059] Wherein, P(n+1) represents the a posteriori error covariance matrix of the number of data packets in the real-time service queue at time (n+1)T, and I represents the identity matrix;

[0060] The calculation formula for the number of data packets in the real-time service queue at time (n+1)T is:

[0061]

[0062] Where Q(n+1) represents the number of packets in the real-time service queue at time (n+1)T. 1 (n+1) represents the first-order differential of the number of data packets in the real-time service queue at time (n+1)T.

[0063] The present invention also provides a high-speed communication transmission rate control device based on Kalman filter prediction, comprising:

[0064] A data packet number calculation module, used to calculate the number of data packets in the real-time service queue at time (n+1)T based on the number of data packets in the real-time service queue at time nT using a Kalman filter algorithm;

[0065] a real-time service queue service rate calculation module, configured to calculate the occupancy ratio of the real-time service queue at time (n+1)T based on the number of data packets in the real-time service queue at time (n+1)T and the length of the real-time service queue, and use the service rate corresponding to the interval where the occupancy ratio of the real-time service queue at time (n+1)T is located as the service rate of the real-time service queue at time (n+1)T;

[0066] A non-real-time service queue service rate calculation module, used to calculate the service rate of the non-real-time service queue at time (n+1)T based on the service rate of the real-time service queue at time (n+1)T, the size of a single data packet in the real-time service queue, the size of a single data packet in the non-real-time service queue, and the size of the available data block;

[0067] The transmission rate adjustment module is used to adjust the transmission rate of each data packet in the real-time service queue and the non-real-time service queue based on the service rate of the real-time service queue at the (n+1)T moment and the service rate of the non-real-time service queue at the (n+1)T moment.

[0068] The present invention also provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the steps of the above-mentioned high-speed communication transmission rate control method based on Kalman filter prediction are implemented.

[0069] The high-speed communication transmission rate control method based on Kalman filter prediction provided by the present application uses the Kalman filter algorithm to predict the number of data packets in the real-time service queue at time (n+1)T, and calculates the occupancy ratio of the real-time service queue at time (n+1)T based on the predicted number of data packets. Since the occupancy ratio of the queue can reflect the load state of the queue, the present application calculates the service rate of the real-time service queue based on the occupancy ratio of the real-time service queue, and calculates the service rate of the non-real-time service queue based on the service rate of the real-time service queue and the available transmission block size, thereby adjusting the data packet transmission rate in the real-time service queue and the non-real-time service queue at time (n+1)T. The present application adjusts the service rate of the real-time service queue based on the queue occupancy ratio of the real-time service, and simultaneously adjusts the service rate of the non-real-time service queue. On the premise of meeting the low delay requirements of the real-time service, the queuing delay of the non-real-time service is reduced, the service fairness between different services is taken into account, and the quality of high-speed communication service is improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0070] In order to make the content of the present invention more clearly understood, the present invention is further described in detail below according to specific embodiments of the present invention in conjunction with the accompanying drawings, wherein:

[0071] Figure 1 Schematic diagram of CRN spectrum sharing architecture of LTE-R system based on C-eNodeB and VG provided in this application;

[0072] Figure 2 Schematic diagram of the LTE EPC data plane protocol stack provided for this application;

[0073] Figure 3 Schematic diagram of IPv4 datagram and DS field provided for this application; wherein, Figure 3 (a) is a schematic diagram of IPv4 datagram. Figure 3 (b) is a schematic diagram showing that DSCP occupies the first 6 bits of the DiffServ DS field;

[0074] Figure 4 A flow chart of a high-speed communication transmission rate control method based on Kalman filter prediction provided in this application;

[0075] Figure 5 A schematic diagram of the high-speed communication transmission rate control principle based on Kalman filter prediction provided in this application;

[0076] Figure 6 A schematic diagram of the prediction results of the number of data packets in the real-time service queue provided by this application; wherein, Figure 6 (a) is a schematic diagram of the prediction result of the number of packets in the telephone service queue. Figure 6 (b) is a schematic diagram of the prediction result of the number of data packets in the video service queue;

[0077] Figure 7 A schematic diagram of telephone service queuing delay after using the method provided by the present application and the method in the prior art; wherein: Figure 7 (a) is a schematic diagram of telephone service queuing delay after using the FIFO queue management strategy, the ToS value priority strategy, the ASREM-3 queue management strategy and the method provided by the present application, Figure 7 (b) is a schematic diagram of telephone service queuing delay after using the ToS value priority strategy, ASREM-3 queue management strategy and the method provided by this application;

[0078] Figure 8 A schematic diagram of video service queuing delay after using the method provided by the present application and the method in the prior art; wherein: Figure 8(a) is a schematic diagram of video service queuing delay after using the FIFO queue management strategy, the ToS value priority strategy, the ASREM-3 queue management strategy and the method provided by the present application. Figure 8 (b) is a schematic diagram of video service queuing delay after using the ToS value priority strategy, ASREM-3 queue management strategy and the method provided by this application;

[0079] Fig. 9 A schematic diagram of non-real-time service queuing delay after using the method provided by the present application and the method in the prior art; wherein, Fig. 9 (a) is a schematic diagram of non-real-time service queuing delay after using the FIFO queue management strategy, the ToS value priority strategy, the ASREM-3 queue management strategy and the method provided by the present application, Fig. 9 (b) is a schematic diagram of non-real-time service queuing delay after using the ToS value priority strategy, ASREM-3 queue management strategy and the method provided by this application;

[0080] Fig.10 A schematic diagram of a single operation time of the method provided in this application and the method in the prior art; wherein, Fig.10 (a) is a schematic diagram of a single running time of the FIFO queue management strategy, the ToS value priority strategy, the ASREM-3 queue management strategy and the method provided by the present application, Fig.10 (b) is a schematic diagram of a single running time of the ToS value priority strategy, the ASREM-3 queue management strategy and the method provided by the present application;

[0081] Fig.11 A schematic diagram of the structure of a high-speed communication transmission rate control device based on Kalman filter prediction provided in this application. DETAILED DESCRIPTION

[0082] The present invention is further described below in conjunction with the accompanying drawings and specific embodiments so that those skilled in the art can better understand the present invention and implement it, but the embodiments are not intended to limit the present invention.

