Dynamic access and backoff method and system based on load perception in large-scale MTC scenarios

Abstract

The present invention discloses a dynamic access and backoff method and system based on load perception. Under the premise of a known device activation model, the number of devices attempting to initiate a random access process in the current RAO and future RAOs (referred to as the number of activated devices) is predicted. Based on the load prediction result of the current RAO, the base station dynamically adjusts the ACB factor to control the probability of the activated device in the current RAO initiating a random access process, so that the number of devices initiating a random access process in the current RAO does not exceed the number of devices that maximize the network throughput. Based on the load prediction result of the future RAO, the scheme dynamically adjusts the distribution of backoff devices to control the number of activated devices in each RAO in the backoff window to be as close to the ideal load as possible, thereby ensuring the adaptability of the scheme in time-varying large-scale access scenarios, and increasing the probability of the activated device passing the ACB test, thereby reducing the number of times the device needs to listen to the ACB factor to access the network, thereby reducing the energy consumption of the device accessing the network without affecting the throughput.

Description

Technical Field

[0001] The present invention belongs to the field of access and transmission technology of cellular Internet of Things, and specifically relates to a dynamic access and backoff method and system based on load perception in a large-scale MTC scenario. Background Art

[0002] Machine Type Communication (MTC) is the main communication method for machine type devices in the Internet of Things. As a key supporting technology for the Internet of Things, massive machine type communication (mMTC) was identified as one of the three typical application scenarios of 5G by the International Telecommunication Union Radiocommunication Bureau (ITU-R) in 2015. In this scenario, a large number of devices can communicate with each other or remote servers without human intervention. The core of the mMTC scenario is to provide large-scale connections. Technologies such as Wi-Fi and ZigBee can be used to support connections, but mobile cellular networks are widely regarded as the core supporting networks for machine type communications due to their wide coverage, flexibility and high communication reliability. At present, the industry has proposed narrowband Internet of Things (NB-IoT) technology embedded in cellular systems and enhanced machine type communication (eMTC) technology to support future mMTC applications.

[0003] The access scheme in the traditional cellular network is a random access scheme based on competition and designed for human-to-human (H2H) communication habits. In this scheme, the base station randomly allocates limited channel resources to the device that initiates the access process, and the device randomly selects channel resources and competes for access. Since the mMTC scenario has the characteristics of massive connections, a large number of devices simultaneously initiating access requests to the base station will cause a large number of collisions between devices. The base station cannot successfully decode the device information, the probability of successful device communication is greatly reduced, and the network communication capacity is significantly reduced, which ultimately leads to serious system overload and congestion. Therefore, studying the access and transmission scheme for mMTC scenarios, solving the problem of network access congestion in mMTC scenarios, and realizing the control of network load, optimizing network performance to the greatest extent and maintaining performance stability are important topics in the study of mMTC.

[0004] In order to solve the problem of network congestion caused by the above-mentioned large-scale MTC access, domestic and foreign researchers have actively carried out research and proposed a series of solutions mainly from the aspects of access class barring (ACB), group paging (GP), resource allocation, etc. The 3GPP standardization organization introduced the ACB verification mechanism to control the access of devices. On the one hand, the ACB mechanism uses the ACB factor to control the probability of activated devices initiating access in the current random access opportunity (RAO); on the other hand, the ACB mechanism controls the devices that fail the ACB test to back off and redistribute the sudden large-scale access requests evenly in time, thereby alleviating instantaneous access collisions. This basic ACB mechanism can alleviate access collisions to a certain extent, but it cannot effectively control the number of devices initiating access in each RAO, and the degree of optimization of network throughput performance is limited. Existing research further proposes a dynamic ACB solution, in which the base station dynamically adjusts the ACB factor based on the load estimation value of the current random access opportunity (RAO) to control the probability of activated devices initiating the random access procedure (RA procedure). At the same time, the base station controls the activated devices that have not passed the ACB test and the activated devices that have access collisions after passing the ACB test to access again in the next RAO. The dynamic ACB solution can achieve near-optimal throughput in large-scale access scenarios, but requires devices to frequently listen to the ACB factor, and cannot initiate the actual random access process through the ACB test, and the energy efficiency of the device access process is extremely low. The basic idea of ​​GP is to group mMTC devices, rely on device-to-device (D2D) links within the group, initiate access requests to the base station as a whole, and reduce the scale of random access requests to reduce access collisions. The basic idea of ​​GP is to group mMTC devices, rely on device-to-device (D2D) links within the group, initiate access requests to the base station as a whole, and reduce the scale of random access requests to reduce access collisions. However, considering that machine type devices may have certain mobility and need to directly connect to the base station for communication, the GP method has certain limitations. The basic idea of ​​resource allocation to solve network congestion is to limit the resource blocks that various types of devices can initiate access to, and to control the number of devices that initiate access to each resource block. However, the resource allocation scheme is not adaptable enough to the device activation model. The device activation time is highly random, which affects the number of devices that actually initiate access in each resource block, resulting in reduced network performance.

[0005] Therefore, it is necessary to study a load control scheme to effectively control the access load of multiple RAOs in a time-varying large-scale random access scenario, so as to maximize the network throughput while reducing the energy consumption of devices accessing the network. Summary of the invention

[0006] In order to solve the problems existing in the prior art, the present invention provides a dynamic access and backoff method and system based on load perception in a large-scale MTC scenario. Compared with the existing research, the present invention realizes the effective control of the access load of each RAO during the random access process, and reduces the energy consumption of the device accessing the network while maximizing the network throughput under a limited number of retransmissions.

