CN112512075A - Pilot frequency collision detection method, device and system - Google Patents

Pilot frequency collision detection method, device and system Download PDF

Info

Publication number
CN112512075A
CN112512075A CN202011342292.9A CN202011342292A CN112512075A CN 112512075 A CN112512075 A CN 112512075A CN 202011342292 A CN202011342292 A CN 202011342292A CN 112512075 A CN112512075 A CN 112512075A
Authority
CN
China
Prior art keywords
pilot
pilot frequency
collision detection
collision
frequency
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Granted
Application number
CN202011342292.9A
Other languages
Chinese (zh)
Other versions
CN112512075B (en
Inventor
丁杰
屈代明
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Huazhong University of Science and Technology
Original Assignee
Huazhong University of Science and Technology
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Huazhong University of Science and Technology filed Critical Huazhong University of Science and Technology
Priority to CN202011342292.9A priority Critical patent/CN112512075B/en
Publication of CN112512075A publication Critical patent/CN112512075A/en
Application granted granted Critical
Publication of CN112512075B publication Critical patent/CN112512075B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W24/00—Supervisory, monitoring or testing arrangements
    • H04W24/06—Testing, supervising or monitoring using simulated traffic
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L27/00—Modulated-carrier systems
    • H04L27/26—Systems using multi-frequency codes
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W24/00—Supervisory, monitoring or testing arrangements
    • H04W24/02—Arrangements for optimising operational condition
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
    • Y02D30/00—Reducing energy consumption in communication networks
    • Y02D30/70—Reducing energy consumption in communication networks in wireless communication networks

Abstract

The invention discloses a pilot frequency collision detection method, a device and a system, comprising the following steps: s1, after the distributed access point in the base station receives the pilot signal, the distributed access point carries out detection on any undetected pilot p in the preset pilot poollCalculating the received sum pilot plThe corresponding pilot signal strength; s2, obtaining pilot frequency p received by all access pointslForming a vector after the corresponding pilot signal intensity; s3, inputting the obtained vector into a pre-trained pilot frequency collision detection model to obtain a pilot frequency plNumber of collisions of
Figure DDA0002798886390000011
And S4, repeating S1-S3 until all the pilots in the preset pilot pool complete collision detection. The invention is oriented to MTC mass terminals randomlyWhen the access is carried out, the pilot frequency collision and the number of conflict terminals can be quickly detected only according to the received pilot frequency signals, so that the extra wireless resources and time overhead in the random access are avoided, the energy consumption and time delay of the access terminal in the random access are reduced, the base station can conveniently and effectively distribute the resources in time, and the random access performance of the terminal is improved.