[0083] See also Figure 1 , Figure 1 The figure shows the CRN spectrum sharing architecture of the LTE-R system based on C-eNodeB and VG provided by the present application. In the railway LTE-R private network, the LTE-R eNodeB only provides wireless connection and services for the OBU and only receives messages from the train ground control center. The passenger mobile terminal devices on the train communicate with the operator through the LTE base station (such as Figure 1Line 1) in the train accesses and uses the network. The railway private network and the operator's network do not interfere with each other to ensure the safety of train operation. In order to improve the communication quality of passengers on the train, the private network LTE-R eNodeB is upgraded to LTE-R C-eNodeB, and a VG that can communicate with the LTE-R private network C-eNodeB is added to the roof of the train. When switching across zones, only the OBU and VG need to switch to complete the switching on behalf of a large number of passengers, reducing the impact of the signaling storm on the train-ground communication caused by the need to switch to the new base station at the same time when the train passes through different base stations using the LTE-R private network. A CRN (such as Figure 1 (shown by the dotted oval box in the middle). After the introduction of CRN, OBU and passenger mobile terminal devices are PU (primary user) and SU (secondary user) in CRN (cognitive radio network) respectively. C-eNodeB is responsible for allocating spectrum resources to PU and SU. In CRN, the data of the train ground center will be forwarded preferentially at C-eNodeB. SU that does not use private network resources through VG will obtain network services from the operator's eNodeB through line 1; SU that obtains network services through VG and C-eNodeB will transmit data through line 2. In the link that obtains Internet services through VG and C-eNodeB, C-eNodeB sends SU data to VG on the train through the air interface if there are remaining spectrum resources. VG forwards it to AP through optical fiber, and finally AP forwards it to SU, completing the spectrum sharing of OBU and passengers in the LTE-R private network. In this architecture, the cognitive LTE-R base station (C-eNodeB) has the ability to intelligently sense and adjust the spectrum. It can dynamically adjust its own parameters according to changes in the external spectrum environment and provide reliable spectrum services for primary users.

[0084] like Figure 2As shown in the figure, in the LTE system, after the UE establishes an end-to-end connection with the remote host through the eBodeB, the function of the eNodeB on the data plane is more like a router, responsible for forwarding data from the remote host to the downlink, or forwarding the UE's data to the remote host. The eNodeB is connected to the SGW / P-GW through the S1-U interface. The SGW / P-GW is a major component of the LTE evolved core network, and is more for maintaining end-to-end routing connections and home management in the network. On the data control plane of the LTE protocol stack, the IP packets arriving at the eNodeB are processed in the radio link control layer (RLC), which is responsible for data caching and cascading or segmenting IP packets in each transmission time interval (TTI). When the size of the transport block (TB) in a TTI is larger than the length of an IP packet, multiple IP packets need to be cascaded to improve resource utilization. When the size of the TB is smaller than the length of an IP packet, the IP packet needs to be segmented to correctly send data. In the LTE protocol stack, resource allocation is based on active flows (data to be sent), that is, only users with data to be forwarded can be allocated resources for sending or receiving data. After each active UE establishes a connection with the eNodeB, the eNodeB will establish an independent packet buffer queue belonging to the UE at its RLC layer. The RLC layer is responsible for the management of IP packets. In each TTI, the scheduler of the MAC layer directly checks the queue status of each user at the RLC layer and completes resource allocation. If the queue length of user i at the RLC layer is 0, that is, user i does not need to send data, the scheduler will not allocate resources to user i.

[0085] In order to provide different types of services on the Internet, IETF proposed the Integrated Services Protocol (IntServ) and the Resource Reservation Protocol (RSVP), but they are both relatively complex and not conducive to implementation in large-scale networks, especially dedicated communication networks such as railway communications that require a simple network structure to reduce network delays and ensure the safe operation of trains.

[0086] In view of the high complexity of IntServ and RSVP protocols, IETF proposed Differentiated Services (DiffServ). DiffServ only needs to add the differentiated services function in the router, without changing the network structure. DiffServ redefines the original 8-bit ToS field of the IPv4 protocol (used to specify the transmission priority and processing method of the data packet in the network, so that the network equipment can process the data packet differently according to its value) and the Traffic Class (TC) field of IPv6 [RFC2460] as the Differentiated Services (DS) field, so that the router can process the arriving packets according to the value of the DS field. At present, according to the recommended standard [RFC 2474], the DS field only uses the first 6 bits, namely the Differentiated Service Code Point (DSCP), and the last two bits are reserved and not used, denoted as CU.

[0087] The ToS field is part of the IP datagram format, such as Figure 3 As shown in (a), ToS occupies 8 bits in the header of the IP datagram and is used to describe the quality of service expected by the IP datagram, including the allocation of priority, delay, throughput and reliability. In view of the particularity of the DS field, in order to solve the problem of excessive delay of real-time services, C-eNodeB uses different ToS values ​​for different services to provide active queue management.