[0007] In order to achieve the above object, the technical solution adopted by the present invention is: a dynamic access and backoff method based on load perception in a large-scale MTC scenario, comprising the following steps:

[0008] S1. Establish a system model for dynamic backoff based on load perception in large-scale MTC scenarios;

[0009] S2, the number of access devices required to maximize the expected throughput of each RAO based on the system model, i.e., the ideal load;

[0010] S3. Predict current RAO and future T based on device activation model m The access load of each RAO is used to obtain the current RAO load forecast value and the future RAO load forecast value;

[0011] S4. Generate an ACB factor based on the load prediction value of the current RAO, and estimate the number of newly generated backoff devices after random access is completed;

[0012] S5. Based on the new activation load and backoff load of RAO, adjust the backoff window and backoff probability vector to minimize the mean square error between the number of backoff devices of each RAO in the backoff window and the ideal number of backoff devices. The optimization problem is constructed as follows:

[0013]

[0014]

[0015]

[0016] Constraint (1) represents the backoff window The possible range of T m is the upper limit of the backoff window length, T m It is artificially given and is affected by the load prediction capability of the base station and the maximum allowable access delay of the device. Constraints (2) and (3) ensure Form a probability space;

[0017] S6. Solve the optimization problem by using a hierarchical optimization method; obtain the optimal backoff probability vector and the optimal backoff window size Effectively control the number of backoff devices for each RAO within the backoff window.

[0018] In S1, consider an mMTC cell with N MTCDs. Assume that N devices are activated according to Beta distribution within T seconds. The probability density function of the device activation at time τ is:

[0019]

[0020] Wherein, B(α, β) represents a beta function with shape parameters α and β. At the beginning of each random access opportunity, the activated device must first perform an ACB test. The device that passes the ACB test initiates a random access request to the base station, otherwise the device enters the backoff process.

[0021] For the activated device of the tth RAO, the base station broadcasts the ACB factor to the device. To limit the probability of the activated device initiating a random access request to the base station, the activated device will generate a random number p between 0 and 1 and the ACB factor Comparison, p is not greater than The corresponding activated device passes the ACB test, otherwise the device enters the backoff process.

[0022] In S2, it is assumed that the only reason why the device failed to successfully complete the random access process is the collision of the preamble code. The access load of any RAO is m, and the number of available preamble codes is R. In this RAO, the probability of successful access of a device P is calculated. s|m for:

[0023]

[0024] Let S be the number of devices that have successfully accessed. When the access load is m, the expected value E(S|m) of the number of devices that have successfully accessed RAO is:

[0025]

[0026] m is a dynamically changeable integer, then

[0027]

[0028] When m = R or m = R-1, the RAO throughput expects E(S|m) to be the maximum, that is, the ideal load L Ideal =R or L Ideal =R-1, further set L Ideal =R.

[0029] In S3, before the current RAOt starts, the future T m The number of activated devices in each RAO is predicted; the activated devices in each RAO are divided into two categories: newly activated devices and backed-off devices. The number of newly activated devices is determined by the activation model, and the number of backed-off devices is determined by the backed-off devices and backed-off probability vectors generated by the past RAOs. The base station predicts the number of activated devices in each RAO based on the device activation model, the number of backed-off devices generated by the past RAOs, and the backed-off probability vectors. Predict current and future T m New activation load for RAO Backoff Load The specific steps include:

[0030] Assume that the current RAO is the tth RAO, then the new activation load prediction value of the base station for the t+ith RAO is for

[0031]

[0032] Where, T RAO represents the interval time between two adjacent RAOs, p(τ) represents the activation probability density of the device at time τ, N represents the total number of devices in the cell, and <·> represents the rounding operation;

[0033] Based on the backoff load prediction value generated by the base station in the (t-1)th RAO, the backoff probability vector, and the estimated number of backoff devices newly generated by the (t-1)th RAO and The backoff load prediction value of the base station in the tth RAO to the (t+i)th RAO is calculated by the backoff load prediction value of the previous RAO to the (t+i)th RAO. Add the expected number of newly transferred backoff devices Perform iterative updates.

[0034]

[0035] New activated load forecast value based on the (t+i)th RAO With backoff load prediction value Get the access load prediction value L of the RAO t (i) are as follows:

[0036]

[0037] In S4, after the base station completes the load prediction, based on the current RAO access load prediction value L t (0), Dynamically adjust the ACB factor And based on the preamble code collision situation in the current RAO, the number of back-off devices generated by the current RAO is estimated;

[0038] ACB Factor Control the probability of the current RAO activated device initiating access, so that the expected number of activated devices that actually initiate the random access process is as close to the ideal load R as possible. The calculation method is as follows:

[0039]

[0040] Where L t (0) is the access load prediction value of the base station for the current RAO;

[0041] Activate the device to The probability of passing the ACB test is that the activated device that passes the ACB test initiates a random access process to the base station. The base station estimates the access load based on the use of the preamble code. for:

[0042]

[0043] in,

[0044]

[0045] In the formula, S t is the number of preambles successfully accessed in the tth RAO, E t is the number of preamble codes not used by users in the tth RAO. Based on the access load estimation result, the number of backoff devices generated by the current RAO is further estimated.