Description

Pilot frequency collision detection method, device and system
Technical Field
The present invention belongs to the field of wireless communication technologies, and in particular, to a pilot collision detection method, apparatus, and system.
Background
As a main application scenario of the fifth generation mobile communication (5G), Machine Type Communication (MTC) has a huge development prospect. It is expected that by 2030, the number of global MTC terminal connections will approach 1 billion, with more than 200 billion in china. In a wireless communication system, an intelligent MTC terminal sends uplink data to a base station in a random access mode, wherein the random access is an indispensable component in the wireless communication system; it refers to a process in which a terminal not allocated with a radio resource requests a base station to allocate a radio resource to perform communication or directly transmit data, using a dedicated timeslot and spectrum resource. The existing random access methods related to the present invention are:
random Access mode of narrowband Internet of things (NB-IoT) communication system, see M.Chen, Y.Miao, Y.Hao and K.Hwang, "Narrow Band Internet of things," IEEE Access, vol.5, pp.20557-20577,2017. The NB-IoT communication system is optimally designed on the basis of LTE and is specially used for meeting the interconnection of massive MTC terminals. The main random access mode is a typical random access based on a grant. In the random access, a terminal having data transmission needs to first transmit a pilot to a base station as an access request, so that the base station can perform terminal identification and resource allocation. However, since MTC generally has a large amount of access requirements and pilot is a scarce resource, multiple MTC terminals may select the same pilot to transmit at the same time, thereby causing pilot collision. The pilot frequency collision is a main reason that the MTC mass terminals cannot obtain timely and reliable access service. And at the base station end, pilot frequency collision information is obtained in time, which is favorable for improving the system performance. Particularly, based on the pilot frequency collision information, the base station can flexibly adjust the resource allocation strategy so as to maximize the resource utilization rate under the pilot frequency collision and improve the access performance of the MTC terminal.
On the other hand, compared with the traditional centralized multi-antenna system, the distributed multi-antenna communication system can better meet the mass access requirement of MTC. A distributed multi-antenna system generally consists of an Access Point (AP) (antenna unit), a base station baseband processing unit (CPU), and a terminal. In a distributed multi-antenna system, APs are evenly distributed in a network covered by a base station and are geographically remotely located. And therefore has better coverage effect compared with a centralized multi-antenna system. A great deal of research confirms that the distributed multi-antenna system has higher system capacity, better spectrum efficiency, larger diversity gain and more energy saving compared with the centralized multi-antenna system. Therefore, the pilot frequency collision detection method, especially the pilot frequency collision detection method applied to the distributed multi-antenna wireless communication system, is significant.
The existing pilot frequency collision detection method needs a base station and a terminal to carry out information interaction for many times. These methods cannot perform effective pilot collision detection only depending on the received pilot signal, although it can be determined whether a certain pilot sequence is transmitted or not by the received signal energy after receiving the pilot signal. Therefore, after receiving the pilot signal, the base station and the terminal need to consume more wireless resources and time for information interaction and feedback to obtain pilot collision information and solve pilot collision. This results in low resource utilization of the system, high signaling overhead, high energy consumption and high access delay when the MTC terminal sends a short data packet.
Disclosure of Invention
Aiming at the defects or the improvement requirements of the prior art, the invention provides a pilot frequency collision detection method of a distributed multi-antenna wireless communication system, aiming at solving the technical problem that the wireless resource and time overhead is large because the pilot frequency collision and the number of collision terminals can not be detected only according to the received pilot frequency signals when the MTC-oriented massive terminals are randomly accessed.
In order to achieve the above object, in a first aspect, the present invention provides a pilot collision detection method, including the following steps:
s1, after receiving the pilot signal, the distributed access point in the base station carries out the collision detection on any pilot p which is not carried out in the preset pilot poollCalculating the received sum pilot plThe corresponding pilot signal strength;
s2, obtaining pilot frequency p received by all access pointslAfter the corresponding pilot signal strength, a vector with length M is formed and is marked as pilot plThe corresponding pilot signal strength characteristics;
s3, pilot plInputting the corresponding pilot signal intensity characteristics into a pre-trained pilot collision detection model to obtain a pilot plNumber of collisions of
Figure BDA0002798886370000031