[0088] The prior art provides a C-eNodeB queue management method based on ToS value priority, which is used to improve the service quality of real-time services and reduce the reception delay of real-time services. Specifically, in the queue management strategy based on ToS value priority, all secondary users have only one buffer queue at the C-eNodeB to store data packets. In order to minimize the average reception delay of real-time services, secondary users using different services are distinguished. In each transmission time interval, the C-eNodeB distinguishes whether the data packet is a real-time service or a non-real-time service based on the ToS value of the data packet, and puts the data packet belonging to the real-time service at the front of the buffer queue so that the real-time service can be served as soon as possible, reducing the queuing delay, and putting the data packet belonging to the non-real-time service at the end of the queue. However, in a high-speed communication environment, data packets belonging to real-time services will arrive in each transmission time interval, and this method continuously puts the arriving real-time services at the front of the buffer queue. Although the queuing delay of the real-time service is reduced, the queuing delay of the non-real-time service is increased, which greatly sacrifices the service quality of the non-real-time service.

[0089] Based on the above problems, this application provides a high-speed communication transmission rate control method based on Kalman filter prediction, please refer to Figure 4, the method specifically comprises:

[0090] S10: Calculate the number of data packets in the real-time service queue at time (n+1)T based on the number of data packets in the real-time service queue at time nT using a Kalman filter algorithm;

[0091] S20: Calculate the occupancy ratio of the real-time service queue at time (n+1)T based on the number of data packets in the real-time service queue at time (n+1)T and the length of the real-time service queue, and use the service rate corresponding to the interval where the occupancy ratio of the real-time service queue at time (n+1)T is located as the service rate of the real-time service queue at time (n+1)T;

[0092] S30: Calculate the service rate of the non-real-time service queue at time (n+1)T based on the service rate of the real-time service queue at time (n+1)T, the size of a single data packet in the real-time service queue, the size of a single data packet in the non-real-time service queue, and the available data block size;

[0093] S40: Based on the service rate of the real-time service queue at time (n+1)T and the service rate of the non-real-time service queue at time (n+1)T, the transmission rate of each data packet in the real-time service queue and the non-real-time service queue is adjusted.

[0094] The present application adjusts the service rate of the real-time service queue based on the queue occupancy ratio of the real-time service, and simultaneously adjusts the service rate of the non-real-time service queue. On the premise of meeting the low delay requirement of the real-time service, the queuing delay of the non-real-time service is reduced, the service fairness between different services is taken into account, and the quality of high-speed communication services is improved.

[0095] Specifically, since the most representative real-time services are telephone services and video services, among which, the telephone service digitizes the analog sound signal and transmits it in real time in the form of data packets on the IP data network; and the video service realizes remote audio-visual communication and information sharing through real-time transmission of video signals, therefore, in some embodiments of the present application, the real-time service is divided into telephone service and video service, and accordingly, the real-time service queue includes a telephone service queue and a video service queue.

[0096] In some embodiments of the present application, before step S10, the process further includes placing each data packet into a real-time service queue or a non-real-time service queue based on the ToS value of each data packet arriving at time nT, which specifically includes:

[0097] If the ToS value of the data packet is 0xff, the data packet is placed in the telephone service queue;

[0098] If the ToS value of the data packet is 0xb0, the data packet is placed in the video service queue;

[0099] If the ToS value of the data packet is 0x00, the data packet is placed in a non-real-time service queue.

[0100] For example, Figure 5 As shown in Figure 1, when the data packet arrives at the RLC layer of the C-eNodeB, after being filtered by the C-eNodeB, the PU data packets enter the queue in the order of arrival. In the RLC layer, C-eNodeB parses each data packet of the SU that arrives, and puts the data packet into the corresponding queue by obtaining the ToS value of the DS field. When the ToS value of the arriving data packet is 0x00, the data packet is put into the non-real-time service queue. When the ToS value of the arriving data packet is 0xff, the data packet is put into the telephone service queue. When the ToS value of the arriving data packet is 0xb0, the data packet is put into the video service queue.

[0101] Specifically, the Kalman filter algorithm refers to a dynamic discrete-time system that can be represented by the state equation of the state variable and the observation equation of the observation variable:

[0102] X(n+1)=F(n+1,n)X(n)+V1(n),

[0103] Y(n)=C(n)X(n)+V2(n),

[0104] Among them, X(n) represents the unobservable state variable at time n, F(n+1,n) represents the matrix describing the system state transition, V1(n) represents the process noise variable at time n, Y(n) represents the observation vector at time n, C(n) represents the observed variable at time n, and V2(n) represents the observation noise;

[0105] Since the number of packets in the queue changes over time, this application uses the Kalman filter algorithm to predict the number of packets in the queue at time (n+1)T based on the number of packets in the queue at time nT:

[0106] This application calculates the state variable posterior estimate of the number of data packets in the real-time service queue at time nT based on the number of data packets in the real-time service queue at time nT:

[0107]

[0108] Where Q(n) represents the number of packets in the real-time service queue at time nT, Q 1 (n) represents the first-order differential of the number of data packets in the real-time service queue at time nT, which is used to describe the rate of change of the number of data packets in the real-time service queue;

[0109] When the sampling interval is small enough, the calculation formula for the number of data packets in the real-time service queue at time (n+1)T is:

[0110] Q(n+1)=Q(n)+TQ 1 (n),

[0111] Q 1 (n) = (γ(n) - μ(n)) / T,

[0112] Among them, Q(n+1) represents the number of data packets in the real-time service queue at time (n+1)T, γ(n) represents the number of data packets arriving at the real-time service queue at time nT, and μ(n) represents the number of data packets served in the real-time service queue at time nT. Since the number of secondary users in the system, service type, duration, channel state change and other factors will affect Q(n), the present application regards these influencing factors as system noise w(n), and the calculation formula of Q(n+1) is:

[0113] Q(n+1)=Q(n)+TQ 1 (n)+u(n),

[0114] Q 1 (n+1)=Q 1 (n)+v(n),

[0115] Then the state variable of the number of data packets in the real-time service queue at time (n+1)T can be expressed as:

[0116] X(n+1)=F(n+1,n)X(n)+w(n),

[0117] Among them, F(n+1,n) represents the state transfer matrix, w(n) represents the system noise, which conforms to the N(0,V(n)) distribution.

[0118] The observation equation for the number of data packets in the real-time service queue at time (n+1)T can be expressed as:

[0119] y(n+1)=CX(n+1)+e(n+1),

[0120] Where C represents the observation matrix, C = [1 0], e(n+1) represents the observation noise at time (n+1)T, which conforms to the Gaussian noise of N(0, R(n+1));

[0121] Based on the above derivation process, the specific implementation of step S20 is:

[0122] S200: obtaining a state variable a posteriori estimate of the number of data packets in the real-time service queue at time nT based on the number of data packets in the real-time service queue at time nT, and calculating a state variable a priori estimate of the number of data packets in the real-time service queue at time (n+1)T based on the state variable a posteriori estimate of the number of data packets in the real-time service queue at time nT and the state transfer matrix;

[0123] Specifically, the calculation formula for the a posteriori estimate of the state variable of the number of data packets in the real-time service queue at time nT is:

[0124]

[0125] Where Q(n) represents the number of packets in the real-time service queue at time nT, Q 1 (n) represents the first-order differential of the number of packets in the real-time service queue at time nT;

[0126] The calculation formula for the a priori estimate of the state variable of the number of data packets in the real-time service queue at time (n+1)T is:

[0127]

[0128] in, represents the state variable prior estimate of the number of packets in the real-time service queue at time (n+1)T, F(n+1,n) represents the state transfer matrix, X(n) represents the state variable a posteriori estimate of the number of packets in the real-time service queue at time nT, and w(n) represents the system noise;

[0129] S201: Calculate the a priori error covariance matrix of the number of data packets in the real-time service queue at time (n+1)T based on the a posteriori error covariance matrix of the number of data packets in the real-time service queue at time nT and the state transfer matrix;

[0130] Specifically, the calculation formula of the prior error covariance matrix of the number of data packets in the real-time service queue at time (n+1)T is:

[0131]

[0132] in, represents the a priori error covariance matrix of the number of packets in the real-time service queue at time (n+1)T, P(n) represents the a posteriori error covariance matrix of the number of packets in the real-time service queue at time nT, T represents the transpose, and V(n) represents the variance of the Gaussian white noise of the number of packets in the real-time service queue at time nT;

[0133] S202: Calculate the Kalman gain of the number of data packets in the real-time service queue at time (n+1)T based on the covariance matrix of the observation noise of the number of data packets in the real-time service queue at time nT and the a priori error covariance matrix of the number of data packets in the real-time service queue at time (n+1)T;

[0134] Specifically, the calculation formula of the Kalman gain of the number of data packets in the real-time service queue at time (n+1)T is:

[0135]

[0136] Among them, G(n+1) represents the Kalman gain of the number of data packets in the real-time service queue at time (n+1)T, represents the a priori error covariance matrix of the number of data packets in the real-time service queue at time (n+1)T, C represents the observation matrix, and R(n) represents the covariance matrix of the observation noise of the number of data packets in the real-time service queue at time nT;

[0137] S203: Calculate a posterior estimate of the state variable of the number of data packets in the real-time service queue at time (n+1)T based on the prior estimate of the state variable of the number of data packets in the real-time service queue at time (n+1)T, the Kalman gain of the number of data packets in the real-time service queue at time (n+1)T, and the observed value of the number of data packets in the real-time service queue at time (n+1)T;

[0138] Specifically, the calculation formula for the a posteriori estimate of the state variable of the number of data packets in the real-time service queue at time (n+1)T is:

[0139]

[0140] Wherein, X(n+1) represents the a posteriori estimate of the state variable of the number of data packets in the real-time service queue at time (n+1)T, and y(n+1) represents the observed value of the number of data packets in the real-time service queue at time (n+1)T;

[0141] S204: Calculate the a posteriori error covariance matrix of the number of data packets in the real-time service queue at time (n+1)T based on the a priori error covariance matrix of the number of data packets in the real-time service queue at time (n+1)T and the Kalman gain of the number of data packets in the real-time service queue at time (n+1)T;

[0142] Specifically, the calculation formula of the a posteriori error covariance matrix of the number of data packets in the real-time service queue at time (n+1)T is:

[0143]

[0144] Wherein, P(n+1) represents the a posteriori error covariance matrix of the number of data packets in the real-time service queue at time (n+1)T, and I represents the identity matrix;

[0145] S205: obtaining the number of data packets in the real-time service queue at time (n+1)T based on the state variable a posteriori estimate of the number of data packets in the real-time service queue at time (n+1)T;

[0146] Specifically, the calculation formula for the number of data packets in the real-time service queue at time (n+1)T is:

[0147]

[0148] Where Q(n+1) represents the number of packets in the real-time service queue at time (n+1)T. 1 (n+1) represents the first-order differential of the number of data packets in the real-time service queue at time (n+1)T.

