High-precision clock calibration and positioning direction finding method based on 5G millimeter wave communication

Abstract

The present invention proposes a high-precision clock calibration and positioning and direction-finding method based on 5G millimeter wave communication, which relates to the technical field of wireless sensor network communication and synchronous positioning, and solves the problem that the current synchronization and positioning planning method of wireless sensors cannot jointly estimate the parameters in position, time parameters, azimuth, and transmission channel gain, resulting in poor synchronization positioning accuracy. First, a network model with unknown target nodes and anchor nodes is established. The transmitter of the anchor node starts to send information about the anchor node position, the antenna direction of the anchor node transmitter, and the local clock parameters of the anchor node to the receiver of the unknown target node. The information is processed using a signal transmission model and transmitted to the unknown target node. The received signal waveform is then used to perform joint synchronous positioning, and the position, time parameters, azimuth, and transmission channel gain are jointly estimated, effectively improving the accuracy of synchronous positioning.

Description

Technical Field

[0001] The present invention relates to the technical field of wireless sensor network communication and synchronous positioning, and in particular to a high-precision clock calibration and positioning direction-finding method based on 5G millimeter wave communication. Background Art

[0002] Wireless Sensor Networks (WSN) are widely used in many fields such as military, civil and commercial due to their advantages such as low cost, low power consumption and self-organization. Positioning and synchronization are important components of each wireless sensor network WSN. In traditional research, synchronization and positioning have been studied separately.

[0003] Synchronization research is mainly carried out from the perspective of protocol design, such as broadcast synchronization protocol and timing synchronization protocol. These protocols require sensor nodes to periodically send high-bandwidth signals, resulting in limited channel capacity and available energy. In addition, the target signal may be blocked, which again limits network performance; positioning research is mainly carried out from the perspective of signal processing, and the most accurate positioning method at present is the positioning method based on the distance between the anchor node and the target node. This method usually uses the arrival time, arrival time difference or received signal strength to calculate the distance between the target node and the anchor point, and then obtains the node position from the calculated distance. However, with the development of 5G technology, the basic technologies of 5G applications such as millimeter wave signals and multi-input and multi-output large antenna arrays can achieve higher-precision positioning and jointly estimate the target node's orientation information while locating the target node. In fact, synchronization and positioning are related, and many studies on wireless sensor networks ignore the relationship between the two.

[0004] For example, in 2014, Vaghefi, Reza Monir and Buehrer, R Michael et al. published a paper titled “Joint TOA-Based Sensor Synchronization and Localization Using Semidefinite Programming", this document considers the close relationship between synchronization and positioning, performs joint synchronous positioning of the target node, estimates the clock parameters of the local clock of the target node while calculating the location information of the target node, and jointly synchronizes the positioning of the known anchor node and the target node signal receiving and sending timestamps. On the one hand, in practical applications, the signal receiving and sending timestamp is an unknown quantity, which is estimated using the receiving and sending signal waveform. This estimation method itself has errors. At the same time, the joint synchronous positioning using the signal receiving and sending timestamps cannot effectively utilize the information about the arrival angle and the transmission angle provided by the basic technologies of 5G applications such as millimeter wave signals and multi-input multi-output large antenna arrays, and cannot realize the estimation of the target node's orientation; on the other hand, most of the current positioning studies based on signal processing regard the transmission channel gain as a known quantity, but in fact the transmission channel gain will change over time, so the transmission channel gain estimated by this method also has deviations. Summary of the invention

[0005] In order to solve the problem that the current synchronization and positioning planning methods of wireless sensors are unable to jointly estimate the parameters in position, time parameters, azimuth, and transmission channel gain, resulting in poor synchronization and positioning accuracy, the present invention proposes a high-precision clock calibration and positioning and direction-finding method based on 5G millimeter wave communication, which directly uses the waveforms of transmitted and received signals for joint synchronization and positioning, and jointly estimates the position, time parameters, azimuth, and transmission channel gain, thereby effectively improving the accuracy of synchronization and positioning.

[0006] In order to achieve the above technical effects, the technical solution of the present invention is as follows:

[0007] A high-precision clock calibration and positioning direction finding method based on 5G millimeter wave communication, comprising the following steps:

[0008] S1. Establish a network model for clock calibration and positioning direction finding, wherein the network model includes an unknown target node to be located and a plurality of anchor nodes randomly distributed around the target node, each anchor node is provided with a transmitter, the unknown target node is provided with a receiver, and both the transmitter and the receiver are provided with antennas;

[0009] S2. The unknown target node sends a positioning request to the anchor node;

[0010] S3. The anchor node receives the positioning request and uses the transmitter to send a first detection signal and a data packet containing the true value of the first detection signal, the anchor node position, the antenna direction of the anchor node transmitter and the anchor node local clock parameters to the unknown target node;

[0011] S4. Processing the first detection signal using the anchor node position, the antenna direction of the anchor node transmitter, the anchor node local clock parameters, and the signal transmission model to obtain a second detection signal transmitted to the unknown target node;

[0012] S5. Assume the position of the receiver of the unknown target node, the antenna direction of the receiver, the local clock parameters of the unknown target node, and the transmission channel gain between the transmitter and the receiver as the parameters to be estimated, and set the initial estimated value and initial error value of the parameters to be estimated;

[0013] S6. Using the first detection signal and the second detection signal to update the initial estimated value and the initial error value of the parameter to be estimated until the updating process converges to obtain a joint estimated value of the parameter to be estimated;

[0014] S7. Set the joint estimation value as the positioning and direction finding result of the unknown target node.