S4, repeating the steps S1-S3 until all the pilot frequencies in the preset pilot frequency pool complete collision detection;
wherein, the preset pilot frequency pool phi ═ p1,p2,...,pL]L is the length of a preset pilot frequency pool; p is a radical oflL is more than or equal to 1 and less than or equal to L; different pilot frequencies in a preset pilot frequency pool are mutually orthogonal; m is the number of access points in the base station; the pilot frequency collision detection model is a model based on machine learning and is used for representing the mapping relation between the pilot frequency signal strength characteristic and the pilot frequency collision number; the number of pilot collision is the number of MTC terminals transmitting a certain pilot at the same time.
Further preferably, if
Figure BDA0002798886370000032
Then pilot plThe MTC terminal is not selected by any activated MTC terminal, and pilot frequency collision does not occur; if it is
Figure BDA0002798886370000033
Then pilot plOnly one activated MTC terminal selects the MTC terminal, and pilot frequency collision does not occur; if it is
Figure BDA0002798886370000034
Then pilot plA collision occurs.
Further preferably, for the mth access point APmReceived at the access point and pilot plThe corresponding pilot signal strengths are:
Figure BDA0002798886370000035
wherein M is more than or equal to 1 and less than or equal to M, and represents the norm of the matrix 2 (·)HFor conjugate transpose operations, YmAs an access point APmThe received pilot signal is a matrix of L × K, K being the access point APmThe number of the transmitting and receiving antennas configured above.
Further preferably, the pilot collision detection model includes: a feed-forward neural network and a maximum selector;
the feedforward neural network being arranged to depend on the pilot plThe corresponding pilot signal intensity characteristic determines the probability value with the pilot collision number of s-1 and outputs the probability value to the maximum selector; wherein S is 1,2, …, Smax+1，SmaxThe maximum number of pilot frequency collision possibly occurring in random access;
the maximum value selector is used for finding out the maximum probability value from the input of the feedforward neural network and taking the pilot frequency collision number corresponding to the maximum probability value as the pilot frequency plThe number of collisions of (2).
Further preferably, the loss function of the pilot collision detection model is a cross entropy loss function.
Further preferably, the method for training the pilot collision detection model includes:
randomly sampling a plurality of groups of pilot signal intensity characteristics and corresponding actual pilot collision numbers S in a distributed communication system to form a training set; and inputting the training set into a pilot frequency collision detection model, and minimizing a loss function of the pilot frequency collision detection model through a back propagation algorithm to obtain optimal model parameters so as to obtain the pilot frequency collision detection model which is pre-trained.
Further preferably, in step S2, the received pilot p at all the access points is obtainedlAfter the corresponding pilot signal intensity, sorting the pilot signal intensities from large to small to form a vector I with the length of M, which is marked as a pilot plThe corresponding pilot signal strength characteristic.
Further preferably, the pilot collision detection method provided by the first aspect of the present invention is applied to a distributed multi-antenna wireless communication system.
In a second aspect, the present invention provides a base station, which includes M access points and a base station baseband processing unit; each access point is uniformly distributed in a network covered by the base station and is connected with the base band processing unit of the base station through a forward return link;
the access point is used for receiving the pilot signal and then carrying out collision detection on any pilot p which is not subjected to collision detection in the preset pilot poollCalculating the received sum pilot plThe corresponding pilot signal intensity is sent to the base station baseband processing unit;
the base station baseband processing unit is used for receiving the pilot signal strength sent by each access point and waiting to acquire the pilot signals p received by all the access pointslAfter the corresponding pilot signal strength, a vector with length M is formed and is marked as pilot plThe corresponding pilot signal strength characteristics; to pilot plInputting the corresponding pilot signal intensity characteristics into a pre-trained pilot collision detection model to obtain a pilot plNumber of collisions of
Figure BDA0002798886370000041
Wherein, the preset pilot frequency pool phi ═ p1,p2,...,pL]L is the length of a preset pilot frequency pool; p is a radical oflL is more than or equal to 1 and less than or equal to L; different pilot frequencies in a preset pilot frequency pool are mutually orthogonal; the pilot frequency collision detection model is a model based on machine learning and is used for representing the mapping relation between the pilot frequency signal strength characteristic and the pilot frequency collision number; the number of pilot collision is the number of MTC terminals transmitting a certain pilot at the same time.
In a third aspect, the present invention provides a pilot collision detection system, including: an MTC terminal and a base station provided by the second aspect of the invention;