[0149] Furthermore, in some embodiments of the present application, the specific implementation of step S30 is as follows:

[0150] Based on the number of data packets in the telephone service queue at time (n+1)T and the length of the telephone service queue, the telephone service queue occupancy ratio at time (n+1)T is calculated, and the service rate corresponding to the interval where the telephone service queue occupancy ratio at time (n+1)T is located is used as the service rate of the telephone service queue at time (n+1)T;

[0151] Specifically, the calculation formula for the telephone service queue occupancy ratio at time (n+1)T is:

[0152] θ voip (n+1)=Q voip (n+1) / Len voip ,

[0153] Among them, θ voip (n) represents the telephone service queue occupancy ratio at time (n+1)T, Q voip (n+1) represents the number of packets in the telephone service queue at time (n+1)T, Len voip Indicates the length of the telephone service queue;

[0154] The service rate of the telephone service queue at time (n+1)T is:

[0155]

[0156] Among them, μ voip (n+1) represents the service rate of the telephone service queue at time (n+1)T;

[0157] Based on the number of data packets in the video service queue at time (n+1)T and the length of the video service queue, the occupancy ratio of the video service queue at time (n+1)T is calculated, and the service rate corresponding to the interval where the occupancy ratio of the video service queue at time (n+1)T is located is used as the service rate of the video service queue at time (n+1)T;

[0158] Specifically, the calculation formula for the video service queue occupancy ratio at time (n+1)T is:

[0159] θ video (n+1)=Q video (n+1) / Len video ,

[0160] Among them, θ video (n+1) represents the video service queue occupancy ratio at time (n+1)T, Q video (n+1) represents the number of packets in the video service queue at time (n+1)T, Len video Indicates the length of the video service queue;

[0161] The service rate of the video service queue at time (n+1)T is:

[0162]

[0163] Among them, μ video (n+1) represents the service rate of the video service queue at time (n+1)T;

[0164] Specifically, In a specific example of this application,

[0165] Although telephone services and video services are both real-time services, due to their different natures, the interval distribution of their queue occupancy ratios is different, and the service rates corresponding to different intervals are also different.

[0166] Furthermore, since the service rate of the telephone service queue, the service rate of the video service queue and the service rate of the non-real-time service queue in a transmission time interval satisfy: TBs ≥ μ voip (n+1)*S voip +μ video (n+1)*S video +μ BE (n+1)*S BE, where TBs represents the available data block size in the transmission time interval, and the value of TBs depends on the number of resource block groups (RBGs) available to the remaining secondary users in the transmission time interval; therefore, the present application calculates the service rate of the non-real-time service queue at time (n+1)T based on the service rate of the telephone service queue at time (n+1)T, the size of a single data packet in the telephone service queue, the service rate of the video service queue, the size of a single data packet in the video service queue, the size of a single data packet in the non-real-time service queue, and the available data block size;

[0167] Specifically, the calculation formula for the service rate of the non-real-time service queue at time (n+1)T is:

[0168]

[0169] Among them, μ BE (n+1) represents the service rate of the non-real-time service queue at time (n+1)T, TBs represents the available data block size, S voip Indicates the size of a single data packet in the telephone service queue, S video Indicates the size of a single data packet in the video service queue, S BE Indicates the size of a single data packet in the non-real-time service queue.

[0170] In order to verify the accuracy of the number of data packets predicted based on the Kalman filter algorithm, the embodiment of the present application also provides a schematic diagram of the prediction result of the number of data packets in the real-time service queue, such as Figure 6 (a) in FIG. 1 is a schematic diagram showing the prediction result of the number of packets in the telephone service queue. Figure 6 (b) is a schematic diagram showing the prediction result of the number of data packets in the video service queue;

[0171] from Figure 6 As can be seen from (a) in the figure, at the beginning of the simulation, the number of packets in the telephone service queue is 0. As more and more packets arrive at the RLC layer buffer of the C-eNodeB, although the number of packets in the telephone service queue jumps to a larger value of 140 at about 500 milliseconds and fluctuates between 120 and 160, it can quickly approach the actual value and fluctuate within a small range with the actual value.

[0172] from Figure 6 As can be seen from (b) in the figure, at the beginning of the simulation, the number of data packets in the video service queue changes slowly and slightly. After 4 seconds, the number of data packets allocated to the video service queue continues to increase, maintaining the same growth trend as the actual value and is very close to the real value.

[0173] In addition, in order to verify the effectiveness of the high-speed communication transmission rate control method based on Kalman filter prediction provided by the present application, the embodiment of the present application also provides a schematic diagram of the queuing delay duration of real-time services and non-real-time services after using different methods;

[0174] like Figure 7 The figure shows the queuing delay duration of telephone services after using different methods. Figure 7 (a) is a schematic diagram of telephone service queuing delay after using the FIFO queue management strategy, the ToS value priority strategy, the ASREM-3 queue management strategy and the method provided by the present application, Figure 7 (b) is a schematic diagram of telephone service queuing delay after using the ToS value priority strategy, ASREM-3 queue management strategy and the method provided by this application;

[0175] from Figure 7 It can be seen that in terms of queuing delay, the FIFO service strategy has the largest queuing delay, especially when there are a large number of secondary users in the system. The average queuing delay of telephone services under the other three methods is relatively small; compared with the ToS value priority strategy and the ASREM-3 algorithm, the queuing delay of telephone services after using the method provided by the present application is relatively large. When the number of SUs in the system is 100, the queuing delay reaches the maximum, close to 30 milliseconds, but is much smaller than the maximum queuing delay of 130 milliseconds required by the system. The reason is that the service rate of the telephone service queue is adjusted according to the predicted number of data packets. On the premise of meeting the queuing delay requirements of the telephone service, the service rate of the telephone service queue is reduced. Therefore, it can be considered that although the present application increases the average queuing delay of the telephone service to a certain extent, it still does not affect the real-time nature of the telephone service.