[0015] In the present technical scheme, a network model for clock calibration and positioning and direction finding is first established to avoid the clock drift effect of frames in communication. The network model includes an unknown target node and an anchor node. Then, the unknown target node is used to send a positioning request to the anchor node. After the anchor node receives the positioning request, the transmitter of the anchor node begins to send a first detection signal, a true value of the first detection signal, the position of the anchor node, the antenna direction of the transmitter of the anchor node, and information on the local clock parameters of the anchor node to the unknown target node. The information is processed using a signal transmission model and transmitted to the unknown target node. Then, the position of the receiver of the unknown target node, the antenna direction of the receiver, the local clock parameters of the unknown target node, and the transmission channel gain between the transmitter and the receiver are set as parameters to be estimated, and the initial estimated values ​​and initial error values ​​of the parameters to be estimated are set. The values ​​of the parameters to be estimated are set and jointly estimated to obtain the joint estimated values ​​of the parameters to be estimated, thereby realizing the joint estimation of the position, time parameters, orientation, and transmission channel gain, and effectively improving the accuracy of synchronous positioning.

[0016] Preferably, in step S1, the number of unknown target nodes to be located is one, the number of anchor nodes is M, and the number of antennas of the transmitter is N. B Each antenna is pointed in the direction b m ∈R 2 The antenna spacing of the transmitter is d a ; Assume the receiver antenna is N M Each antenna is pointed in the direction bp ∈R 2 Arrangement, the antenna spacing of the receiver and the antenna spacing of the transmitter d a same.

[0017] Preferably, in step S3, assuming that the first detection signal is x[k,m,n], the calculation expression of the first detection signal x[k,m,n] is:

[0018]

[0019] Among them, x[k,m,n] is the first detection signal, M t is the sampling number; k is the signal transmitted by the kth antenna of the transmitter, k=1,…,N B , m is the signal sent by the mth transmitter, m=1,…,M, n is the nth subcarrier, n=1,…,N;

[0020] Let the anchor node position be q m , its calculation expression is:

[0021]

[0022] Among them, q m is the position of the Mth anchor node;

[0023] Assume that the antenna direction of the anchor node's transmitter is b m , the local clock parameter of the anchor node is the skewness α of the mth transmitter clock m and the offset θ of the mth transmitter clock m , the first local time of the transmitter of the mth anchor node is t m , its calculation expression is:

[0024] t m =α m t+θ m

[0025] Among them, t is the input setting parameter.

[0026] Preferably, in step S4, the position of the receiver of the unknown target node is set to p and the antenna direction of the receiver is set to b by uniform random generation. p , the local clock parameter of the unknown target node is α p and θ p , the transmission channel gain between the transmitter and the receiver is h. During the transmission of the second detection signal to the unknown target node, the local time of the receiver of the unknown target node is t p , its calculation expression is:

[0027] t p =α pt+θ p

[0028] Among them, α p ,θ p is the local clock parameter of the unknown target node; the transmitter of the mth anchor node is at the specific local time R m Send information and the receiver of the target node receives it at a specific local time T m The information is received, the transmission speed of the information is c, and the second detection signal received by the receiver is y[i,k,m,n]. The expression of the second detection signal y[i,k,m,n] is:

[0029]

[0030] Where i represents the signal received by the i-th antenna of the receiver, i = 1, ..., N M ;

[0031] Observation errors need to be taken into account when a transmitter sends information to a receiver:

[0032]

[0033] Consider the second detection signal observation error:

[0034]

[0035] Among them, ε[i,k,m,n] is the observation error, ε l [i, k, m, n] are independent and identically distributed under a Gaussian distribution with zero mean and variance N0, l=1,...,M t .

[0036] Preferably, the expression of the signal transmission model is:

[0037] y=Zh+ε

[0038] in

[0039]

[0040]

[0041]

[0042]

[0043] in, are y[i,k,m,n], z[i,k,m,n], and ε[i,k,m,n] respectively, which are vectors strung together in the order of i,k,n. z[i,k,m,n] is the signal theoretically received by the target node receiver without considering the transmission channel gain h.