the MTC terminal is used for randomly selecting a pilot frequency from a preset pilot frequency pool phi in a random access time slot when data transmission is required, and transmitting the pilot frequency to the base station;
wherein, the preset pilot frequency pool phi ═ p1,p2,...,pL]L is the length of a preset pilot frequency pool; different pilots in the preset pilot pool are mutually orthogonal.
In general, compared with the prior art, the above technical solution contemplated by the present invention can achieve the following beneficial effects:
1. the invention provides a pilot frequency collision detection method, a device and a system, which are used for detecting a pilot frequency p to be detected received by all access pointslAnd a vector formed by corresponding pilot signal strength is used as a pilot signal strength characteristic, and a mapping relation between the pilot signal strength characteristic and the pilot collision number is represented by a pilot collision detection model. In detecting pilot plIn the case of collision number of (2), the pilot frequency plInputting the corresponding pilot signal intensity characteristics into the pilot collision detection model to obtain a pilot plNumber of collisions of
Figure BDA0002798886370000051
Through the process, when the MTC mass terminal is randomly accessed, the invention can quickly detect the pilot frequency collision and the number of the conflict terminals according to the received pilot frequency signal, thereby avoiding the extra wireless resource and time overhead in the random access, reducing the energy consumption and time delay of the access terminal in the random access, facilitating the base station to carry out timely and effective resource allocation and improving the random access performance of the terminal.
2. The pilot frequency collision detection method provided by the invention is based on a distributed multi-antenna system, and the detection of the pilot frequency collision is realized by utilizing the signal information of two dimensions, namely the pilot frequency signal strength received by each access point and the distributed access point position, so that more accurate pilot frequency collision number can be obtained.
3. The pilot frequency collision detection method provided by the invention can complete detection by depending on a machine learning algorithm, does not need to pass through the traditional complex mathematical modeling, only needs to establish a pilot frequency collision detection model through the neural network self-adaptive learning, and can rapidly and accurately complete the detection of the number of pilot frequency collisions.
Drawings
Fig. 1 is a flowchart of a pilot collision detection method provided in embodiment 1 of the present invention;
fig. 2 is a flow chart of an exemplary neural network provided in embodiment 1 of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention is described in further detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention. In addition, the technical features involved in the embodiments of the present invention described below may be combined with each other as long as they do not conflict with each other.
Examples 1,
A pilot frequency collision detection method is applied to a distributed multi-antenna wireless communication system (comprising a base station and an MTC terminal), wherein the base station consists of M Access Points (APs) and 1 base station baseband processing unit (CPU), and the AP serial numbers are 1 to M in sequence; each AP is provided with K receiving and transmitting antennas and is uniformly distributed in a network covered by the base station; all APs are connected to the base station CPU via a fronthaul link. The base station CPU can collect the received signals on all APs for signal centralized processing. In addition, in a network covered by the distributed AP, there are MTC terminals distributed randomly; in each random access time slot, each MTC terminal can access a network with a certain activation probability; the probability of active access is typically much less than 1. In the random access process, an activated MTC terminal (which has a data transmission requirement) may randomly select a pilot from a preset orthogonal pilot pool to transmit.
Specifically, as shown in fig. 1, the pilot collision detection method includes the following steps:
s1, after receiving the pilot signal, the distributed access point in the base station carries out the collision detection on any pilot p which is not carried out in the preset pilot poollCalculating the received sum pilot plThe corresponding pilot signal strength (denoted as received pilot signal strength); wherein, the preset pilot frequency pool phi ═ p1,p2,...,pL]L is the length of a preset pilot frequency pool; p is a radical oflIs the first pilot frequency which is a vector with the length of L, and L is more than or equal to 1 and less than or equal to L; different pilot frequencies in a preset pilot frequency pool are mutually orthogonal;
specifically, each distributed AP receives a pilot signalAfter the number, for pilot plCalculating the intensity of the pilot signal corresponding to the pilot signal; and sends the resulting pilot signal strength to the base station CPU.
S2, obtaining pilot frequency p received by all access pointslAfter the corresponding pilot signal strength, a vector with length M is formed and is marked as pilot plThe corresponding pilot signal strength characteristics; wherein M is the number of access points in the base station;
in particular, for the mth access point APmReceived at the access point and pilot plThe corresponding pilot signal strengths are:
Figure BDA0002798886370000071
wherein M is more than or equal to 1 and less than or equal to M, and represents the norm of the matrix 2 (·)HFor conjugate transpose operations, YmAs an access point APmThe received pilot signal is a matrix of L × K, K being the access point APmThe number of the transmitting and receiving antennas configured above.
Preferably, the base station CPU acquires the received pilot p at all access pointslAfter the corresponding pilot signal intensity, sorting the pilot signal intensities from large to small to form a vector I with the length of M, which is marked as a pilot plAnd the corresponding pilot signal strength characteristic is used for reducing the calculation complexity of a subsequent pilot collision detection model.
S3, pilot plInputting the corresponding pilot signal intensity characteristics into a pre-trained pilot collision detection model to obtain a pilot plNumber of collisions of
Figure BDA0002798886370000072
In particular, if
Figure BDA0002798886370000073
Then pilot plThe MTC terminal is not selected by any activated MTC terminal, and pilot frequency collision does not occur; if it is
Figure BDA0002798886370000074
Then pilot plOnly one activated MTC terminal selects the MTC terminal, and pilot frequency collision does not occur; if it is
Figure BDA0002798886370000081
Then pilot plA collision occurs.
The pilot frequency collision detection model is a model based on machine learning and used for representing the mapping relation between the pilot frequency signal intensity characteristic and the pilot frequency collision number; the number of pilot collision is the number of MTC terminals transmitting a certain pilot at the same time. It should be noted that, the pilot signal strength characteristic includes not only the pilot signal strength information received at each AP but also the AP distributed location information, and a more accurate pilot collision number can be obtained by using the information of the two dimensions.
Preferably, the pilot collision detection model comprises: a feed-forward neural network and a maximum selector. Wherein the feedforward neural network is used for deriving the pilot plThe corresponding pilot signal intensity characteristic determines the probability value with the pilot collision number of s-1 and outputs the probability value to the maximum selector; wherein S is 1,2, …, Smax+1，SmaxIs the maximum number of pilot collisions that can occur in random access. The maximum value selector is used for finding out the maximum probability value from the input of the feedforward neural network and taking the pilot frequency collision number corresponding to the maximum probability value as the pilot frequency plThe number of collision and impact.
Further, a typical neural network flow diagram is shown in FIG. 2; it consists of an input layer (layer 0), J hidden layers (J is more than or equal to 1) (layers 1 to J) and an output layer (layer J + 1); each layer is composed of NjJ is more than or equal to 0 and less than or equal to J + 1. Wherein J and NjThe size of the system is set according to system parameters and requirements; in neural networks, the pilot collision detection function is implemented by training known signal samples. In the present invention, N0＝M；I0Input for layer 0, which consists of M received pilot signal strengths; preferably, in order to reduce the computational complexity of the neural network, the base station may rank the M received pilot signal strengthsForm I0。
Based on I0The construction of the hidden layer is an iterative process; specifically, it can be written as:
Ij＝f(Wj-1Ij-1+bj-1),j＝1,2,...,J
wherein IjThe output of the j layer of the neural network is NjA vector of lengths; wj-1Weighting parameter matrix for j-1 layer of neural network with size of Nj×Nj-1；bj-1For the neural network layer j-1 paranoia parameter, is NjA vector of lengths; f (-) is an activation function of the neural network, and has the function of adding some non-linear factors to the neural network, so that the neural network can better solve the more complex problem; preferably, the activation function f (-) may be a sigmoid function.
In the present invention, IJ+1Is the output of layer J +1, which is an NJ+1＝SmaxA +1 length vector; wherein S ismaxThe maximum number of pilot frequency collision possibly occurring in random access; in particular, IJ+1Can be written as:
IJ+1＝softmax(WJIJ+bJ)
wherein softmax (·) is a softmax function; thus, IJ+1Middle element IJ+1,s(1≤s≤SmaxThe size of +1) represents the probability value that the number of pilot collisions estimated by the neural network is s-1.
Neural network based output IJ+1Maximum value selector from IJ+1Finding out the pilot frequency collision number corresponding to the element with the maximum probability value from all the elements, and using the pilot frequency collision number as the output of the pilot frequency collision detection module
Figure BDA0002798886370000091
Further, the training method of the pilot collision detection model includes: randomly sampling a plurality of groups of pilot signal intensity characteristics and corresponding actual pilot collision numbers S in a distributed communication system to form a training set; and inputting the training set into the pilot frequency collision detection model in batches, and minimizing a loss function of the pilot frequency collision detection model through a back propagation algorithm to obtain optimal model parameters so as to obtain the pilot frequency collision detection model which is pre-trained.