[0046]

[0047] In S5, the ideal number of backoff devices C for the (t+i)th RAO t (i) To minimize the number of activated devices L t (i) The number of backoff devices with a difference from the ideal load R, i = 1, 2, ..., T m ,

[0048] For any and Estimated value of the number of backed-off devices in the (t+i)th RAO after back-off is completed for:

[0049]

[0050] The hierarchical optimization method is used to solve the optimization problem, which specifically includes the following steps:

[0051] S601, initialization of input parameters as well as Set the initial backoff window

[0052] S602, using the water injection algorithm to solve the Lagrange multiplier λ:

[0053]

[0054] S603, Solution The optimal number of backoff devices for each RAO is

[0055]

[0056] S604, Solution The number of backoff devices for each RAO is When the optimization problem objective function takes the value,

[0057]

[0058] S605, when When , return to step S602;

[0059] S606, Comparison The value of , and output the corresponding value when the objective function is minimum and

[0060] S607, according to And solve the optimal backoff probability vector

[0061]

[0062] At the same time, a dynamic backoff system based on load perception in large-scale MTC scenarios is provided, including a model building module, an ideal load calculation module, a load prediction module, an estimation module, a problem construction module and a solution module;

[0063] The model building module is used to establish a system model for dynamic backoff based on load perception in large-scale MTC scenarios;

[0064] The ideal load calculation module maximizes the number of access devices required for the expected throughput of each RAO based on the system model, that is, the ideal load;

[0065] The load prediction module predicts the current RAO and future T based on the equipment activation model. m The access load of each RAO is used to obtain the current RAO load forecast value and the future RAO load forecast value;

[0066] The estimation module generates an ACB factor based on the load prediction value of the current RAO, and estimates the number of newly generated backoff devices after random access is completed;

[0067] The problem construction module adjusts the backoff window and backoff probability vector based on the new activation load and backoff load of the RAO, minimizes the mean square error between the number of backoff devices of each RAO within the backoff window and the ideal number of backoff devices, and constructs the optimization problem as follows:

[0068]

[0069] Constraint (1) represents the backoff window The possible range of T m is the upper limit of the backoff window length, T m It is artificially given and is affected by the load prediction capability of the base station and the maximum allowable access delay of the device. Constraints (2) and (3) ensure Form a probability space;

[0070] The solution module uses a hierarchical optimization method to solve the optimization problem; the optimal backoff probability vector is obtained and the optimal backoff window size Effectively control the number of backoff devices for each RAO within the backoff window.

[0071] The present invention also provides a large-scale MTC scenario communication system, including a base station and a plurality of user terminals. The base station can execute the dynamic access and backoff method based on load perception in the large-scale MTC scenario described in the present invention when running.

[0072] Compared with the prior art, the present invention has at least the following beneficial effects:

[0073] Based on the calculation of network throughput and energy consumption of device access to the network, the number of activated devices performing ACB checks in each RAO is controlled to be equal to the number of available preamble codes in the RAO, which can maximize network throughput while reducing the energy consumption required for device access to the network. In order to achieve effective control of activated devices in each RAO, the present invention predicts the number of newly activated devices and the number of backoff devices in the current RAO and future RAOs based on the device activation model, the number of backoff devices generated by the past RAO, and the backoff probability vector. Based on the load prediction results of the current RAO, the present invention dynamically adjusts the ACB factor to control the probability of the activated device in the current RAO initiating a random access process, so that the number of devices initiating a random access process in the current RAO does not exceed the number of devices that maximize network throughput, that is, the ideal load. Based on the load prediction results of the future RAO, the present invention takes minimizing the mean square error of the number of backoff devices of each RAO in the backoff window and the ideal backoff load as the optimization goal, dynamically adjusts the backoff window size and the backoff probability vector, and controls the probability of the backoff devices generated in the current RAO backing off to each RAO in the backoff window. The present invention dynamically adjusts the access and backoff processes of devices based on load prediction results, and effectively controls the number of activated devices of each RAO, thereby maximizing network throughput while reducing device access energy consumption.

[0074] The technical solution of the present invention is further described in detail below through the accompanying drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0075] Figure 1 This is a system model diagram of a large-scale MTC network.

[0076] Figure 2 Schematic diagram of the principle of dynamic access and dynamic backoff method based on load perception.

[0077] Figure 3 This is a curve chart showing how the throughput of a large-scale MTC network changes with the total number of devices.

[0078] Figure 4 This is a curve chart showing how the energy consumption required for devices to access the network changes with the total number of devices in a large-scale MTC network. DETAILED DESCRIPTION

[0079] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0080] In the description of the present invention, it should be understood that the terms “include” and “comprises” indicate the presence of described features, wholes, steps, operations, elements and / or components, but do not exclude the presence or addition of one or more other features, wholes, steps, operations, elements, components and / or collections thereof.