[0044] The calculation expression of z[i,k,m,n] is:

[0045] z[i,k,m,n]=a R [i,m,n]a T [k,m,n]Γ[m,n]x[m,n]

[0046] Among them, a R [i,m,n] is the emission angle gain, a T [k,m,n] is the arrival angle gain, Γ[m,n] is the loss caused by the transmission distance, a R The calculation expression of [i,m,n] is

[0047]

[0048] in, is the transmission angle, that is, the transmission direction of the line of sight path from the mth anchor node to the target node and the direction of the mth anchor node transmitter toward b m The angle, λ n is the signal wavelength of the nth subcarrier, The calculation expression is:

[0049]

[0050] λ n The calculation expression is:

[0051]

[0052] Where c is the signal transmission speed, is the sampling period, f c is the carrier frequency, B is the bandwidth;

[0053] a T The calculation expression of [k,m,n] is:

[0054]

[0055] in, is the transmission angle, that is, the transmission direction of the line of sight path from the mth anchor node to the target node and the direction of the mth anchor node transmitter toward b m The angle of

[0056] The calculation expression is:

[0057]

[0058] The calculation expression of Γ[m,n] is:

[0059]

[0060] Among them, τ m is the transmission time, and its calculation expression is:

[0061]

[0062] Among them, d m is the distance between the mth anchor node and the unknown target node, and its calculation expression is:

[0063]

[0064] Preferably, in step S5, the position p of the receiver of the unknown target node and the antenna direction b of the receiver are assumed to be p , unknown target node local clock parameter α p ,θ p The transmission channel gain h between the transmitter and the receiver is the parameter to be estimated, and the initial estimated value and initial error value of the parameter to be estimated are set. The specific processing process is as follows:

[0065] Assume that the estimated parameters p0,b p,0 ,α p,0 ,θ p,0 , the initial values ​​of h are p0,b p,0 ,α p,0 ,θ p,0 ,h0, for p,b p , α p ,θ p , the initial value of h is jointly estimated, and the initial estimated value of Z is calculated as Z0 and the initial error value is r0, where the calculation expression of the initial estimated value of Z0 is:

[0066] Z0=Z(p0,b p,0 ,α p,0 ,θ p,0 )

[0067] Where Z is (p, b p ,α p ,θ p ) function;

[0068] The calculation expression of the initial value h0 of the transmission channel gain is:

[0069]

[0070] The calculation expression of the initial error value r0 is:

[0071] r0=y-Z0h0.

[0072] Preferably, in step S6, the specific steps of updating are:

[0073] S61.Set r -1 N B *N M *M*N*M t The dimension 0 vector is, t is 0, the threshold is δ, and the number of cutoffs is N T ;

[0074] S62. Judgment ‖r t -r t-1 Is ‖2 less than δ, or is t greater than or equal to N? T If so, let p, b p , α p ,θ p , the estimated values ​​of h are p t , b p,t , α p,t ,θ p,t ,h t , output joint estimate Otherwise, let t=t+1 and execute step S63;

[0075] S63. Using gradient descent to estimate parameters p, b p ,α p ,θ p The value is updated and b p The updated value is normalized to obtain (p t ,b p,t ,α p,t ,θ p,t ), execute step S64;

[0076] S64. Calculate the estimated value Z after the tth iteration of Z t , where Z t The calculation expression is:

[0077] Z t =Z(p t ,b p,t ,α p,t ,θ p,t )

[0078] Use linear regression to update the estimated value of h to h t ;

[0079] S65. Calculate the error value after the tth iteration, and the calculation expression is:

[0080] r t =yZ t h t

[0081] The error value r after the tth iteration t Calculate the square and get: S t =r t T r t , return to step S62.

[0082] Preferably, in step S63, during the gradient descent process, the step size is set to β, and for the tth iteration, the Sth iteration is calculated. t-1 Step about p t-1 ,b p,t-1 ,α p,t-1 ,θ p,t-1 Gradient For the estimated parameters p,b p ,α p ,θ p The estimated value of is updated, and the update expression is:

[0083]

[0084]

[0085]

[0086]

[0087] For b p Updated b p,t * The value is normalized, and the normalized calculation expression is:

[0088]

[0089] Preferably, in step S64, the calculation expression for updating the estimated parameter h by linear regression is:

[0090] h t =(Z t T Z t ) -1 Z t T y.

[0091] Preferably, the transmitter position is Q, and the calculation expression of Q is:

[0092]

[0093] The transmitter has clock parameters

[0094] Compared with the prior art, the technical solution of the present invention has the following beneficial effects:

[0095] The present invention proposes a high-precision clock calibration and positioning direction-finding method based on 5G millimeter wave communication. First, a network model for clock calibration and positioning direction-finding is established. Then, an unknown target node is used to send a positioning request to an anchor node. After the anchor node receives the positioning request, the transmitter of the anchor node starts to send a first detection signal, a true value of the first detection signal, the anchor node position, the antenna direction of the anchor node's transmitter, and information on the local clock parameters of the anchor node to the unknown target node. The information is processed using a signal transmission model and transmitted to the unknown target node. Then, the position of the receiver of the unknown target node, the antenna direction of the receiver, the local clock parameters of the unknown target node, and the transmission channel gain between the transmitter and the receiver are set as parameters to be estimated, and the initial estimated values ​​and initial error values ​​of the parameters to be estimated are set. The values ​​of the parameters to be estimated are set and jointly estimated to obtain the joint estimated values ​​of the parameters to be estimated, thereby realizing the joint estimation of the position, time parameters, orientation, and transmission channel gain, and effectively improving the accuracy of synchronous positioning. BRIEF DESCRIPTION OF THE DRAWINGS