To further illustrate the pilot collision detection method proposed by the present invention, the following detailed description is made with reference to specific embodiments:
this embodiment is to detect pilot p1The pilot collision of (2) is explained as an example. In the distributed antenna communication system of the present embodiment, M is 8, and K is 1; the APs are randomly distributed in the network coverage range; in random access, the maximum number of possible pilot collisions Smax2; in the pilot collision detection model, the neural network comprises three layers, wherein the number of ganglion points of an input layer (layer 0) is N0The number of hidden layers J equals 1 and the number of ganglion points of this layer (layer 1) is N, 81The number of ganglion points in the output layer (layer 2) is N, 162＝3。
For any APmM is more than or equal to 1 and less than or equal to 8, and the received pilot signal is marked as ymIs an L × 1 vector; pilot p1The corresponding received pilot signal strength is expressed as:
Figure BDA0002798886370000101
after receiving 8 signal strengths, the base station CPU forms a signal strength vector I ═ E11,E21,...,E81](ii) a In order to reduce the computation complexity of the neural network under the condition of ensuring certain detection accuracy, the elements in the I can be preferably sorted from large to small to generate a new signal intensity vector I0Is specifically shown as I0＝sortD(I) (ii) a Wherein sortD(. cndot.) is a descending sort function.
Based on I0The constructed neural network output is expressed as:
I2＝softmax(W1I1+b1)
wherein I1＝sigmoid(W0I0+b0)；W0And W1Weighting parameters for layer 1 and layer 2 of the neural network, respectively16 × 8 and 8 × 3 matrices; b0And b1The bias term parameters of the neural network layer 1 and layer 2 are vectors with the length of 16 and 3 respectively; the optimal values of these network parameters are determined by the neural network offline sample training; specifically, the training set is input into the neural network, and the optimal network parameters can be obtained by minimizing a cross entropy loss function through a back propagation algorithm.
In this embodiment, I2＝[I2,1,I2,2,I2,3](ii) a Wherein I2,1+I2,2+I2,31, element I2,1Pilot p representing a neural network estimate1A probability value that the number of collisions is 0; element I2,2Pilot p representing a neural network estimate1A probability value that the number of collisions is 1; element I2,3Pilot p representing a neural network estimate1The number of collisions is a probability value of 2. Based on I2The maximum selector finds out the pilot frequency collision number corresponding to the maximum probability value element from all three elements, namely the detected pilot frequency p1The number of collisions of (2).
Examples 2,
A base station comprises M access points and a base station baseband processing unit; each access point is uniformly distributed in a network covered by the base station and is connected with the base band processing unit of the base station through a forward return link;
the access point is used for receiving the pilot signal and then carrying out collision detection on any pilot p which is not subjected to collision detection in the preset pilot poollCalculating the received sum pilot plThe corresponding pilot signal intensity is sent to the base station baseband processing unit;
the base station baseband processing unit is used for receiving the pilot signal strength sent by each access point and waiting to acquire the pilot signals p received by all the access pointslAfter the corresponding pilot signal strength, a vector with length M is formed and is marked as pilot plThe corresponding pilot signal strength characteristics; to pilot plInputting the corresponding pilot signal intensity characteristics into a pre-trained pilot collision detection model to obtain a pilot plNumber of collisions of
Figure BDA0002798886370000111
Wherein, the preset pilot frequency pool phi ═ p1,p2,...,pL]L is the length of a preset pilot frequency pool; p is a radical oflL is more than or equal to 1 and less than or equal to L; different pilot frequencies in a preset pilot frequency pool are mutually orthogonal; the pilot frequency collision detection model is a model based on machine learning and is used for representing the mapping relation between the pilot frequency signal strength characteristic and the pilot frequency collision number; the number of pilot collision is the number of MTC terminals transmitting a certain pilot at the same time.
The related technical scheme is the same as embodiment 1, and is not described herein.
Examples 3,
A pilot collision detection system, comprising: an MTC terminal and a base station provided in embodiment 2 of the present invention;
the MTC terminal is used for randomly selecting a pilot frequency from a preset pilot frequency pool phi in a random access time slot when data transmission is required, and transmitting the pilot frequency to the base station;
wherein, the preset pilot frequency pool phi ═ p1,p2,...,pL]L is the length of a preset pilot frequency pool; different pilots in the preset pilot pool are mutually orthogonal.
The related technical scheme is the same as embodiment 1, and is not described herein.
It will be understood by those skilled in the art that the foregoing is only a preferred embodiment of the present invention, and is not intended to limit the invention, and that any modification, equivalent replacement, or improvement made within the spirit and principle of the present invention should be included in the scope of the present invention.