[0081] The present invention proposes a dynamic access and backoff method based on load perception, and proposes a load prediction mechanism. Under the premise of a known device activation model, the number of devices attempting to initiate a random access process in the current RAO and future RAO (referred to as the number of activated devices) is predicted. Based on the load prediction result of the current RAO, the base station dynamically adjusts the ACB factor to control the probability of the activated device in the current RAO initiating a random access process, so that the number of devices initiating a random access process in the current RAO does not exceed the number of devices that maximize the network throughput, that is, the ideal load. Based on the load prediction result of the future RAO, the scheme dynamically adjusts the distribution of backoff devices to control the number of activated devices in each RAO in the backoff window to be as close to the ideal load as possible. The load prediction mechanism ensures the adaptability of the scheme in time-varying large-scale access scenarios. The dynamic access mechanism based on load perception ensures that the network throughput is as close to the theoretical maximum as possible. At the same time, the dynamic backoff mechanism based on load perception distributes the backoff devices as dispersedly as possible, and the probability of the activated device passing the ACB test is increased, thereby reducing the number of times the device needs to listen to the ACB factor to access the network, thereby reducing the energy consumption of the device accessing the network without affecting the throughput.

[0082] See also Figure 1 The present invention proposes a dynamic access and backoff method based on load perception in a large-scale MTC scenario, considering the following system model:

[0083] The present invention considers a large-scale cellular access network with N devices, wherein N machine type communication terminals are activated in time T according to beta distribution, that is, the activation probability of the device at time t is:

[0084]

[0085] Wherein, B(α, β) represents a beta function with shape parameters α and β, and the activation model base station is known.

[0086] In the mMTC network, there are dedicated time-frequency resource blocks to transmit random access requests of devices. This type of resource block is called a random access opportunity (RAO). RAOAppears periodically. The device that attempts to initiate a random access request in the RAO is called an activated device. The activated device first performs an ACB test. Among them, the activated device that passes the ACB test can initiate the random access process, and the activated device that fails the ACB test enters the backoff process. After a random backoff period, the nearest RAO is selected to initiate the random access request again. For the device that initiates the random access process, if the transmission preamble does not collide with other devices, the device can successfully access the network and transmit subsequent data packets; if the transmission collides with the preamble, the device will declare that the random access process has failed and enter the backoff process. After a random backoff period, the nearest RAO is selected to initiate the random access request again.

[0087] Specifically, at the tth RAO, the base station (BS) broadcasts the ACB factor to the device To limit the probability of a device initiating a random access request to the base station. For an activated device in the tth RAO, when the ACB factor is detected After that, a random number p between 0 and 1 is generated and compared with the ACB factor. If the activated device passes the ACB test, otherwise, the activated device enters the backoff process; after a random backoff period, the ACB test is performed again. After passing the ACB test, the activated device will initiate a contention-based random access process and send an access request to the base station. The contention-based random access process is mainly divided into four steps. The specific process is as follows Figure 1 As shown, UE stands for user equipment and eNB stands for base station. The four-way handshake of the contention-based random access process is as follows.

[0088] Step 1: The user equipment UE randomly selects an available preamble to initiate access. The number of available preambles in the contention-based random access process is R. The device randomly selects an available preamble as a random access request message (Msg1) and sends it to the base station through the Physical Random Access Channel (PRACH);

[0089] Step 2: The base station receives and detects the message in the PRACH channel, and generates a random access response message (RAR) based on the decoded information. The RAR contains the preamble code number detected by the base station, the uplink transmission timing advance (TA), the uplink channel resources corresponding to the preamble code, the cell radio network temporary identifier (C-RNTI), and the backoff indication parameter. The base station sends Msg2 to each cell device through the Physical Downlink Shared Channel (PDSCH). After the preamble transmission is completed, the activated device listens to the RAR message carrying the device transmission preamble number in the PDSCH channel. When the activated device hears the corresponding RAR message in the RAR window, step 3 is executed. Otherwise, the device declares that the random access has failed.

[0090] Step 3: UE schedules transmission according to the guidance of Msg2. If Msg2 includes the preamble information sent by the UE, Msg3 is transmitted in the uplink channel specified in Msg2. If not, it is backed off according to the instructions of Msg2. Msg3 contains the device identity, which is used by the eNB to register the source of the message;

[0091] Step 4: Contention resolution: If no collision occurs, the base station successfully decodes Msg3 to obtain the device identity information, and transmits an access success message (Msg4) through the downlink control channel (PDCCH).

[0092] Only the random access process failure caused by the preamble collision is considered. Specifically, when multiple devices transmit the same preamble at the same time, the base station cannot correctly decode the preamble. As a result, the device cannot detect the corresponding RAR message in the RAR window, declares the random access process failed, and enters the backoff process. After a random backoff period, the ACB test is performed again. For the activated devices that failed to access successfully, including the devices that failed the ACB test and the activated devices that passed the ACB test but had access collision, according to the backoff probability vector Given a probability distribution, each RAO in the backoff window is randomly selected to re-perform the ACB check and initiate a random access request.

[0093] The throughput of each RAO is set to the number of devices that successfully access the RAO. Assume that the only reason why the device failed to successfully complete the random access process is the collision of the preamble code. Let the access load of any RAO be m, and the number of available preamble codes be R. Then, when the preamble code selected by a device is different from the preamble codes selected by the remaining m-1 devices, the device successfully accesses. Therefore, in this RAO, the probability of a device successfully accessing is P s|m for

[0094]

[0095] Let S be the number of devices that successfully access. When the access load is m, the expected value of the number of devices that successfully access RAO, E(S|m), is:

[0096]

[0097] Since m is a dynamically changeable integer,

[0098]

[0099] Therefore, when m = R or m = R-1, the RAO throughput expects E(S|m) to be maximum, that is, the ideal load L Ideal =R or L Ideal =R-1. To simplify the analysis, we directly set L Ideal =R.