[0176] like Figure 8 The figure shows the queuing delay duration of video services after using different methods. Figure 8 (a) is a schematic diagram of video service queuing delay after using the FIFO queue management strategy, the ToS value priority strategy, the ASREM-3 queue management strategy and the method provided by the present application. Figure 8 (b) is a schematic diagram of video service queuing delay after using the ToS value priority strategy, ASREM-3 queue management strategy and the method provided by this application;

[0177] from Figure 8It can be seen that, similar to the telephone service, in terms of queuing delay, the FIFO service strategy has the largest queuing delay, especially when there are a large number of secondary users in the system. The average queuing delay of the video service under the other three methods is relatively small; compared with the ToS value priority strategy and the ASREM-3 algorithm, the queuing delay of the video service is relatively large after using the method provided by the present application. When the number of SUs in the system is 70, the queuing delay reaches the maximum, close to 270 milliseconds, but is much smaller than the maximum queuing delay of 480 milliseconds required by the system. The reason is that the service rate of the video service queue is adjusted according to the predicted number of data packets. On the premise of meeting the queuing delay requirements of the video service, the service rate of the video service queue is reduced. Therefore, it can be considered that although the present application increases the average queuing delay of the video service to a certain extent, it still does not affect the real-time nature of the video service.

[0178] like Fig. 9 The figure shows the queuing delay duration of non-real-time services after using different methods. Fig. 9 (a) is a schematic diagram of non-real-time service queuing delay after using the FIFO queue management strategy, the ToS value priority strategy, the ASREM-3 queue management strategy and the method provided by the present application, Fig. 9 (b) is a schematic diagram of non-real-time service queuing delay after using the ToS value priority strategy, ASREM-3 queue management strategy and the method provided by this application;

[0179] from Fig. 9 It can be seen from the figure that when the number of SUs in the system is less than 40, due to the light network load, the average queuing delay of non-real-time services under various methods is small. When the number of SUs is greater than 40, the average queuing delay of non-real-time services begins to increase, among which the delay based on the FIFO queue management strategy is the smallest. This is because under the simulation parameters given in this embodiment, non-real-time services arrive at the C-eNodeB first and are served first. In addition, combined with Figure 7 and Figure 8 It can be seen that when the FIFO queue management strategy is used, the queuing delays of telephone services, video services and non-real-time services are all very large when the number of system SUs is large; when the method of the present application is used, the service rate of the real-time service queue is controlled under the premise of satisfying the queuing delay of the real-time service, so that the non-real-time service can be served as early as possible, thereby reducing the queuing delay of the non-real-time service and improving the service fairness of C-eNodeB to the real-time service and the non-real-time service.

[0180] In addition, this embodiment also compares the time complexity of the method provided by the present application with the time complexity of the other three queue management strategies, such as Fig.10 As shown, Fig.10(a) is a schematic diagram of a single running time of the FIFO queue management strategy, the ToS value priority strategy, the ASREM-3 queue management strategy and the method provided by the present application, Fig.10 (b) is a schematic diagram of a single running time of the ToS value priority strategy, the ASREM-3 queue management strategy and the method provided by the present application;

[0181] from Fig.10 It can be seen that when the number of SUs in the system is less than 60, the time complexity of various methods is not much different. When the number of SUs is greater than 60, the time complexity of the ToS value priority strategy increases dramatically. This is because the increase in the number of SUs leads to an increase in the number of data packets received by the C-eNodeB, an increase in the number of data packets in the telephone service queue and the video service queue, an increase in the complexity and number of sorting, and an increase in the overall time complexity. The method provided in the present application has a better time complexity than the FIFO queue management strategy except for individual number points. This is because the method provided in the present application can process data packets of different service types in a timely manner at each transmission time interval, thereby reducing the computational complexity.

[0182] based on Figure 6 to Figure 10 As can be seen from the schematic diagram, using the method provided by the present application on C-eNodeB to control the service rates of different service queues can effectively improve the communication service quality in a high-speed communication environment.

[0183] Based on the high-speed communication transmission rate control method based on Kalman filter prediction provided in the above embodiment, the embodiment of the present application also provides a high-speed communication transmission rate control device based on Kalman filter prediction, such as Fig.11 As shown, the device specifically includes:

[0184] The data packet number calculation module 10 is used to calculate the number of data packets in the real-time service queue at time (n+1)T based on the number of data packets in the real-time service queue at time nT using a Kalman filter algorithm;

[0185] The real-time service queue service rate calculation module 20 is used to calculate the occupancy ratio of the real-time service queue at time (n+1)T based on the number of data packets in the real-time service queue at time (n+1)T and the length of the real-time service queue, and use the service rate corresponding to the interval where the real-time service queue occupancy ratio at time (n+1)T is located as the service rate of the real-time service queue at time (n+1)T;

[0186] The non-real-time service queue service rate calculation module 30 is used to calculate the service rate of the non-real-time service queue at time (n+1)T based on the service rate of the real-time service queue at time (n+1)T, the size of a single data packet in the real-time service queue, the size of a single data packet in the non-real-time service queue, and the available data block size;

[0187] The transmission rate adjustment module 40 is used to adjust the transmission rate of each data packet in the real-time service queue and the non-real-time service queue based on the service rate of the real-time service queue at time (n+1)T and the service rate of the non-real-time service queue at time (n+1)T.