Claims (10)

1. A pilot collision detection method is characterized by comprising the following steps:
s1, after receiving the pilot signal, the distributed access point in the base station carries out the collision detection on any pilot p which is not carried out in the preset pilot poollCalculating the received pilot frequency plThe corresponding pilot signal strength;
s2, the sum of the received sum and the received sum to be acquired at all the access pointsThe pilot frequency plAfter the corresponding pilot signal strength, a vector with length M is formed and is marked as pilot plThe corresponding pilot signal strength characteristics;
s3, the pilot frequency plInputting the corresponding pilot signal intensity characteristics into a pre-trained pilot collision detection model to obtain a pilot plNumber of collisions of
Figure FDA0002798886360000011
S4, repeating the steps S1-S3 until all the pilot frequencies in the preset pilot frequency pool complete collision detection;
wherein the preset pilot pool phi ═ p1,p2,...,pL]And L is the length of the preset pilot frequency pool; p is a radical oflL is more than or equal to 1 and less than or equal to L; different pilot frequencies in the preset pilot frequency pool are mutually orthogonal; m is the number of access points in the base station; the pilot frequency collision detection model is a model based on machine learning and is used for representing the mapping relation between the pilot frequency signal strength characteristic and the pilot frequency collision number; the number of pilot frequency collisions is the number of MTC terminals simultaneously transmitting a certain pilot frequency.
2. The pilot collision detection method according to claim 1, wherein the collision detection is performed if
Figure FDA0002798886360000012
Then the pilot plThe MTC terminal is not selected by any activated MTC terminal, and pilot frequency collision does not occur; if it is
Figure FDA0002798886360000013
Then the pilot plOnly one activated MTC terminal selects the MTC terminal, and pilot frequency collision does not occur; if it is
Figure FDA0002798886360000014
Then the pilot plA collision occurs.
3. According toThe pilot collision detection method of claim 1, characterized in that for the mth access point APmReceived at the access point with the pilot plThe corresponding pilot signal strengths are:
Figure FDA0002798886360000021
wherein M is more than or equal to 1 and less than or equal to M, and represents the norm of the matrix 2 (·)HFor conjugate transpose operations, YmAs an access point APmThe received pilot signal is a matrix of L × K, K being the access point APmThe number of the transmitting and receiving antennas configured above.
4. The pilot collision detection method according to claim 1, wherein the pilot collision detection model comprises: a feed-forward neural network and a maximum selector;
the feedforward neural network is used for according to the pilot frequency plThe corresponding pilot signal intensity characteristic determines the probability value with the pilot collision number of s-1 and outputs the probability value to the maximum selector; wherein S is 1,2, …, Smax+1，SmaxThe maximum number of pilot frequency collision possibly occurring in random access;
the maximum value selector is used for finding out a maximum probability value from the input of the feedforward neural network and taking the pilot frequency collision number corresponding to the maximum probability value as a pilot frequency plThe number of collisions of (2).
5. The pilot collision detection method according to claim 1, wherein the loss function of the pilot collision detection model is a cross entropy loss function.
6. The method of claim 1, wherein the method for training the pilot collision detection model comprises:
randomly sampling a plurality of groups of pilot signal intensity characteristics and corresponding actual pilot collision numbers S in a distributed communication system to form a training set; and inputting the training set into a pilot frequency collision detection model, and minimizing a loss function of the pilot frequency collision detection model through a back propagation algorithm to obtain optimal model parameters so as to obtain the pilot frequency collision detection model which is pre-trained.
7. The method as claimed in any one of claims 1 to 6, wherein in step S2, the pilot collision detection method is obtained by obtaining the collision information p with the pilot received at all access pointslAfter the corresponding pilot signal intensity, sorting the pilot signal intensities from large to small to form a vector I with the length of M, which is marked as a pilot plThe corresponding pilot signal strength characteristic.
8. The pilot collision detection method according to claim 1, applied to a distributed multi-antenna wireless communication system.
9. A base station comprising M access points and a base station baseband processing unit; the access points are uniformly distributed in a network covered by the base station and are connected with the base band processing unit of the base station through a forward return link;
the access point is used for receiving the pilot signals and then carrying out collision detection on any pilot p which is not subjected to collision detection in the preset pilot poollCalculating the received pilot frequency plThe corresponding pilot signal intensity is sent to the base station baseband processing unit;
the base station baseband processing unit is used for receiving the pilot signal strength sent by each access point and obtaining the pilot p received by all the access pointslAfter the corresponding pilot signal strength, a vector with length M is formed and is marked as pilot plThe corresponding pilot signal strength characteristics; the pilot frequency plInputting the corresponding pilot signal intensity characteristics into a pre-trained pilot collision detection model to obtain a pilot plNumber of collisions of
Figure FDA0002798886360000031
Wherein the preset pilot pool phi ═ p1,p2,...,pL]And L is the length of the preset pilot frequency pool; p is a radical oflL is more than or equal to 1 and less than or equal to L; different pilot frequencies in the preset pilot frequency pool are mutually orthogonal; the pilot frequency collision detection model is a model based on machine learning and is used for representing the mapping relation between the pilot frequency signal strength characteristic and the pilot frequency collision number; the number of pilot frequency collisions is the number of MTC terminals simultaneously transmitting a certain pilot frequency.
10. A pilot collision detection system, comprising: an MTC terminal and a base station according to claim 9;
the MTC terminal is used for randomly selecting a pilot frequency p from a preset pilot frequency pool phi in a random access time slot when data transmission is requiredlAnd sending to the base station;
wherein the preset pilot pool phi ═ p1,p2,...,pL]And L is the length of the preset pilot frequency pool; and different pilot frequencies in the preset pilot frequency pool are mutually orthogonal.
CN202011342292.9A 2020-11-26 2020-11-26 Pilot frequency collision detection method, device and system Active CN112512075B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202011342292.9A CN112512075B (en) 2020-11-26 2020-11-26 Pilot frequency collision detection method, device and system