[0100] When the number of devices initiating random access processes for each RAO is controlled as close to the ideal load as possible, the network throughput can be maximized. Furthermore, the access load of each RAO, that is, the number of activated devices, can be controlled to be as close to the ideal load as possible, thereby increasing the probability of the activated device passing the ACB test, thereby reducing the number of times the device needs to listen to the ACB factor to access the network, thereby maximizing the network throughput while reducing the energy consumption of the device access process.

[0101] See also Figure 2 The present invention discloses a dynamic access and backoff method based on load perception in a large-scale MTC scenario, which is specifically divided into three steps: load prediction, generation and broadcasting of ACB factors and backoff parameters, and random access and dynamic backoff.

[0102] Step 11: Load perception. Before the current RAOt starts, the base station first performs load prediction. The base station uses the device activation model information, backoff parameters, and device access status to predict the current and future T m New activation load for RAO Backoff Load (in Represent the new activation load and backoff load of the (t+i)th RAO respectively), providing information basis for the base station to generate device access adjustment parameters in the next step;

[0103] Step 12: Generation and broadcast of ACB factors and backoff parameters. On the one hand, based on the load prediction results of the current RAO, the base station generates and broadcasts the ACB factor Control the probability of the activated device initiating the random access process to ensure the throughput of the current RAO; on the other hand, based on the load prediction results of the future RAO, the base station generates and broadcasts the backoff probability vector Controls the backoff device selection backoff window generated by the tth RAO The probability distribution of access retransmission for each RAO in the backoff window is calculated to minimize the gap between the access load of each RAO in the backoff window and the ideal load. The backoff devices include the devices whose current RAO has not passed the ACB test and the devices that have access collision after passing the ACB test. Represents the probability that the backoff device generated by the t-th RAO backs off to the (t+i)-th RAO.

[0104] Step 13: Device random access and dynamic backoff. The activated device of the tth RAO initiates access and completes backoff according to the ACB factor and backoff parameter broadcast by the base station. The activated device randomly generates a number p between 0 and 1, which is equal to the ACB factor broadcast by the base station. Compare, if p is not greater than The device performs a random access process. Otherwise, the device The backoff RAO is randomly selected with a given probability distribution. For an activated device that passes the ACB test, if the preamble code it transmits does not collide with other devices, the device successfully accesses the network. Otherwise, the device is The backoff RAO is randomly selected with a given probability distribution.

[0105] Next, we take the tth RAO as an example to conduct a specific analysis and design of the dynamic access and backoff method based on load prediction.

[0106] First, before the start of the tth RAO, the base station will compare the current RAO with the future RAO. m The number of activated devices in each RAO is predicted. The activated devices in each RAO can be divided into two categories: newly activated devices and backed-off devices. Newly activated devices are devices that have not yet undergone ACB verification, and backed-off devices are devices that choose RAO to initiate random access requests again after backing off. Therefore, the base station predicts the current RAO and the future T m The number of newly activated devices and the number of backed-off devices in a RAO. Assuming that the base station knows the activation model of the device, N machine type communication terminals are activated in time T according to the beta distribution, that is, the activation probability of the device at time τ is

[0107]

[0108] Where B(α, β) represents the beta function with shape parameters α and β. Then the estimated value of the number of newly activated devices in the (t+i)th RAO is for

[0109]

[0110] In addition, the backoff load in the (t+i)th RAO is affected by the number of backoff devices generated by the past RAO and the iteration of the backoff probability vector. Therefore, the backoff load prediction value in the (t+i)th RAO by the tth RAO can be iteratively updated by adding the backoff load prediction value of the previous RAO to the (t+i)th RAO plus the number of newly transferred backoff devices. Specifically, it is known that the backoff load prediction value of the (t+i)th RAO generated in the (t-1)th RAO is The resulting backoff load is The backoff probability vector generated by the base station is Therefore, the backoff load prediction value of the (t+i)th RAO time slot generated by the tth RAO is for:

[0111]

[0112] New activated load forecast value based on the (t+i)th RAO With backoff load prediction value The access load prediction value L of the RAO can be obtained t (i) are as follows:

[0113]

[0114] At the tth RAO, after the base station completes the load prediction, based on the current RAO access load prediction value L t (0), Dynamically adjust the ACB factor The probability of the activated device of the current RAO initiating access is controlled so that the expected number of activated devices that actually initiate the random access process is as close to the ideal load R as possible. Therefore, The calculation method is as follows:

[0115]

[0116] After the device completes the random access process, the base station detects that there is S in the tth RAO. t A successful access preamble, E t The number of preambles that are not used by users. When the number of devices initiating the random access process is M and the number of available preambles is R, the number of successful access and idle preambles is S respectively. t and E t The probability is as follows:

[0117]

[0118] Therefore, the maximum likelihood estimation method is used to estimate the number of devices that actually initiate random access in the tth RAO: Right now:

[0119]

[0120] Based on the access load estimation result, the base station further estimates the number of backoff devices generated by the current RAO

[0121]

[0122] After the number of backoff devices is estimated, the base station constructs an optimization problem, adjusts the backoff window and the backoff probability vector, and minimizes the square error between the number of backoff devices of each RAO within the backoff window and the ideal number of backoff devices. Specifically, the activated devices in each RAO can be divided into newly activated devices and backoff devices, where the number of newly activated devices is determined by the device activation model and is not regulated by the base station. Therefore, the present invention minimizes the difference between the number of activated devices in each RAO and the ideal load by controlling the number of backoff devices in each RAO. Define the ideal number of backoff devices C for the (t+i)th RAO t (i)(i=1,2,…,T m ) is to minimize the number of activated devices L t (i) The number of back-off devices with a difference from the ideal load R, then

[0123]

[0124] Design backoff probability vector Controls the probability of the backoff device generated by the tth RAO to re-initiate a random access request to each RAO within the backoff window. and The expected value of the number of backed-off devices in the (t+i)th RAO after backoff is completed for

[0125]

[0126] Based on the ideal number of backoff devices C t (i) Expected value of the number of backoff devices in each RAO after backoff parameter adjustment An optimization problem is proposed to minimize the ideal backoff load C of each RAO within the backoff window. t (i) Estimated number of backoff devices The square sum of the errors is the optimization target, and the backoff window size is adjusted dynamically. and the backoff probability vector The specific optimization problem is as follows:

[0127]

[0128] Among them, constraint (1) represents the backoff window The possible range of T mis the upper limit of the backoff window length, T m It is manually set and is affected by the load prediction capability of the base station and the maximum allowable access delay of the device. Constraints (2) and (3) ensure Form a probability space.

[0129] because When given, the optimization problem (OP) is is a convex function, and and One-to-one correspondence, so first convert the optimization problem (OP) into the following optimization problem:

[0130]

[0131]

[0132] The present invention regards this optimization problem as a two-level optimization. First, given Based on the Lagrange multiplier method, we can find the The optimal solution as follows:

[0133]

[0134] where λ is the solution of

[0135]

[0136] Then, use one-dimensional search to change Find the optimal backoff window size and the corresponding backoff probability vector Based on the optimal and The optimal backoff probability vector can be obtained as follows:

[0137]

[0138] According to the above process, That is, the optimal backoff probability vector that maximizes network throughput while reducing the energy consumption of devices accessing the network.

[0139] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all of the embodiments. The components of the embodiments of the present invention described and shown in the drawings here can usually be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present invention provided in the drawings is not intended to limit the scope of the claimed invention, but merely represents selected embodiments of the present invention. Based on the embodiments in the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0140] Numerical analysis is used to evaluate the throughput and energy consumption performance of the load-aware dynamic access and backoff method in large-scale scenarios. If not explicitly stated, the parameters of the simulation experiment are set as follows: MTCDs device activation model in the cell α = 3, β = 4, device activation time T A =10s; activation equipment transmission energy consumption P Tx =460mW, listening power consumption P Rx =80mW, idle power consumption is P Idle =3WW; RAO duration in the network is 1ms and occurs every 5ms, that is, T RAO =5ms; During random access, the RAR window length is W RAR =5ms, the length of the CR window is W CR =48ms, the maximum number of retransmissions of the device J = 9; the load prediction window size T m = 1000. To illustrate the superiority of the proposed scheme, we compare the RACH throughput and RACH energy consumption in the following schemes: dynamic ACB and uniform backoff method (abbreviated as D-ACB scheme), where the backoff window W B They are set to 1, 128, 512 or 1024 respectively, and the dynamic access and backoff method based on load perception (abbreviated as LAB scheme). Figure 3 and Figure 4 The relationship between the average throughput and average access energy consumption of D-ACB of the LAB solution and the number of MTCD devices N in the network is shown respectively. Figure 3 and Figure 4 It can be seen that the D-ACB solution with backoff window WB=1 has the highest throughput, but the highest access energy consumption. This is because the backoff MTCD needs to continuously monitor the updated ACB factor in each RAO and perform ACB checks without performing the actual random access process. BThe larger D-ACB scheme reduces the access energy consumption of the device by sacrificing the throughput performance. More importantly, the load-aware dynamic access and backoff method proposed in the present invention can achieve the lowest RACH energy consumption while maintaining almost the highest RACH throughput.

[0141] In summary, the present invention is applicable to large-scale MTC scenarios, and designs a dynamic access and backoff method based on load perception. Based on the theoretical analysis of maximizing throughput, when the scheme controls the number of activated devices in each RAO to be equal to the available preamble code in the RAO, it can achieve maximizing network throughput while reducing the energy consumption of devices accessing the network. The scheme is specifically divided into three steps: load perception, random access, and dynamic backoff. Based on the device activation model, the number of backoff devices generated by the past RAO, and the backoff probability vector, the base station predicts the current and future T m The number of activated devices in each RAO provides information basis for the base station to generate the ACB factor and backoff probability vector; based on the load prediction results of the current RAO, the base station generates and broadcasts the ACB factor Control the probability of the activated device initiating the random access process to ensure the throughput of the current RAO; based on the load prediction results of the future RAO, the base station generates and broadcasts the backoff probability vector Controls the backoff device selection backoff window generated by the tth RAO The probability distribution of access retransmission of each RAO in the backoff window is calculated to minimize the gap between the access load of each RAO in the backoff window and the ideal load. The dynamic access and backoff method based on load perception can reduce the energy consumption of device access to the network while maximizing network throughput.

[0142] The present invention also provides a large-scale MTC scenario communication system, including a base station and a plurality of user terminals. The base station can execute the dynamic access and backoff method based on load perception in the large-scale MTC scenario described in the present invention when running.

[0143] 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.

[0144] 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.

[0145] 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 including 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.

[0146] 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 in the computer or other programmable device. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0147] The above contents are only for explaining the technical idea of ​​the present invention and cannot be used to limit the protection scope of the present invention. Any changes made on the basis of the technical solution in accordance with the technical idea proposed by the present invention shall fall within the protection scope of the claims of the present invention.

Claims

1. A dynamic access and backoff method based on load perception in a large-scale MTC scenario, characterized in that: The following steps are involved: S1. Establish a system model for dynamic backoff based on load perception in large-scale MTC scenarios; S2, the number of access devices required to maximize the expected throughput of each RAO based on the system model, i.e., the ideal load; S3. Predict current RAO and future T based on device activation model m The access load of each RAO is calculated to obtain the load forecast value of the current RAO and the load forecast value of the future RAO. Before the current RAO starts, the future RAO is predicted. m The number of activated devices in each RAO is predicted; the activated devices in each RAO are divided into two categories: newly activated devices and backed-off devices. The number of newly activated devices is determined by the activation model, and the number of backed-off devices is determined by the backed-off devices and backed-off probability vectors generated by the past RAOs. The base station predicts the number of activated devices in each RAO based on the device activation model, the number of backed-off devices generated by the past RAOs, and the backed-off probability vectors. Predict current and future T m New activation load for RAO Backoff Load The specific steps include: Assume that the current RAO is the tth RAO, then the new activation load prediction value of the base station for the t+ith RAO is for Where, T RAO represents the interval time between two adjacent RAOs, p(τ) represents the activation probability density of the device at time τ, N represents the total number of devices in the cell, and <·> represents the rounding operation; Based on the backoff load prediction value generated by the base station in the (t-1)th RAO, the backoff probability vector, and the estimated number of backoff devices newly generated by the (t-1)th RAO i=1,2,...T m and The backoff load prediction value of the base station in the tth RAO to the (t+i)th RAO is calculated by the backoff load prediction value of the previous RAO to the (t+i)th RAO. Add the expected number of newly transferred backoff devices Perform iterative updates. New activated load forecast value based on the (t+i)th RAO With backoff load prediction value Get the access load prediction value L of the RAO t (i) are as follows: S4, based on the load prediction value of the current RAO, an ACB factor is generated, and the number of newly generated back-off devices is estimated after random access is completed; after the base station completes the load prediction, based on the current RAO access load prediction value L t (0), Dynamically adjust the ACB factor And based on the preamble code collision situation in the current RAO, the number of back-off devices generated by the current RAO is estimated; ACB Factor Control the probability of the current RAO activated device initiating access, so that the expected number of activated devices that actually initiate the random access process is as close to the ideal load R as possible. The calculation method is as follows: Where L t (0) is the access load prediction value of the base station for the current RAO; Activate the device to The probability of passing the ACB test is that the activated device that passes the ACB test initiates a random access process to the base station. The base station estimates the access load based on the use of the preamble code. for: in, In the formula, S t is the number of preambles successfully accessed in the tth RAO, E t is the number of preamble codes not used by users in the tth RAO. Based on the access load estimation result, the number of backoff devices generated by the current RAO is further estimated. S5. Based on the new activation load and backoff load of RAO, adjust the backoff window and backoff probability vector to minimize the mean square error between the number of backoff devices of each RAO in the backoff window and the ideal number of backoff devices. The optimization problem is constructed as follows: Constraint (1) represents the backoff window The possible range of T m is the upper limit of the backoff window length, T m It is artificially given and is affected by the load prediction capability of the base station and the maximum allowable access delay of the device. Constraints (2) and (3) ensure Form a probability space; S6. Solve the optimization problem by using a hierarchical optimization method; obtain the optimal backoff probability vector and the optimal backoff window size Effectively control the number of backoff devices for each RAO within the backoff window.

2. The method for dynamic access and backoff based on load perception in a large-scale MTC scenario according to claim 1, characterized in that: In S1, consider an mMTC cell with N MTCDs. Assume that N devices are activated according to Beta distribution within T seconds. The probability density function of the device activation at time τ is: Wherein, B(α, β) represents a beta function with shape parameters α and β. At the beginning of each random access opportunity, the activated device must first perform an ACB test. The device that passes the ACB test initiates a random access request to the base station, otherwise the device enters the backoff process.

3. The method for dynamic access and backoff based on load perception in a large-scale MTC scenario according to claim 2, characterized in that: For the activated device of the tth RAO, the base station broadcasts the ACB factor to the device. To limit the probability of the activated device initiating a random access request to the base station, the activated device will generate a random number p between 0 and 1 and the ACB factor Comparison, p is not greater than The corresponding activated device passes the ACB test, otherwise the device enters the backoff process.

4. The method for dynamic access and backoff based on load perception in a large-scale MTC scenario according to claim 1, characterized in that: In S2, it is assumed that the only reason why the device failed to successfully complete the random access process is the collision of the preamble code. The access load of any RAO is m, and the number of available preamble codes is R. In this RAO, the probability of successful access of a device P is calculated. s|m for: Let S be the number of devices that have successfully accessed. When the access load is m, the expected value E(S|m) of the number of devices that have successfully accessed RAO is: m is a dynamically changeable integer, then When m = R or m = R-1, the RAO throughput expects E(S|m) to be the maximum, that is, the ideal load L Ideal =R or L Ideal =R-1, further set L Ideal =R.

5. The method for dynamic access and backoff based on load perception in a large-scale MTC scenario according to claim 1, characterized in that: In S5, the ideal number of backoff devices C for the (t+i)th RAO t (i) To minimize the number of activated devices L t (i) The number of backoff devices with a difference from the ideal load R, i = 1, 2, ..., T m , For any and Estimated value of the number of backed-off devices in the (t+i)th RAO after back-off is completed for:

6. The method for dynamic access and backoff based on load perception in a large-scale MTC scenario according to claim 1, characterized in that: The hierarchical optimization method is used to solve the optimization problem, which specifically includes the following steps: S601, initialization of input parameters as well as Set the initial backoff window S602, using the water injection algorithm to solve the Lagrange multiplier λ: S603, Solution The optimal number of backoff devices for each RAO is S604, Solution The number of backoff devices for each RAO is When the optimization problem objective function takes the value, S605, when When , return to step S602; S606, Comparison The value of , and output the corresponding value when the objective function is minimum and S607, according to And solve the optimal backoff probability vector 7. A dynamic backoff system based on load perception in a large-scale MTC scenario, characterized in that: It includes model building module, ideal load calculation module, load prediction module, estimation module, problem building module and solution module; The model building module is used to establish a system model for dynamic backoff based on load perception in large-scale MTC scenarios; The ideal load calculation module maximizes the number of access devices required for the expected throughput of each RAO based on the system model, that is, the ideal load; The load prediction module predicts the current RAO and future T based on the equipment activation model. m The access load of each RAO is calculated to obtain the load forecast value of the current RAO and the load forecast value of the future RAO. Before the current RAO starts, the future RAO is predicted. m The number of activated devices in each RAO is predicted; the activated devices in each RAO are divided into two categories: newly activated devices and backed-off devices. The number of newly activated devices is determined by the activation model, and the number of backed-off devices is determined by the backed-off devices and backed-off probability vectors generated by the past RAOs. The base station predicts the number of activated devices in each RAO based on the device activation model, the number of backed-off devices generated by the past RAOs, and the backed-off probability vectors. Predict current and future T m New activation load for RAO Backoff Load The specific steps include: Assume that the current RAO is the tth RAO, then the new activation load prediction value of the base station for the t+ith RAO is for Where, T RAO represents the interval time between two adjacent RAOs, p(τ) represents the activation probability density of the device at time τ, N represents the total number of devices in the cell, and <·> represents the rounding operation; Based on the backoff load prediction value generated by the base station in the (t-1)th RAO, the backoff probability vector, and the estimated number of backoff devices newly generated by the (t-1)th RAO and The backoff load prediction value of the base station in the tth RAO to the (t+i)th RAO is calculated by the backoff load prediction value of the previous RAO to the (t+i)th RAO. Add the expected number of newly transferred backoff devices Perform iterative updates. New activated load forecast value based on the (t+i)th RAO With backoff load prediction value Get the access load prediction value L of the RAO t (i) are as follows: The estimation module generates the ACB factor based on the load prediction value of the current RAO and estimates the number of newly generated backoff devices after random access is completed; after the base station completes the load prediction, it accesses the load prediction value L based on the current RAO t (0), Dynamically adjust the ACB factor And based on the preamble code collision situation in the current RAO, the number of back-off devices generated by the current RAO is estimated; ACB Factor Control the probability of the current RAO activated device initiating access, so that the expected number of activated devices that actually initiate the random access process is as close to the ideal load R as possible. The calculation method is as follows: Where L t (0) is the access load prediction value of the base station for the current RAO; Activate the device to The probability of passing the ACB test is that the activated device that passes the ACB test initiates a random access process to the base station. The base station estimates the access load based on the use of the preamble code. for in, In the formula, S t is the number of preambles successfully accessed in the tth RAO, E t is the number of preamble codes not used by users in the tth RAO. Based on the access load estimation result, the number of backoff devices generated by the current RAO is further estimated. The problem construction module adjusts the backoff window and backoff probability vector based on the new activation load and backoff load of the RAO, minimizes the mean square error between the number of backoff devices of each RAO within the backoff window and the ideal number of backoff devices, and constructs the optimization problem as follows: Constraint (1) represents the backoff window The possible range of T m is the upper limit of the backoff window length, T m It is artificially given and is affected by the load prediction capability of the base station and the maximum allowable access delay of the device. Constraints (2) and (3) ensure Form a probability space; The solution module uses a hierarchical optimization method to solve the optimization problem; the optimal backoff probability vector is obtained and the optimal backoff window size Effectively control the number of backoff devices for each RAO within the backoff window.

8. A large-scale MTC scenario communication system, characterized in that: The invention comprises a base station and a user terminal, and the base station can execute the dynamic access and backoff method based on load perception in a large-scale MTC scenario described in any one of claims 1 to 6 when running.

Images