[0188] An embodiment of the present application also provides a computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, the steps of the above-mentioned high-speed communication transmission rate control method based on Kalman filter prediction are implemented.

[0189] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented in one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that include computer-usable program code.

[0190] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0191] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0192] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0193] Obviously, the above embodiments are merely examples for the purpose of clear explanation and are not intended to limit the implementation methods. For those skilled in the art, other different forms of changes or modifications can be made based on the above description. It is not necessary and impossible to list all the implementation methods here. The obvious changes or modifications derived therefrom are still within the scope of protection of the present invention.

Claims

1. A high-speed communication transmission rate control method based on Kalman filter prediction, characterized in that: include: Using the Kalman filter algorithm, the number of data packets in the real-time service queue at time nT is calculated based on the number of data packets in the real-time service queue at time (n+1)T; Calculate the occupancy ratio of the real-time service queue at time (n+1)T based on the number of data packets in the real-time service queue at time (n+1)T and the length of the real-time service queue, and use the service rate corresponding to the interval where the occupancy ratio of the real-time service queue at time (n+1)T is located as the service rate of the real-time service queue at time (n+1)T; The service rate of the non-real-time service queue at time (n+1)T is calculated based on the service rate of the real-time service queue at time (n+1)T, the size of a single data packet in the real-time service queue, the size of a single data packet in the non-real-time service queue, and the available data block size; wherein the service rate of the non-real-time service queue at time (n+1)T is calculated as follows: ， in, It represents the service rate of the non-real-time service queue at time (n+1)T. Indicates the available data block size, represents the service rate of the telephone service queue at time (n+1)T, Indicates the size of a single data packet in the telephone service queue. represents the service rate of the video service queue at time (n+1)T, Indicates the size of a single data packet in the video service queue. Indicates the size of a single data packet in the non-real-time service queue; Based on the service rate of the real-time service queue at time (n+1)T and the service rate of the non-real-time service queue at time (n+1)T, the transmission rate of each data packet in the real-time service queue and the non-real-time service queue is adjusted.

2. The high-speed communication transmission rate control method based on Kalman filter prediction according to claim 1 is characterized in that: The real-time service queue includes a telephone service queue and a video service queue.

3. The high-speed communication transmission rate control method based on Kalman filter prediction according to claim 2 is characterized in that: Based on the number of data packets in the real-time service queue at time (n+1)T and the length of the real-time service queue, calculating the occupancy ratio of the real-time service queue at time (n+1)T, and taking the service rate corresponding to the interval where the occupancy ratio of the real-time service queue at time (n+1)T is located as the service rate of the real-time service queue at time (n+1)T, including: Calculate the telephone service queue occupancy ratio at time (n+1)T based on the number of data packets in the telephone service queue at time (n+1)T and the length of the telephone service queue, and use the service rate corresponding to the interval where the telephone service queue occupancy ratio at time (n+1)T is located as the service rate of the telephone service queue at time (n+1)T; Based on the number of data packets in the video service queue at time (n+1)T and the length of the video service queue, the occupancy ratio of the video service queue at time (n+1)T is calculated, and the service rate corresponding to the interval in which the occupancy ratio of the video service queue at time (n+1)T is located is used as the service rate of the video service queue at time (n+1)T.

4. The high-speed communication transmission rate control method based on Kalman filter prediction according to claim 3 is characterized in that: The calculation formula for the telephone service queue occupancy ratio at time (n+1)T is: ， in, represents the telephone service queue occupancy ratio at (n+1)T time, represents the number of packets in the telephone service queue at time (n+1)T, Indicates the length of the telephone service queue; The service rate of the telephone service queue at time (n+1)T is: ， in, Indicates the service rate of the telephone service queue at time (n+1)T; , , ; The calculation formula for the video service queue occupancy ratio at time (n+1)T is: ， in, Indicates the video service queue occupancy ratio at (n+1)T time. Indicates the number of data packets in the video service queue at time (n+1)T, Indicates the length of the video service queue; The service rate of the video service queue at time (n+1)T is: ， in, represents the service rate of the video service queue at time (n+1)T, , , .

5. The high-speed communication transmission rate control method based on Kalman filter prediction according to claim 2 is characterized in that: Using the Kalman filter algorithm, based on the number of packets in the real-time service queue at time nT, and also based on the ToS value of each packet arriving at time nT, each packet is placed in the real-time service queue or the non-real-time service queue, which specifically includes: If the ToS value of the data packet is 0xff, the data packet is placed in the telephone service queue; If the ToS value of the data packet is 0xb0, the data packet is placed in the video service queue; If the ToS value of the data packet is 0x00, the data packet is placed in a non-real-time service queue.

6. The high-speed communication transmission rate control method based on Kalman filter prediction according to claim 1 is characterized in that: Calculating the number of data packets in the real-time service queue at time (n+1)T using the Kalman filter algorithm based on the number of data packets in the real-time service queue at time nT includes: Based on the number of data packets in the real-time service queue at time nT, a state variable a posteriori estimate of the number of data packets in the real-time service queue at time nT is obtained; based on the state variable a posteriori estimate of the number of data packets in the real-time service queue at time nT and the state transfer matrix, a state variable a priori estimate of the number of data packets in the real-time service queue at time (n+1)T is calculated; Based on the a posteriori error covariance matrix and the state transition matrix of the number of data packets in the real-time service queue at time nT, calculate the a priori error covariance matrix of the number of data packets in the real-time service queue at time (n+1)T; Based on the covariance matrix of the observation noise of the number of data packets in the real-time service queue at time nT and the a priori error covariance matrix of the number of data packets in the real-time service queue at time (n+1)T, the Kalman gain of the number of data packets in the real-time service queue at time (n+1)T is calculated; Calculate a posterior estimate of the state variable of the number of data packets in the real-time service queue at time (n+1)T based on the prior estimate of the state variable of the number of data packets in the real-time service queue at time (n+1)T, the Kalman gain of the number of data packets in the real-time service queue at time (n+1)T, and the observed value of the number of data packets in the real-time service queue at time (n+1)T; Calculate the a posteriori error covariance matrix of the number of data packets in the real-time service queue at time (n+1)T based on the a priori error covariance matrix of the number of data packets in the real-time service queue at time (n+1)T and the Kalman gain of the number of data packets in the real-time service queue at time (n+1)T; The number of data packets in the real-time service queue at time (n+1)T is obtained based on the state variable a posteriori estimate of the number of data packets in the real-time service queue at time (n+1)T.

7. The high-speed communication transmission rate control method based on Kalman filter prediction according to claim 6 is characterized in that: The calculation formula for the a posteriori estimate of the state variable of the number of packets in the real-time service queue at time nT is: ， in, Indicates the number of data packets in the real-time service queue at time nT. Represents the first-order differential of the number of packets in the real-time service queue at time nT; The calculation formula for the a priori estimate of the state variable of the number of data packets in the real-time service queue at time (n+1)T is: ， in, represents the prior estimate of the state variable of the number of packets in the real-time service queue at time (n+1)T, represents the state transfer matrix, The a posteriori estimate of the state variable representing the number of packets in the real-time service queue at time nT, represents the system noise; The calculation formula of the prior error covariance matrix of the number of data packets in the real-time service queue at time (n+1)T is: ， in, represents the a priori error covariance matrix of the number of packets in the real-time service queue at time (n+1)T, represents the a posteriori error covariance matrix of the number of packets in the real-time service queue at time nT, represents transpose, The variance of the Gaussian white noise representing the number of packets in the real-time service queue at time nT; The calculation formula of the Kalman gain of the number of data packets in the real-time service queue at time (n+1)T is: ， in, The Kalman gain representing the number of packets in the real-time service queue at time (n+1)T, represents the a priori error covariance matrix of the number of packets in the real-time service queue at time (n+1)T, represents the observation matrix, The covariance matrix representing the observation noise of the number of packets in the real-time service queue at time nT; The calculation formula for the a posteriori estimate of the state variable of the number of packets in the real-time service queue at time (n+1)T is: ， in, represents the a posteriori estimate of the state variable of the number of packets in the real-time service queue at time (n+1)T, represents the observed value of the number of data packets in the real-time service queue at time (n+1)T; The calculation formula of the a posteriori error covariance matrix of the number of data packets in the real-time service queue at time (n+1)T is: ， in, represents the a posteriori error covariance matrix of the number of packets in the real-time service queue at time (n+1)T, represents the identity matrix; The calculation formula for the number of data packets in the real-time service queue at time (n+1)T is: ， in, Indicates the number of data packets in the real-time service queue at time (n+1)T. Represents the first-order differential of the number of data packets in the real-time service queue at time (n+1)T.

8. A high-speed communication transmission rate control device based on Kalman filter prediction, characterized in that: include: A data packet number calculation module, used to calculate the number of data packets in the real-time service queue at time (n+1)T based on the number of data packets in the real-time service queue at time nT using a Kalman filter algorithm; a real-time service queue service rate calculation module, configured to calculate the occupancy ratio of the real-time service queue at time (n+1)T based on the number of data packets in the real-time service queue at time (n+1)T and the length of the real-time service queue, and use the service rate corresponding to the interval where the occupancy ratio of the real-time service queue at time (n+1)T is located as the service rate of the real-time service queue at time (n+1)T; The non-real-time service queue service rate calculation module is used to calculate the service rate of the non-real-time service queue at time (n+1)T based on the service rate of the real-time service queue at time (n+1)T, the size of a single data packet in the real-time service queue, the size of a single data packet in the non-real-time service queue, and the size of the available data block; wherein the calculation formula for the service rate of the non-real-time service queue at time (n+1)T is: ， in, It represents the service rate of the non-real-time service queue at time (n+1)T. Indicates the available data block size, represents the service rate of the telephone service queue at time (n+1)T, Indicates the size of a single data packet in the telephone service queue. represents the service rate of the video service queue at time (n+1)T, Indicates the size of a single data packet in the video service queue. Indicates the size of a single data packet in the non-real-time service queue; The transmission rate adjustment module is used to adjust the transmission rate of each data packet in the real-time service queue and the non-real-time service queue based on the service rate of the real-time service queue at the (n+1)T moment and the service rate of the non-real-time service queue at the (n+1)T moment.

9. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the high-speed communication transmission rate control method based on Kalman filter prediction described in any one of claims 1 to 7 are implemented.

Images