[0096] Figure 1 A schematic diagram showing a flow chart of a high-precision clock calibration and positioning direction finding method based on 5G millimeter wave communication proposed in an embodiment of the present invention;

[0097] Figure 2 A structural diagram showing a network model for high-precision clock calibration and positioning direction finding based on 5G millimeter wave communication;

[0098] Figure 3 A schematic diagram showing a receiver antenna receiving a signal;

[0099] Figure 4 CDF diagram showing the position estimation error when there is 20dB noise;

[0100] Figure 5 CDF diagram showing the azimuth estimation error when there is 20dB noise;

[0101] Figure 6 CDF diagram showing the error in clock parameter offset estimation when there is 20dB noise;

[0102] Figure 7 CDF diagram showing the error in clock parameter skew estimation when there is 20dB noise;

[0103] Figure 8 CDF diagram showing the root mean square error of transmission channel gain estimation when there is 20dB noise;

[0104] Fig. 9 A plot showing the relationship between the estimation error and the signal-to-noise ratio for each estimated parameter;

[0105] Fig.10 A diagram showing a comparison of the square error of the position estimation of the method of the present invention and the square error of the position estimation of the positioning method based on the waveform of the received and transmitted signal without considering the clock deviation;

[0106] Fig.11 A diagram showing a comparison of the square error of the position estimation by the method of the present invention and the square error of the position estimation by the method based on the sending and receiving timestamp; DETAILED DESCRIPTION

[0107] The drawings are for illustrative purposes only and should not be construed as limiting the present patent;

[0108] In order to better illustrate the present embodiment, some parts of the drawings may be omitted, enlarged or reduced, and do not represent actual sizes. The description of the directions of parts such as "upper" and "lower" does not limit the present patent;

[0109] It is understandable to those skilled in the art that some well-known contents may be omitted in the drawings;

[0110] The positional relationships described in the drawings are only for illustrative purposes and should not be construed as limiting the present patent.

[0111] The technical solution of the present invention is further described below in conjunction with the accompanying drawings and embodiments.

[0112] Example 1

[0113] like Figure 1 As shown, a high-precision clock calibration and positioning direction finding method based on 5G millimeter wave communication includes the following steps:

[0114] S1. Establish a network model for clock calibration and positioning direction finding, wherein the network model includes an unknown target node to be located and a plurality of anchor nodes randomly distributed around the target node, each anchor node is provided with a transmitter, the unknown target node is provided with a receiver, and both the transmitter and the receiver are provided with antennas;

[0115] In step S1, see Figure 1 and Figure 2 , assuming that the number of unknown target nodes to be located is one, the number of anchor nodes is M, and the number of transmitter antennas is N B Each antenna is pointed in the direction b m ∈R 2 The antenna spacing of the transmitter is d a ; Assume the receiver antenna is N M Each antenna is pointed in the direction b p ∈R 2 Arrangement, the antenna spacing of the receiver and the antenna spacing of the transmitter d aSimilarly, the transmitter position is Q, and the calculation expression of Q is:

[0116]

[0117] The transmitter has clock parameters

[0118] S2. The unknown target node sends a positioning request to the anchor node;

[0119] S3. The anchor node receives the positioning request and uses the transmitter to send a first detection signal and a data packet containing the true value of the first detection signal, the anchor node position, the antenna direction of the anchor node transmitter and the anchor node local clock parameters to the unknown target node;

[0120] In step S3, see Figure 1 and Figure 3 , assuming that the first detection signal is x[k,m,n], the calculation expression of the first detection signal x[k,m,n] is:

[0121]

[0122] Among them, x[k,m,n] is the first detection signal, M t is the sampling number; k is the signal transmitted by the kth antenna of the transmitter, k=1,…,N B , m is the signal sent by the mth transmitter, m=1,…,M, n is the nth subcarrier, n=1,…,N;

[0123] Let the anchor node position be q m , its calculation expression is:

[0124]

[0125] Among them, q m is the position of the Mth anchor node;

[0126] Assume that the local clock parameter of the anchor node is the skewness α of the mth transmitter clock m and the offset θ of the mth transmitter clock m , the local time of the transmitter of the mth anchor node is t m , its calculation expression is:

[0127] t m =α m t+θ m

[0128] Among them, t is the standard time.

[0129] S4. Processing the first detection signal using the anchor node position, the antenna direction of the anchor node transmitter, the anchor node local clock parameters, and the signal transmission model to obtain a second detection signal transmitted to the unknown target node;

[0130] S5. Assume the position of the receiver of the unknown target node, the antenna direction of the receiver, the local clock parameters of the unknown target node, and the transmission channel gain between the transmitter and the receiver as the parameters to be estimated, and set the initial estimated value and initial error value of the parameters to be estimated;

[0131] In step S5, let the position of the receiver of the unknown target node p and the antenna direction of the receiver b p , unknown target node local clock parameter α p ,θ p The transmission channel gain h between the transmitter and the receiver is the parameter to be estimated, and the initial estimated value and initial error value of the parameter to be estimated are set. The specific processing process is as follows:

[0132] Assume that the estimated parameters p0,b p,0 ,α p,0 ,θ p,0 , the initial values ​​of h are p0,b p,0 ,α p,0 ,θ p,0 ,h0, for p,b p , α p ,θ p , the initial value of h is jointly estimated, and the initial estimated value of Z is calculated as Z0 and the initial error value is r0, where the calculation expression of the initial estimated value of Z0 is:

[0133] Z0=Z(p0,b p,0 ,α p,0 ,θ p,0 )

[0134] Where Z is (p, b p ,α p ,θ p ) function;

[0135] The calculation expression of the initial value h0 of the transmission channel gain is:

[0136]

[0137] The calculation expression of the initial error value r0 is:

[0138] r0=y-Z0h0.

[0139] S6. Using the first detection signal and the second detection signal to update the initial estimated value and the initial error value of the parameter to be estimated until the updating process converges to obtain a joint estimated value of the parameter to be estimated;

[0140] S7. Set the joint estimation value as the positioning and direction finding result of the unknown target node.

[0141] See also Figure 2 First, the transmitter and receiver are used to directly transmit and receive signal waveforms to perform synchronous positioning to improve the accuracy of synchronous positioning. Then, see Figure 3 Since the interval between antennas is much smaller than the distance between the transmitter and the receiver, the transmission signal can be regarded as arriving at each antenna in parallel. The transmission distance between adjacent antennas can be regarded as the antenna spacing multiplied by the cosine value of the angle. Therefore, the angle can be estimated to perform antenna direction finding and reduce the influence of the antenna orientation on the received waveform. Finally, considering the possible clock deviation and the inaccurate prior transmission channel gain, these two parameters are jointly estimated to reduce the deviation caused by the clock deviation and the inaccurate prior channel gain, thereby improving the positioning accuracy.

[0142] Example 2

[0143] See also Figure 1 and Figure 3 In step S4, the position of the receiver of the unknown target node is set to p and the antenna direction of the receiver is set to b by uniform random generation. p , the local clock parameter of the unknown target node is α p and θ p , the transmission channel gain between the transmitter and the receiver is h. During the transmission of the second detection signal to the unknown target node, the local time of the receiver of the unknown target node is t p , its calculation expression is:

[0144] t p =α p t+θ p

[0145] Among them, α p ,θ p is the local clock parameter of the unknown target node; the transmitter of the mth anchor node is at the specific local time R m Send information and the receiver of the target node receives it at a specific local time T m The information is received, the transmission speed of the information is c, and the second detection signal received by the receiver is y[i,k,m,n]. The expression of the second detection signal y[i,k,m,n] is:

[0146]

[0147] Where i represents the signal received by the i-th antenna of the receiver, i = 1, ..., N M ;

[0148] Observation errors need to be taken into account when a transmitter sends information to a receiver:

[0149]

[0150] Consider the second detection signal observation error:

[0151]

[0152] Among them, ε[i,k,m,n] is the observation error, ε l [i, k, m, n] are independent and identically distributed under a Gaussian distribution with zero mean and variance N0, l=1,...,M t .

[0153] The expression of the signal transmission model is:

[0154] y=Zh+ε

[0155] in

[0156]

[0157]

[0158]

[0159]

[0160] in, are y[i,k,m,n], z[i,k,m,n], and ε[i,k,m,n] respectively, which are vectors strung together in the order of i,k,n. z[i,k,m,n] is the signal theoretically received by the target node receiver without considering the transmission channel gain h.

[0161] The calculation expression of z[i,k,m,n] is:

[0162] z[i,k,m,n]=a R [i,m,n]a T [k,m,n]Γ[m,n]x[m,n]

[0163] Among them, a R [i,m,n] is the emission angle gain, a T [k,m,n] is the arrival angle gain, Γ[m,n] is the loss caused by the transmission distance, a R The calculation expression of [i,m,n] is

[0164]

[0165] in, is the transmission angle, that is, the transmission direction of the line of sight path from the mth anchor node to the target node and the direction of the mth anchor node transmitter toward b m The angle, λ n is the signal wavelength of the nth subcarrier, The calculation expression is:

[0166]

[0167] λ n The calculation expression is:

[0168]

[0169] Where c is the signal transmission speed, is the sampling period, f c is the carrier frequency, B is the bandwidth;

[0170] a T The calculation expression of [k,m,n] is:

[0171]

[0172] in, is the transmission angle, that is, the transmission direction of the line of sight path from the mth anchor node to the target node and the direction of the mth anchor node transmitter toward b m The angle of

[0173] The calculation expression is:

[0174]

[0175] The calculation expression of Γ[m,n] is:

[0176]

[0177] Among them, τ m is the transmission time, and its calculation expression is:

[0178]

[0179] Among them, d m is the distance between the mth anchor node and the unknown target node, and its calculation expression is:

[0180]

[0181] Example 3

[0182] Table 1 shows the setting values ​​of various known parameters. Referring to Table 1, a 100m×100m site is considered, and a network model of high-precision clock calibration and positioning direction finding based on 5G millimeter wave communication is established in the site. Figure 2 and Figure 3 In this network model, the carrier frequency f is set c =60GHz, bandwidth B = 100MHz, signal transmission speed c = 3×10 8 m / s, x is the generated discrete time domain signal converted into the frequency domain signal by Fourier transform. t =10 sampling points, the number of subcarriers N = 5, the transmitter clock parameter α q =1,θ q = 0, that is, in the simulation experiment, the transmitter clock time is set to the standard time, the transmitter position Q is uniformly randomly generated in a 100m×100m place, the receiver position p real value and initial value are uniformly randomly generated in a 100m×100m place, and the direction b is uniformly randomly generated on the unit circle. p True value and initial value, uniformly randomly generate clock parameter α between [0,1] p True value and initial value, uniformly randomly generate clock parameter θ between [0.95,1.05] p The actual value and initial value are uniformly randomly generated on [0.8, 0.95], and the component h of the transmission channel gain h between the transmitter and the target receiver is m The true value and initial value of ; consider that the signal-to-noise ratio of the frequency domain signal received by the receiver is 20dB to randomly generate noise.

[0183] Table 1

[0184]

[0185]

[0186] See also Figure 1 In step S6, the specific steps of updating the first initial estimate value and the first initial error value of the initial value are:

[0187] S61.Set r -1 N B *N M *M*N*M t The dimension 0 vector is, t is 0, the threshold is δ, and the number of cutoffs is N T ;

[0188] S62. Judgment ‖r t -r t-1 Is ‖2 less than δ, or is t greater than or equal to N? TIf so, let p, b p , α p ,θ p , the estimated values ​​of h are p t , b p,t , α p,t ,θ p,t ,h t , output joint estimate Otherwise, let t=t+1 and execute step S63;

[0189] S63. Using gradient descent to estimate parameters p, b p ,α p ,θ p The value is updated and b p The updated value is normalized to obtain (p t ,b p,t ,α p,t ,θ p,t ), execute step S64;

[0190] In step S63, during the gradient descent process, the step size is set to β=5, and for the tth iteration, the error of the tth iteration is:

[0191] r t [i,k,m,n,l]=y[i,k,m,n,l]-h m,t z t [i,k,m,n,l]

[0192] z t [i,k,m,n,l]=a R,t [i,m,n]a T,t [k,m,n]Γ t [m,n]x[k,m,n,l]

[0193] The emission angle gain in the tth iteration is:

[0194]

[0195] The arrival angle gain in the tth iteration is:

[0196]

[0197] The gain caused by the corrected transmission distance in the tth iteration is:

[0198]

[0199] The cosine of the emission angle in the tth iteration is:

[0200]

[0201] The cosine of the angle reached in the tth iteration is:

[0202] The error of the tth iteration is:

[0203] r t [i,k,m,n,l]=y[i,k,m,n,l]-h m,t z t [i,k,m,n,l]

[0204] z t [i,k,m,n,l]=a R,t [i,m,n]a T,t [k,m,n]Γ t [m,n]x[k,m,n,l]

[0205] The emission angle gain in the tth iteration is:

[0206]

[0207] The arrival angle gain in the tth iteration is:

[0208]

[0209] The gain caused by the corrected transmission distance in the tth iteration is:

[0210]

[0211] The cosine of the emission angle in the tth iteration is:

[0212]

[0213] The cosine of the angle reached in the tth iteration is:

[0214]

[0215] Calculate S t-1 Step about p t-1 ,b p,t-1 ,α p,t-1 ,θ p,t-1 Gradient For the estimated parameters p,b p ,α p ,θ p The estimated value of is updated, and the update expression is:

[0216]

[0217]

[0218]

[0219]

[0220] in

[0221] U p,t , The values ​​of each component are:

[0222]

[0223]

[0224]

[0225]

[0226] in

[0227]

[0228]

[0229]

[0230]

[0231]

[0232]

[0233]

[0234] For b p Updated b p,t * The value is normalized, and the normalized calculation expression is:

[0235]

[0236] S64. Calculate the estimated value Z after the tth iteration of Z t , where Z t The calculation expression is:

[0237] Z t =Z(p t ,b p,t ,α p,t ,θ p,t )

[0238] Use linear regression to update the estimated value of h to h t;

[0239] In step S64, the calculation expression for updating the estimated parameter h by linear regression is:

[0240] h t =(Z t T Z t ) -1 Z t T y.

[0241] S65. Calculate the error value after the tth iteration, and the calculation expression is:

[0242] r t =yZ t h t

[0243] The error value r after the tth iteration t Calculate the square and get: S t =r t T r t , return to step S62.

[0244] This embodiment conducts 3000 experiments at a signal-to-noise ratio of 20 dB, and obtains the cumulative distribution function diagrams of the unknown target node position estimation error, orientation estimation error, bias estimation error, skewness estimation error, and transmission channel gain estimation root mean square error, see Figure 4 , the position estimation error is basically below 0.1m; see Figure 5 , the azimuth estimation error is basically below 0.05rad; see Figure 6 , the clock parameter offset estimation error is basically below 0.1ns; see Figure 7 , the error of clock parameter skew estimation is basically below 0.002; see Figure 8 , the root mean square error of transmission channel gain estimation is basically below 0.02.

[0245] Fig. 9 is the relationship between the estimation error and signal-to-noise ratio of each estimated parameter, see Figure 6 , the logarithm of the estimation error of each estimation parameter decreases linearly with the increase of the signal-to-noise ratio.

[0246] See also Fig.10 , the square error of the position estimation of this method is compared with the square error of the position estimation of the positioning method based on the waveform of the receiving and transmitting signal without considering the clock deviation. The horizontal axis is the number of iterations, and the vertical axis is the square error of the position estimation. Depend on Fig.10It can be seen that the positioning accuracy with clock deviation considered is significantly better than that without considering it, which means that if the clock deviation is not considered for synchronization calibration, the obtained positioning information will have a large error.

[0247] See also Fig.11 , the square error of the position estimation of this method is compared with the square error of the position estimation method using the sending and receiving timestamps for joint estimation. The horizontal axis is the number of iterations, and the vertical axis is the square error of the position estimation. Depend on Fig.11 It can be seen that the effect of joint estimation using the received and transmitted frequency domain signals is better than that using the timestamps. In addition, the joint estimation using the received and transmitted frequency domain signals can estimate the orientation information of the target device, while the method using the received and transmitted timestamps cannot estimate the orientation of the target device.

[0248] Obviously, the above embodiments of the present invention are only examples for clearly illustrating the present invention, and are not intended to limit the embodiments of the present invention. For those skilled in the art, other different forms of changes or modifications can be made based on the above description. It is not necessary and impossible to list all the embodiments here. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the claims of the present invention.

Claims

1. A high-precision clock calibration and positioning direction finding method based on 5G millimeter wave communication, characterized in that: The following steps are involved: S1. Establish a network model for clock calibration and positioning direction finding, wherein the network model includes an unknown target node to be located and a plurality of anchor nodes randomly distributed around the target node, each anchor node is provided with a transmitter, the unknown target node is provided with a receiver, and both the transmitter and the receiver are provided with antennas; S2. The unknown target node sends a positioning request to the anchor node; S3. The anchor node receives the positioning request and uses the transmitter to send a first detection signal and a data packet containing the true value of the first detection signal, the anchor node position, the antenna direction of the anchor node transmitter and the anchor node local clock parameters to the unknown target node; S4. Processing the first detection signal using the anchor node position, the antenna direction of the anchor node transmitter, the anchor node local clock parameters, and the signal transmission model to obtain a second detection signal transmitted to the unknown target node; S5. Assume the position of the receiver of the unknown target node, the antenna direction of the receiver, the local clock parameters of the unknown target node, and the transmission channel gain between the transmitter and the receiver as the parameters to be estimated, and set the initial estimated value and initial error value of the parameters to be estimated; S6. Using the first detection signal and the second detection signal to update the initial estimated value and the initial error value of the parameter to be estimated until the updating process converges to obtain a joint estimated value of the parameter to be estimated; S7. Set the joint estimation value as the positioning and direction finding result of the unknown target node.

2. The high-precision clock calibration and positioning direction finding method based on 5G millimeter wave communication according to claim 1 is characterized in that: In step S1, assume that the number of unknown target nodes to be located is one, the number of anchor nodes is M, and the number of transmitter antennas is N. B Each antenna is pointed in the direction b m ∈R 2 The antenna spacing of the transmitter is d a ; Assume the receiver antenna is N M Each antenna is pointed in the direction b p ∈R 2 Arrangement, the antenna spacing of the receiver and the antenna spacing of the transmitter d a same.

3. The high-precision clock calibration and positioning direction finding method based on 5G millimeter wave communication according to claim 2 is characterized in that: In step S3, assuming that the first detection signal is x[k,m,n], the calculation expression of the first detection signal x[k,m,n] is: Among them, x[k,m,n] is the first detection signal, M t is the sampling number; k is the signal transmitted by the kth antenna of the transmitter, k=1,…,N B , m is the signal sent by the mth transmitter, m=1,…,M, n is the nth subcarrier, n=1,…,N; Let the anchor node position be q m , its calculation expression is: Among them, q m is the position of the Mth anchor node; Assume that the local clock parameter of the anchor node is the skewness α of the mth transmitter clock m and the offset θ of the mth transmitter clock m , the local time of the transmitter of the mth anchor node is t m , its calculation expression is: t m =a m t+θ m Among them, t is the standard time.

4. The high-precision clock calibration and positioning direction finding method based on 5G millimeter wave communication according to claim 3 is characterized in that: In step S4, the position of the receiver of the unknown target node is set to p and the antenna direction of the receiver is set to b by uniform random generation. p , the local clock parameter of the unknown target node is α p and θ p , the transmission channel gain between the transmitter and the receiver is h. During the transmission of the second detection signal to the unknown target node, the local time of the receiver of the unknown target node is t p , its calculation expression is: t p =a p t+θ p Among them, α p ,θ p is the local clock parameter of the unknown target node; the transmitter of the mth anchor node is at the specific local time R m Send information and the receiver of the target node receives it at a specific local time T m The information is received, the transmission speed of the information is c, and the second detection signal received by the receiver is y[i,k,m,n]. The expression of the second detection signal y[i,k,m,n] is: Where i represents the signal received by the i-th antenna of the receiver, i = 1, ..., N M ; Observation errors need to be taken into account when a transmitter sends information to a receiver: Consider the second detection signal observation error: Among them, ε[i,k,m,n] is the observation error, ε l [i, k, m, n] are independent and identically distributed under a Gaussian distribution with zero mean and variance N0, l=1,...,M t .

5. The high-precision clock calibration and positioning direction finding method based on 5G millimeter wave communication according to claim 4 is characterized in that: The expression of the signal transmission model is: y=Zh+ε in in, are y[i,k,m,n], z[i,k,m,n], and ε[i,k,m,n] respectively, which are vectors strung together in the order of i,k,n. z[i,k,m,n] is the signal theoretically received by the target node receiver without considering the transmission channel gain h. The calculation expression of z[i,k,m,n] is: z[i,k,m,n]=a R [i,m,n]a T [k,m,n]Γ[m,n]x[m,n] Among them, a R [i,m,n] is the emission angle gain, a T [k,m,n] is the arrival angle gain, Γ[m,n] is the loss caused by the transmission distance, a R The calculation expression of [i,m,n] is in, is the transmission angle, that is, the transmission direction of the line of sight path from the mth anchor node to the target node and the direction of the mth anchor node transmitter toward b m The angle, λ n is the signal wavelength of the nth subcarrier, The calculation expression is: λ n The calculation expression is: Where c is the signal transmission speed, is the sampling period, f c is the carrier frequency, B is the bandwidth; a T The calculation expression of [k,m,n] is: Among them, φ m is the transmission angle, that is, the transmission direction of the line of sight path from the mth anchor node to the target node and the direction of the mth anchor node transmitter toward b m The angle of cosφ m The calculation expression is: The calculation expression of Γ[m,n] is: Among them, τ m is the transmission time, and its calculation expression is: Among them, d m is the distance between the mth anchor node and the unknown target node, and its calculation expression is:

6. The high-precision clock calibration and positioning direction finding method based on 5G millimeter wave communication according to claim 5, characterized in that: In step S5, let the position of the receiver of the unknown target node p and the antenna direction of the receiver b p , unknown target node local clock parameter α p ,θ p The transmission channel gain h between the transmitter and the receiver is the parameter to be estimated, and the initial estimated value and initial error value of the parameter to be estimated are set. The specific processing process is as follows: Assume that the parameters to be estimated are p, b p , α p ,θ p , the initial values ​​of h are p0,b p,0 ,α p,0 ,θ p,0 ,h0, for p,b p , α p ,θ p , the initial value of h is jointly estimated, and the initial estimated value of Z is calculated as Z0 and the initial error value is r0, where the calculation expression of the initial estimated value of Z0 is: Z0=Z(p0,b p,0 ,a p,0 ,i p,0 ) Where Z is (p, b p ,α p ,θ p ) function; The calculation expression of the initial value h0 of the transmission channel gain is: The calculation expression of the initial error value r0 is: r0=y-Z0h0.

7. The high-precision clock calibration and positioning direction finding method based on 5G millimeter wave communication according to claim 6, characterized in that: In step S6, the specific steps of updating are: S61.Set r -1 N B *N M *M*N*M t dimensional 0 vector, t is 0, threshold is δ, and cutoff number is N T ; S62. Judgement || r t -r t-1 ||2 is less than δ, or t is greater than or equal to N T If so, let p, b p , α p ,θ p , the estimated values ​​of h are p t , b p,t , α p,t ,θ p,t , g t , output joint estimate Otherwise, let t=t+1 and execute step S63; S63. Using gradient descent to estimate parameters p, b p ,α p ,θ p The value is updated and b p The updated value is normalized to obtain (p t ,b p,t ,α p,t ,θ p,t ), execute step S64; S64. Calculate the estimated value Z after the tth iteration of Z t , where Z t The calculation expression is: WITH t =Z(p t ,b p,t ,α p,t ,θ p,t ) Use linear regression to update the estimated value of h to h t ; S65. Calculate the error value after the tth iteration, and the calculation expression is: r t =y-Z t h t The error value r after the tth iteration t Calculate the square and get: S t =r t T r t , return to step S62.

8. The high-precision clock calibration and positioning direction finding method based on 5G millimeter wave communication according to claim 7, characterized in that: In step S63, during the gradient descent process, the step size is set to β, and for the tth iteration, the Sth t-1 Step about p t-1 ,b p,t-1 ,α p,t-1 ,θ p,t-1 Gradient For the estimated parameters p,b p ,α p ,θ p The estimated value of is updated, and the update expression is: For b p Updated b p,t * The value is normalized, and the normalized calculation expression is:

9. The high-precision clock calibration and positioning direction finding method based on 5G millimeter wave communication according to claim 8, characterized in that: In step S64, the calculation expression for updating the estimated parameter h by linear regression is: h t =(Z t T Z t ) -1 Z t T y。 10. The high-precision clock calibration and positioning direction finding method based on 5G millimeter wave communication according to any one of claims 1 to 9, characterized in that: The transmitter position is Q, and the calculation expression of Q is: The transmitter has clock parameters

Images