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202011342292.9A CN112512075B (en) 2020-11-26 2020-11-26 Pilot frequency collision detection method, device and system

Publications (2)

Publication Number Publication Date
CN112512075A true CN112512075A (en) 2021-03-16
CN112512075B CN112512075B (en) 2022-02-15

Family

ID=74966203

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202011342292.9A Active CN112512075B (en) 2020-11-26 2020-11-26 Pilot frequency collision detection method, device and system

Country Status (1)

Country Link
CN (1) CN112512075B (en)

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JPWO2007125910A1 (en) * 2006-04-25 2009-09-10 パナソニック株式会社 Wireless communication terminal device, wireless communication base station device, and wireless communication method
CN102148675A (en) * 2011-05-06 2011-08-10 北京邮电大学 Random access protocol based on hybrid automatic retransmission
CN110139392A (en) * 2019-05-06 2019-08-16 安徽继远软件有限公司 LTE electric power wireless private network random access channel multiple conflict detection method
CN110611627A (en) * 2018-06-15 2019-12-24 维沃移动通信有限公司 Signal detection method and receiving end
CN111031611A (en) * 2020-01-03 2020-04-17 浙江工业大学 Pilot random access method based on non-authorization
CN111865845A (en) * 2020-06-04 2020-10-30 国网江西省电力有限公司供电服务管理中心 MIMO user detection and channel estimation device and method

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JPWO2007125910A1 (en) * 2006-04-25 2009-09-10 パナソニック株式会社 Wireless communication terminal device, wireless communication base station device, and wireless communication method
CN102148675A (en) * 2011-05-06 2011-08-10 北京邮电大学 Random access protocol based on hybrid automatic retransmission
CN110611627A (en) * 2018-06-15 2019-12-24 维沃移动通信有限公司 Signal detection method and receiving end
CN110139392A (en) * 2019-05-06 2019-08-16 安徽继远软件有限公司 LTE electric power wireless private network random access channel multiple conflict detection method
CN111031611A (en) * 2020-01-03 2020-04-17 浙江工业大学 Pilot random access method based on non-authorization
CN111865845A (en) * 2020-06-04 2020-10-30 国网江西省电力有限公司供电服务管理中心 MIMO user detection and channel estimation device and method

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
钟思洁: ""基于机器学习的MTC网络接入策略"", 《中国优秀博硕士学位论文全文数据库(硕士)，信息科技辑（月刊）》 *

Also Published As

Publication number Publication date
CN112512075B (en) 2022-02-15

Similar Documents

Publication Publication Date Title
CN110401964B (en) Power control method based on deep learning for user-oriented center network
CN109005551B (en) Multi-user NOMA downlink power distribution method of non-ideal channel state information
CN113014295B (en) Uplink joint receiving method for large-scale de-cellular MIMO system
Yoshida et al. MAB-based client selection for federated learning with uncertain resources in mobile networks
CN111698045B (en) Energy efficiency power distribution method in millimeter wave communication system based on non-orthogonal multiple access
CN100547945C (en) A kind of method of down beam shaping and device
CN111181612A (en) Cooperative beamforming method of large-scale MIMO system
CN109714091B (en) Iterative hybrid precoding method based on hierarchical design in millimeter wave MIMO system
CN106972880B (en) Low-complexity joint precoding method for transmitting end and relay based on SWIPT technology
CN113411105B (en) AP selection method of non-cell large-scale antenna system
CN113098584B (en) Multi-beam satellite communication power control method for large-scale radiation array
CN111405596A (en) Resource optimization method for large-scale antenna wireless energy-carrying communication system under Rice channel
Huang et al. A Full Duplex D2D Clustering Resource Allocation Scheme Based on a-Means Algorithm
Bhardwaj et al. Deep learning-based MIMO and NOMA energy conservation and sum data rate management system
CN112512075B (en) Pilot frequency collision detection method, device and system
CN110149133B (en) Large-scale uplink transmission method based on beam space
CN109951219B (en) Low-cost large-scale non-orthogonal multi-access method
CN107346985B (en) Interference alignment method combined with transmitting antenna selection technology
CN106793120B (en) Splitting number-based rapid joint resource allocation method in virtual MIMO system
CN105451350A (en) Combined unicast and multicast mechanism-based resource allocation method
CN112243283B (en) Cell-Free Massive MIMO network clustering calculation method based on successful transmission probability
CN107835043A (en) A kind of method of information transfer outage probability in rapid evaluation wireless power communication
Huang et al. Wireless federated learning over MIMO networks: Joint device scheduling and beamforming design
CN109194375B (en) FD-MIMO multi-cell downlink interference coordination method
JP2003304192A (en) Uplink data communication equipment and method thereof

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant