KR20180072506A - Method and Apparatus for Estimating a Clock Frequency Error and a Clock Time Error in a PTP Slave - Google Patents

Abstract

The present invention relates to a device for estimating a clock frequency error in a precision time protocol (PTP) slave. The device comprises: a delay detecting unit which forms a plurality of first delay time samples and a plurality of transmitting time samples respectively related to a plurality of sync messages, and forms a plurality of second delay time samples and a plurality of receiving time samples respectively related to a plurality of delay request messages; and a clock frequency error estimating unit which estimates a clock frequency error based on a selected portion from the plurality of first delay time samples, and transmitting time samples of a portion of the plurality of transmitting time samples associated with the selected portion from the plurality of first delay time samples, and based on a selected portion from the plurality of second delay time samples, and receiving time samples of a portion of the plurality of receiving time samples associated with the selected portion from the plurality of second delay time samples. The plurality of transmitting time samples are associated with the plurality of first delay time samples, respectively, while the plurality of receiving time samples are associated with the plurality of second delay time samples, respectively. Therefore, the method and the device for estimating a clock frequency error and a clock time error in a PTP slave are able to provide increased accuracy and rapid calculation compared to the existing technology.

Description

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method and an apparatus for estimating a clock frequency error and a clock time error in a PTP slave,

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a technology for adjusting clock synchronization in a PTP (Precision Time Protocol) slave, and more particularly, to a technique for estimating a clock frequency error and a clock time error in a PTP slave

In recent radio access networks, a small cell having a small size such as a micro cell, a pico cell, or a femto cell is a relatively large macro cell macro cell) to improve service quality and expand coverage. The small cell, which is a small mobile communication base station, is a wireless access node with a low power and has a relatively narrow service area than a general cell. It is similar to a DSL modem and can be connected to a wired IP network in the home to freely use wired / Do it.

One of the basic requirements for a small cell, especially a small cell based on a TDD system, to operate as a small mobile communication base station is frequency and time synchronization with a macro base station. IEEE 1588 Precision Time Protocol (PTP) is known as one such synchronization method. According to the PTP, for example, a PTP slave such as a small cell receives periodic sync messages sent from a PTP master (master server), and also transmits delay request messages (Delay Request messages) And receives Delay Response messages from the PTP master. The PTP slave generally performs clock frequency and time synchronization by estimating a clock frequency error and a clock time error by applying an averaging method to the delay times of the sync messages and the delay times of the delay request messages. There is a problem that it is difficult to provide an accurate estimation result due to a packet delay variation (PDV) of the network.

SUMMARY OF THE INVENTION It is an object of the present invention to provide a method and an apparatus for estimating a clock frequency error and a clock time error in a PTP slave,

The problems to be solved by the present invention are not limited to the above-mentioned problems, and other problems not mentioned can be clearly understood by those skilled in the art from the following description.

In one aspect, a method for estimating a clock frequency error in a PTP slave (Precision Time Protocol slave) is provided. The method includes configuring a plurality of first delay time samples by determining a plurality of first delay times respectively associated with a plurality of sync messages transmitted sequentially from a PTP master to arrive at the PTP slave Identifying a plurality of transmission times of a plurality of sink messages sequentially transmitted from the PTP master to construct a plurality of transmission time samples, the plurality of transmission time samples being associated with the plurality of first delay time samples - determining a plurality of second delay times respectively associated with a plurality of delay request messages transmitted sequentially from the PTP slave to the PTP master to construct a plurality of second delay time samples, A plurality of delay times, in which a plurality of delay request messages sequentially transmitted from the PTP slave arrive at the PTP master, Wherein the plurality of receive time samples are each associated with the plurality of second delay time samples, and wherein the selected plurality of receive time samples comprises a selected portion of the plurality of first delay time samples, Estimating the clock frequency error based on a selected portion of time samples, a selected portion of the plurality of second delay time samples, and a selected portion of the plurality of receive time samples.

In one embodiment, a selected portion of the plurality of first delay time samples, a selected portion of the plurality of transmission time samples, a selected portion of the plurality of second delay time samples, and a selected portion of the plurality of receive time samples Wherein estimating the clock frequency error comprises: averaging each of a plurality of first delay time sample groups selected from the plurality of first delay time samples to determine a plurality of first sample average values, Determining a plurality of second sample mean values by averaging each of a plurality of transmission time sample groups selected from time samples and associated with each of the plurality of first delay time sample groups, Averaging each of a plurality of second delay time sample groups to determine a plurality of third sample average values, Determining a plurality of fourth sample mean values by averaging each of a plurality of receive time sample groups selected from time samples and associated with each of the plurality of second delay time sample groups, Constructing a plurality of first samples and a plurality of second samples from a second sample mean of the plurality of first samples and a plurality of second samples, Dimensional coordinate system using the second sample average value of the at least one outlier, and if there is at least one outlier among the points of the first group, The first sample average value and the second sample average value are respectively removed from the plurality of first sample average values and the plurality of second sample average values, Constructing a plurality of second samples, - constructing a plurality of third samples and a plurality of fourth samples from the plurality of third sample mean values and the plurality of fourth sample mean values, The step of constructing the third sample and the plurality of fourth samples may include defining a second group of points in the two dimensional coordinate system using the plurality of third sample average values and the plurality of fourth sample average values, A third sample average value and a fourth sample mean value corresponding to the at least one outlier are removed from the plurality of third sample average values and the plurality of fourth sample average values, respectively, when there is at least one outlier among the plurality And constructing the plurality of fourth samples, and using the plurality of first samples to the plurality of fourth samples to obtain the clock frequency error And a step of estimating.

In one embodiment, the step of estimating the clock frequency error using the plurality of first samples to the plurality of fourth samples may include estimating the clock frequency error using the plurality of first samples and the plurality of second samples, Defining a third group of points in the third group and defining points of the third group as a first one-dimensional function, determining a slope of the first one-dimensional function as a first slope value, Defining a fourth group of points in a two-dimensional coordinate system using the plurality of fourth samples and defining the points of the fourth group as a second one-dimensional function, calculating a slope of the second one-dimensional function as a second slope value Calculating a standard deviation for the plurality of first samples as a first standard deviation, computing a standard deviation for the plurality of third samples as a second standard deviation, It may include a value group, the method comprising estimating the second slope value, the first standard deviation and said clock frequency error by using the second standard deviation.

In one embodiment, the method further comprises calculating an average value for the plurality of first samples as a first average value, calculating an average value for the plurality of second samples as a second average value, Calculating a mean value for the sample as a third mean value, calculating an average value for the plurality of fourth samples as a fourth mean value, calculating the first mean value to the fourth mean value, the estimated clock frequency error, And estimating a clock time error using a value indicating a clock time error.

In one embodiment, the step of averaging each of a plurality of first delay time sample groups selected from the plurality of first delay time samples to determine a plurality of first sample average values comprises: N _sub of first delay time sample comprising: dividing a plurality of the first sub-block composed of each said N _sub is a natural number - at least a first plurality of M _sub of first delay sample from each of the sub-block , _Wherein M _sub is a natural number smaller than the N _sub , and selecting N _ran second sub-blocks that are valid among the plurality of second sub-blocks, wherein the N _ran is a sub- And determining an average value of M _sub first delay time samples included in each of the selected N _ran second sub-blocks as the corresponding first sample average value, have.

In one embodiment, the step of averaging each of a plurality of second delay time samples selected from the plurality of second delay time samples to determine a plurality of third sample average values comprises: N _sub of second delay time, a sample comprising: dividing a plurality of the first sub-block composed of each said N _sub is a natural number - at least a first plurality of M _sub of second delay time samples from each of the sub-block , _Wherein M _sub is a natural number smaller than the N _sub , and selecting N _ran second sub-blocks that are valid among the plurality of second sub-blocks, wherein the N _ran is a sub- And determining an average value of M _sub second delay time samples included in each of the selected N _ran second sub-blocks as the corresponding third sample average value. have.

In one embodiment, the step of selecting N _ran second valid subblocks among the plurality of second subblocks may include determining a standard for M _sub first delay time samples included in each of the plurality of second subblocks, And determining the second sub-block as a valid second sub-block if the deviation is smaller than the predetermined threshold value.

In one embodiment, the step of selecting valid N _ran second sub-blocks among the plurality of second sub-blocks may include determining a standard for M _sub second delay time samples included in each of the plurality of second sub- And determining the second sub-block as a valid second sub-block if the deviation is smaller than the predetermined threshold value.

In one embodiment, constructing a plurality of first samples and a plurality of second samples from the plurality of first sample mean values and the plurality of second sample mean values may comprise: Setting the values of the plurality of first samples and the values of the plurality of first samples to the values of the vertical axis coordinates of the first group of points, and determining whether the at least one outlier is present among the first group of points (RANSAC) algorithm using the RANSAC ( Random Sample Consensus ) algorithm.

In one embodiment, the step of constructing the plurality of third samples and the plurality of fourth samples from the plurality of third sample average values and the plurality of fourth sample average values may further comprise: Setting values of a plurality of fourth samples and setting values of the vertical axis coordinates of the second group of points to values of the plurality of third samples and determining whether the at least one outlier is among the points of the second group And determining whether the RANSAC algorithm is used.

In one embodiment, defining a third group of points in a two-dimensional coordinate system using the plurality of first samples and the plurality of second samples and defining the points of the third group as a first one-dimensional function , And applying a curve fitting algorithm to the third group of points.

In one embodiment, defining a fourth group of points in a two-dimensional coordinate system using the plurality of third samples and the plurality of fourth samples and defining the points of the fourth group as a second one- , And applying a curve fitting algorithm to the fourth group of points.

In one embodiment, estimating the clock frequency error using the first slope value, the second slope value, the first standard deviation, and the second standard deviation comprises:

는 상기 제1 표준 편차를 나타내며,

는 상기 제2 표준 편차를 나타냄 - 에 의해 상기 클럭 주파수 오차를 추정하는 단계를 포함할 수 있다.Wherein E _Freq represents the clock frequency error, f _t2t1 represents the first slope value, f _t4t3 represents the second slope value,

Represents the first standard deviation,

And estimating the clock frequency error by the second standard deviation.

In one embodiment, estimating the clock time error using the first mean value to the fourth mean value, the estimated clock frequency error, and the current time value,

Wherein E _Time represents the clock time error, E _Freq represents the clock frequency error, D _t2t1 represents the first average value, D _t4t3 represents the third average value, T ₁ represents the second average value, T ₄ represents the fourth average value, T _current represents the current time, And estimating the clock time error by the clock time error.

In one aspect, an apparatus is provided for estimating a clock frequency error in a PTP slave. The apparatus comprises a plurality of first delay time samples and a plurality of transmission time samples each associated with a plurality of sink messages, a plurality of second delay time samples and a plurality of reception time samples, Wherein the plurality of transmission time samples are each associated with the plurality of first delay time samples and the plurality of reception time samples are respectively associated with the plurality of second delay time samples, A selected portion of the plurality of first delay time samples, a selected portion of the plurality of second delay time samples and a portion of the plurality of transmission time samples associated with the selected portion of the plurality of first delay time samples, Based on some of the plurality of received time samples associated with a selected portion of the plurality of second delay time samples And a clock frequency error estimator configured to estimate the clock frequency error.

In one aspect, an apparatus is provided for estimating a clock frequency error in a PTP slave. The present apparatus determines a plurality of transmission time samples by identifying a plurality of transmission times from which a plurality of sync messages are transmitted from the PTP master and calculates delay times of the plurality of sync messages from the PTP master to the PTP slave Determining a plurality of first delay time samples, determining a plurality of second delay time samples by calculating delay times of a plurality of delay request messages from the PTP master to the PTP master, A delay detector configured to identify a plurality of reception time samples arriving at a PTP master to determine a plurality of reception time samples, the plurality of transmission time samples being each associated with the plurality of first delay time samples, Are respectively associated with the plurality of second delay time samples, and the delay detector Each cycle N _sub of the first delay time samples, N _sub transmit time samples, N _sub of second delay samples, and is configured to output the N _sub reception time samples, the N _sub is a natural number - and the selection The N _sub first delay time samples and the N _sub transmission time samples from the delay detector for each period, and to output N _ran first sample average values and N _ran second sample average values, filter-one, each cycle of the selected receiving the N _sub of second delay samples and the N _sub reception time samples from the delay detection unit N _ran of third sample average value and the N _ran - the N _ran is a natural number A second sub-block filter configured to output a fourth sample mean value, a second sub-block filter configured to output N _ran first sample mean values and N _ran second sample mean values from the first sub- Power take M _ran of first samples and M _ran of the claim configured to provide a second sample 1 RANSAC filter, the M _ran of the first group in the first sample with the M _ran of the two-dimensional coordinate system using the two-sample the defining point, the first defining first group points to the first one-dimensional functions, the first one-dimensional function, a first least-squares curve fitting configured to determine a slope of a first inclination value of the calculation unit and the M _ran of A first sub-block selection and block averaging unit including a first delay and a time averaging unit configured to determine a standard deviation for a first sample as a first standard deviation, the M _ran being a natural number equal to or less than the N _ran - the second sub-block receives the filter input to N _ran of third sample average value and the N _ran of the fourth sample average value from M _ran of third sample and the second RANSAC filter, the adapted to provide M _ran of the fourth sample The second one-dimensional definition of M _ran of the points of the second group in a two-dimensional coordinate system using the three samples and the M _ran of the fourth sample, and define the points of the second group with a second one-dimensional functions, and A second least squares curve fitting calculator configured to determine a slope of the function as a second slope value and a second delay and a time averaging unit configured to determine a standard deviation of the M _ran third samples as a second standard deviation A frequency error and a time error configured to estimate the clock frequency error using the first slope value, the second slope value, the first standard deviation, and the second standard deviation, And an estimation unit.

In one embodiment, the first sub-block filters, and the N _sub of the alignment is configured the size of the first delay time samples to select a small M _sub of first delay samples and selection unit, wherein the M _sub is the N count register for storing, rollover de delay storage unit, the count value (R _cnt) - - less than _sub natural number initial value of the count value (R _cnt) being stored in the count register is 0 -, and the M _sub of first case to validate the delay sample the M _sub of it is determined that the first delay sample is valid, the M _sub of first determining a take an average of the first sample mean value for the delay sample, wherein effective said M _{sub-determined} number of first delay taking the average of the M _sub transmit time samples associated with the time samples and determining the second sample mean value, the M number of first _sub delay samples If it is determined not valid the M _sub of storing a first delay samples in the roll-over de delay storage unit, and may include a standard deviation checking unit configured to increase a count value (R _cnt) of the count register by one , which the sorting and selecting unit, the following in response to which the N _sub of first delay sample input to the period by checking the count value (R _cnt) of the count register, the count value of the count register (R _cnt) 0 The first delay time sample of N _sub that is input in the next period and the first delay time sample of M _{sub that} is the smallest among the M _sub first delay time samples stored in the rollover delay storage unit Lt; / RTI >

In one embodiment, the standard deviation inspection unit, the M _sub of claim may be further arranged that the M _sub of first delay sample is less than the threshold standard deviation is selected for the first delay time samples so as to determine that the effective have.

In one embodiment, the first RANSAC filter, the N _ran one of the first sample average value and the N _ran of second sample mean value using the defined points of the third group in a two-dimensional coordinate system, and the third group If there is at least one outlier from the points to each removal of the first sample average value and the second sample average value corresponding to the at least one outlier from the N _ran of first sample average value and the N _ran of second sample mean value _{It ran} the M number of first samples, and to provide a second sample of the M _ran can be further adapted.

In one embodiment, the first delay time and the average unit configuration further determines the mean value for the M _ran of the first sample as the first average value, and to determine a mean value for the M _ran of the second sample into a second mean value may be, the second delay time and the average unit determines the mean value for the M _ran of third samples as the third average value and is further configured to determine an average value for the three fourth sample the M _ran as a fourth average value And the frequency error and time error estimator may be further configured to estimate a clock time error using a value indicating the first to fourth mean values, the estimated clock frequency error, and the current time.

In one embodiment, the frequency error and time error estimator include:

는 상기 제1 표준 편차를 나타내며,

는 상기 제2 표준 편차를 나타냄 - 에 의해 상기 클럭 주파수 오차를 추정하도록 더 구성될 수 있다.Wherein E _Freq represents the clock frequency error, f _t2t1 represents the first slope value, f _t4t3 represents the second slope value,

Represents the first standard deviation,

May be further configured to estimate the clock frequency error by: < EMI ID = 1.0 >

In one embodiment, the frequency error and time error estimator include:

Wherein E _Time represents the clock time error, E _Freq represents the clock frequency error, D _t2t1 represents the first average value, D _t4t3 represents the third average value, T ₁ represents the second average value, T ₄ represents the fourth average value, T _current represents the current time, Lt; RTI ID = 0.0 > time-error < / RTI >

According to embodiments of the present invention, only some of the delay time samples selected in accordance with the size and homogeneity of the samples for the packet delay time between the PTP master and the PTP slave are used for clock frequency error and clock time error estimation, There is a technical effect that accuracy of clock frequency error and clock time error estimation and calculation speed can be increased.

1 is a diagram illustrating that a PTP slave exchanges messages with a PTP master to perform clock synchronization with a PTP master in accordance with the PTP protocol.
2 is a block diagram of an embodiment of an apparatus for estimating clock frequency error and clock time error based on PTP according to the present invention.
3 is a block diagram of a sub-block filter according to an embodiment of the present invention.
FIG. 4 is a block diagram illustrating a sub-block selection and block averaging unit according to an embodiment of the present invention. Referring to FIG.
5 is a block diagram of a frequency and time error estimator according to an embodiment of the present invention.
6 is a flowchart illustrating an embodiment of a method for estimating a clock frequency error and a clock time error based on PTP according to the present invention

BRIEF DESCRIPTION OF THE DRAWINGS The advantages and features of the present invention and the manner of attaining them will become apparent with reference to the embodiments described in detail below with reference to the accompanying drawings. The present invention may, however, be embodied in many different forms and should not be construed as limited to the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the concept of the invention to those skilled in the art. To fully disclose the scope of the invention to a person skilled in the art, and the invention is only defined by the scope of the claims.

The terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the present invention. For example, an element expressed in singular < Desc / Clms Page number 5 > terms should be understood to include a plurality of elements unless the context clearly dictates a singular value. In addition, in the specification of the present invention, it is to be understood that terms such as "include" or "have" are intended to specify the presence of stated features, integers, steps, operations, components, The use of the term does not exclude the presence or addition of one or more other features, numbers, steps, operations, elements, parts or combinations thereof. Further, in the embodiments described herein, 'module' or 'sub-unit' may mean at least one function or a functional part performing an operation.

In addition, all terms used herein, including technical or scientific terms, unless otherwise defined, have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs. Terms such as those defined in commonly used dictionaries should be construed as meaning consistent with meaning in the context of the related art and may be interpreted in an ideal or overly formal sense unless explicitly defined in the specification of the present invention It does not.

Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings. In the following description, well-known functions or constructions will not be described in detail if they obscure the subject matter of the present invention.

1 is a diagram illustrating that a PTP slave exchanges messages with a PTP master to perform clock synchronization with a PTP master in accordance with the PTP protocol.

The IEEE 1588 system, which is a method for adjusting clock frequency / time synchronization on a packet network, includes a PTP master 110 and a PTP master 110, which transmit packets containing time information to all nodes of the network. And a PTP slave 120 for synchronizing the clock of the master using the transmitted timing packets. The PTP master 110 receives the time information from GPS (Global Positioning System) or an accurate clock source and uses it as a base clock in the IEEE 1588 algorithm, and the PTP slave 120 includes the time packet received from the PTP master 110 The local time is adjusted to the clock of the master. In one embodiment, the PTP slave 120 may be a small cell such as a microcell, a picocell, a femtocell, or the like.

In the IEEE 1588 system, the PTP master 110 and the PTP slave 120 are generally synchronized through the offset correction process and the transmission delay measurement process.

The offset correction process corrects a clock time error between the PTP master 110 and the PTP slave 120. For this purpose, the PTP master 110 periodically For example, every 2 seconds, to the PTP slave 120. The Sync message, which is a synchronization message, is transmitted to the PTP slave 120 every 2 seconds. In this case, the PTP master 110 measures the precise time at which the sink message is transmitted using hardware or software, and transmits the measured transmission time to the PTP slave 120 again using a follow-up message do. The PTP slave 120 measures the reception time of the sink message using hardware or software, and calculates the delay time of the sink message based on the transmission time and the reception time of the sink message.

In order to measure the transmission delay between the PTP master 110 and the PTP slave 120, the PTP slave 120 periodically transmits a delay request message to the PTP master 110 as shown in FIG. 1 The PTP master 110 measures the reception time of the delay request message using hardware or software and informs the PTP slave 120 of the reception time of the measured delay request message through a delay response message. The PTP slave 120 receiving the delay response message measures the transmission delay of the network based on the transmission time of the delay request message and the reception time of the delay request message.

The PTP slave 120 may determine the time at which the PTP master 110 transmits the sink message, the time at which the PTP slave 120 receives the sink message, the time at which the PTP slave 120 transmits the delay request message, Can determine the clock time error and the transmission delay time between the PTP master and the PTP slave as shown in the following equations (1) and (2) using the time when the delay request message is received.

T1, t2, t3 and t4 are the time when the PTP master 110 transmitted the sink message, the time when the PTP slave 120 received the sink message, and the time when the PTP slave 120 The time at which the delay request message is transmitted and the time at which the PTP master 110 received the delay request message, respectively.

Where TD represents the transmission delay time.

Since the PTP is based on a wired packet network, a sink message transmitted from the PTP master 110 and a delay request message transmitted from the PTP slave 120 necessarily have a packet delay time and the delay time is not constant (i.e., There is a packet delay variance (PDV)). For this reason, it is difficult to accurately estimate the clock frequency error and the clock time error with Equation (1) above, and a large number of packet samples are required for the accurate estimation.

In the exemplary embodiments of the present invention described below, all of the samples of the packet delay times (the sync packet delay time and the delay request packet delay time) between the PTP master 110 and the PTP slave 120, There is provided a method of selecting a specific sample according to size and homogeneity from the above samples and estimating a clock frequency error and a clock time error using the frequency samples and the clock time errors instead of estimating the frequency error and the clock time error.

2 is a block diagram of an embodiment of an apparatus for estimating clock frequency error and clock time error based on PTP according to the present invention.

2, the apparatus 200 identifies a plurality of transmission times tm1 (i) from which a plurality of sync packets have been transmitted from the PTP master 110 to generate a plurality of transmission time samples t1 (i) And determines from the PTP master 110 to the PTP slave 120 based on the plurality of transmission times tm1 (i) and the times tm2 (i) at which the plurality of sync packets are received in the PTP slave 120, And a delay detector 210 configured to calculate delay times of the plurality of sync packets up to the first delay time _dt2t1 (i) and determine a plurality of first delay time samples _dt2t1 (i) according to Equation (3) below.

Here, d _t2t1 (i), tm2 (i), tm1 (i), t1 (i), and i is a first delay time samples, the sync packet received from the PTP the slave time, sending a sync packet transmitted from the PTP master Time, transmission time sample, and sample index, respectively.

In one embodiment, identification of a plurality of transmission times tm1 (i) may be accomplished by identifying the timestamps included in the sync messages. In another embodiment, the identification of the plurality of transmission times tm1 (i) may be accomplished by identifying the timestamps included in the follow-up messages.

The delay detection unit 210 identifies a plurality of reception times tm4 (i) at which a plurality of delay request packets transmitted from the PTP slave 120 arrive at the PTP master 110 and generates a plurality of reception time samples t4 ) From the PTP slave 120 based on the transmission times tm3 (i) and the plurality of reception times tm4 (i) at which the plurality of delay request packets are transmitted from the PTP slave 120, To determine a plurality of second delay time samples (d _t4t3 (i)) by calculating delay times of a plurality of delay request packets up to the master (110).

Here, d _t4t3 (i), tm4 (i), tm3 (i), t4 (i), and i is a second delay time samples, delay request packet is the reception time, the delay request packets reaches a PTP master in PTP slave The transmitted time, the received time sample, and the sample index, respectively.

In one embodiment, identification of the plurality of reception times tm4 (i) may be accomplished by identifying the timestamps contained in the plurality of delay response messages transmitted by the PTP master 110. [ Wherein the plurality of transmission time samples t1 (i) are each associated with the plurality of first delay time samples _dt2t1 (i), and the plurality of reception time samples t4 (i) And a second delay time sample _dt4t3 (i), respectively. The delay detector 210 includes N _sub (natural number) first delay time samples d _t2t1 (i), N _sub transmission time samples t1 (i), N _sub second delay time samples (d _t4t3 (i)) and N _sub receive time samples t4 (i). In one embodiment, the predetermined period may be a period corresponding to a clock period of the system clock (not shown) or a multiple of the system clock (not shown) inside the apparatus 200.

및 N_ran 개의 제2 샘플 평균값

을 출력하도록 구성될 수 있다. 제2 서브블록 필터(224)는 N_sub 개의 제2 지연 시간 샘플(d_t4t3(i)) 및 N_sub 개의 수신 시간 샘플(t4(i))을 입력 받아 그 중에서 동질성 요건을 만족하는 소정의 샘플들을 선별하여 선별된 샘플들을 평균화하는 기능을 수행하도록 구성될 수 있다. 즉, 제2 서브블록 필터(224)는 선정된 주기 마다 딜레이 검출부(210)로부터 N_sub 개의 제2 지연 시간 샘플(d_t4t3(i)) 및 N_sub 개의 수신 시간 샘플(t4(i))을 입력 받아 N_ran 개의 제3 샘플 평균값

및 N_ran 개의 제4 샘플 평균값

을 출력하도록 구성될 수 있다.The apparatus 200 may further include a first sub-block filter 222 and a second sub-block filter 224 connected to the delay detection unit 210. The first sub-block filter 222 receives N _sub first delay time samples d _t2t1 (i) and N _sub transmission time samples t1 (i) And averaging the selected samples. That is, the first sub-block filter 222 receives N _sub first delay time samples d _t2t1 (i) and N _sub transmission time samples t1 (i) from the delay detector 210 at predetermined intervals The first sample average value N _ran (natural number)

And N _ran second sample average values

As shown in FIG. The second sub-block filter 224 receives N _sub second delay time samples d _t4t3 (i) and N _sub reception time samples t4 (i) And averaging the selected samples. That is, the second sub-block filter 224 receives N _sub second delay time samples d _t4t3 (i) and N _sub reception time samples t4 (i) from the delay detector 210 every predetermined period The third sample average value of N _ran

And N _ran fourth sample average values

As shown in FIG.

및 N_ran 개의 제2 샘플 평균값

을 입력받아 제1 기울기값(f_t2t1), 제1 표준 편차

제1 평균값(D_t2t1) 및 제2 평균값(T₁)을 출력하도록 구성될 수 있다. 제2 서브블록 선택 및 블록 평균부(254)는 제2 서브블록 필터(224)로부터 N_ran 개의 제3 샘플 평균값

및 N_ran 개의 제4 샘플 평균값

을 입력받아 제2 기울기값(f_t4t3), 제2 표준 편차

,제3 평균값(D_t4t3) 및 제4 평균값(T₄)을 출력하도록 구성될 수 있다.The apparatus 200 includes a first subblock selection and block averaging unit 252 and a second subblock selection and block averaging unit 254 connected to the first subblock filter 222 and the second subblock filter 224, ). Each of the first sub-block selection and block averaging unit 252 and the second sub-block selection and block averaging unit 254 uses the remaining selected sample average values excluding the singular value (s) among the plurality of sample average values to be input And calculate a mean value and a standard deviation for the slope values. The first sub-block selection and block averaging unit 252 receives N _ran first sample average values

And N _ran second sample average values

And receives a first slope value f _t2t1 , a first standard deviation

And may output the first average value D _t2t1 and the second average value T ₁ . The second sub-block selection and block averaging unit 254 receives N _ran third sample average values

And N _ran fourth sample average values

And receives a second slope value (f _t4t3 ), a second standard deviation

, A third average value (D _t4t3 ), and a fourth average value (T ₄ ).

및 제2 표준 편차

를 이용하여 클럭 주파수 오차(E_Freq)를 추정하도록 구성될 수 있다. 주파수 오차 및 시간 오차 추정부(270)는 제1 평균값 내지 제4 평균값(D_t2t1, T₁, D_t4t3, T₄), 추정된 클럭 주파수 오차(E_Freq) 및 현재 시간을 나타내는 값(T_current)을 이용하여 클럭 시간 오차(E_Time)를 추정하도록 더 구성될 수 있다.The apparatus 200 may further include a frequency error and time error estimator 270 connected to the first subblock selection and block averaging unit 252 and the second subblock selection and block averaging unit 254. The frequency error and time error estimator 270 estimates the frequency error and the time error based on the first slope value f _t2t1 , the second slope value f _t4t3 ,

And a second standard deviation

To estimate the clock frequency error (E _Freq ). Frequency offset and timing offset estimation unit 270, a first average through fourth average value _{(D t2t1, T 1, D} t4t3, T 4), indicating the estimated clock frequency error (E _Freq), and the current time value (T _current ) To estimate the clock time error (E _Time ).

3 is a block diagram of a sub-block filter according to an embodiment of the present invention.

을 결정하고, 아래 수학식 8에 따라 유효하다고 결정된 M_sub 개의 제1 지연 시간 샘플(d'_t2t1(i))에 연관된 M_sub 개의 송신 시간 샘플(t1'(j))에 대해 평균을 취해 제2 샘플 평균

을 결정하도록 구성될 수 있다.3, the first sub-block filter 222 may include an alignment and selection unit 310, a standard deviation checking unit 330, and a rollover delay storage unit 350. The first sub-block filter 222 may further include a count register (not shown). The count register is used to store the count value ( _Rcnt ) and its initial value can be set to zero. Sorting and selection unit 310 N _sub of the first delay time samples (d _t2t1 (i)), the sorted according to size the small M _sub of first delay size sample of which (d _'t2t1 (i) ). ≪ / RTI > Where M _sub is a natural number less than N _sub . In one embodiment, M _sub may be set to about 20-30% of N _sub . Standard deviation checking unit 330 if it is determined that M _sub of the first delay time samples (d _'t2t1 (i)) to validate M _sub of first delay sample of (d' _t2t1 (i)) is valid, The average of the first delay time samples d ' _t2t1 (i) of M _{sub is calculated} according to the following equation (7)

Taking the average for the determination, and determined to be valid in accordance with the equation (8) below the M _sub of the first delay time samples _{(d 't2t1 (i))} M sub transmit time samples (t1 associated with the' (j)) the 2 sample averages

. ≪ / RTI >

In one embodiment, if the standard deviation for the M _sub first delay time samples (d ' _t2t1 (i)) is less than the predetermined threshold, then the standard deviation checker 330 determines M _sub first delay time samples d ' _t2t1 (i)) is valid. In one embodiment, the selected threshold may be a time error limit set by the user. The time error limit may be set based on a range of time errors that must be satisfied in implementing the clock time error estimation apparatus 200 of the present invention. For example, if the apparatus 200 of the present invention is based on an LTE TDD system, the time error threshold may be 3 microseconds. On the other hand M _sub of the case that are not right the first delay time samples (d _'t2t1 (i)) is valid, the standard deviation checking unit 330 first samples the average value and a without determining the second sample mean value M _sub of first storing the delay time samples (d _'t2t1 (i)) to the roll-over de delay storage unit 350, and may be further configured to increase a count value (R _cnt) of the count register by one. As will be described later, the first delay time samples stored in the rollover delay storage unit 350 are reused for the calculation in the first sub-block filter 222 in the next period.

The sorting and selecting unit 310 may be configured to check the count value _Rcnt stored in the count register in the next cycle. The count value _Rcnt of the count register is not valid because the M _sub first delay time samples d ' _t2t1 (i) selected by the sorting and selecting unit 310 in the previous period (s) (350). As described above, the initial value of the count value _Rcnt is 0, which is determined by the standard deviation checker 330 as being valid for M _sub first delay time samples d ' _t2t1 (i) in the corresponding period It can be reinitialized to zero whenever it is. On the other hand, each time the standard deviation checking unit 330 determines that the first delay time samples d ' _t2t1 (i) of M _sub of the corresponding period are not valid and transmits the same to the rollover delay storage unit 350, the count value R _cnt ) may be incremented by one. The count value (R _cnt ) can be initialized to 0 again when the value is equal to the maximum rollover delay count (R _MAX ) set by the user. The reason why the R _MAX value is set to initialize the count value R _cnt is that the value of a first delay time sample abnormally approaches zero or becomes negative due to an occasional abnormal time stamp value This value is repeatedly selected by the sorting and selecting unit 310 over several cycles and the first delay time samples d ' _t2t1 (i) of M _sub selected as this value are _validated by the standard deviation checking unit 330 And is stored in the rollover delay storage unit 350, so that the operation in the first sub-block filter 222 may not be terminated.

및 제2 샘플 평균값

을 결정하거나 M_sub 개의 제1 지연 시간 샘플(d'_t2t1(i))을 롤오버드 딜레이 저장부(350)에 저장하고 카운트 레지스터의 카운트 값(R_cnt)을 1만큼 증가시킬 수 있다.Sorting and selection unit 310, when the count value of the count register (R _cnt) is determined to be non-0 N _sub input to the periodic first delay time samples (d _t2t1 (i)) and rollover de delay in consideration with the M _sub of the first delay time samples (d _'t2t1 (i)) stored in the storage unit 350, that is, the small the size from that in consideration of the total N _sub + M _sub of first delay sample And to select M _sub first delay time samples (d ' _t2t1 (i)). However, if it is determined that the count value _Rcnt of the count register is 0, the sorting and selecting unit 310 considers only the N _sub first delay time samples d _t2t1 (i) (D ' _t2t1 (i)) of the smallest M _sub among the first delay time samples (d' _t2t1 (i)). As in the first period, the standard deviation checking unit 330 checks the validity of M _sub first delay time samples d ' _t2t1 (i) in the corresponding period and determines whether the first sample average value

And the second sample average value

A decision or may store the samples of first delay (d _'t2t1 (i)) to the roll-over M _sub de delay storage unit 350, and increases the count value (R _cnt) of the count register by one.

및 N_ran 개의 제2 샘플 평균값

이 모두 결정되면 이를 출력하도록 구성될 수 있다.In this manner, the first sub-block filter 222 _extracts N _sub first delay time samples d _t2t1 (i) and N _sub transmission time samples t1 (i) from the delay detector 210 every predetermined period And repeatedly performs the above-described operations. As a result, N _ran first sample average values

And N _ran second sample average values

Can be configured to output it.

및 N_ran 개의 제4 샘플 평균값

을 출력하도록 구성될 수 있다. 설명의 반복을 피하기 위해 제2 서브블록 필터(224)에 관한 상세한 설명은 생략한다.The second sub-block filter 224 receives N _sub second delay time samples d _t4t3 (i) and N _sub reception time samples t4 (i) 222 in the same manner. That is, the second sub-block filter 224 receives N _sub second delay time samples d _t4t3 (i) and N _sub reception time samples t4 (i) from the delay detector 210 every predetermined period And performs the same operation as in the first sub-block filter 222 to obtain N _ran third sample average values

And N _ran fourth sample average values

As shown in FIG. A detailed description of the second sub-block filter 224 is omitted in order to avoid repetition of description.

FIG. 4 is a block diagram illustrating a sub-block selection and block averaging unit according to an embodiment of the present invention. Referring to FIG.

및 N_ran 개의 제2 샘플 평균값

을 입력 받아 M_ran 개의 제1 샘플

및 M_ran 개의 제2 샘플

을 제공하도록 구성될 수 있다. 여기서, M_ran은 N_ran 보다 작거나 같은 자연수이다. RANSAC 필터(410)는 RANSAC 알고리즘에 기반한 처리를 수행하도록 구현될 수 있다. 즉, RANSAC 필터(410)는, N_ran 개의 제1 샘플 평균값

과 N_ran 개의 제2 샘플 평균값

을 이용하여 2차원 좌표계에서의 일 군의 점들을 정의하고 이 일 군의 점들 중에서 적어도 하나의 아웃라이어(outlier)가 있는 경우 상기 적어도 하나의 아웃라이어에 해당하는 제1 샘플 평균값과 제2 샘플 평균값을 N_ran 개의 제1 샘플 평균값

및 N_ran 개의 제2 샘플 평균값

으로부터 각각 제거하여 M_ran 개의 제1 샘플

및 M_ran 개의 제2 샘플

을 제공하도록 구성될 수 있다.4, the first sub-block selection and block averaging unit 252 includes a Random Sample Consensus (RANSAC) filter 410, a minimum square curve fitting calculator 420, a delay and time averaging unit 430, . ≪ / RTI > RANSAC filter 410 receives N _ran first sample mean values < RTI ID = 0.0 >

And N _ran second sample average values

The first sample of M _ran

And M _ran second samples

. ≪ / RTI > Here, M _ran is a natural number less than or equal to N _ran . The RANSAC filter 410 may be implemented to perform processing based on the RANSAC algorithm. In other words, the RANSAC filter 410 calculates N _ran first sample average values

And N _ran second sample average values

Dimensional coordinate system and if there is at least one outlier among the points of the group, a first sample average value corresponding to the at least one outlier and a second sample mean value corresponding to the at least one outlier, To N _ran first sample average values

And N _ran second sample average values

Respectively, to obtain M _ran first samples

And M _ran second samples

. ≪ / RTI >

중 어느 한 제1 샘플 평균값과 N_ran 개의 제2 샘플 평균값

중 위 어느 한 제1 샘플 평균값에 연관된 제2 샘플 평균값의 쌍일 수 있다. 데이터 세트는 총 N_ran 개의 데이터를 포함한다. RANSAC 필터(410)는 다음과 같은 연산 처리를 수행할 수 있다: 데이터 세트에서 원하는 함수를 구하는데 필요한 최소 개수의 데이터를 선택한다. 본 발명의 실시예에서와 같이 원하는 함수가 직선인 경우 데이터 세트에서 2개의 데이터를 선택할 수 있다. 선택한 데이터를 이용해 함수를 구한다. 이 함수로부터 나머지 데이터들이 떨어진 거리들을 구하고 이 거리들 중, 예컨대 사용자에 의해 설정되는 거리(T)보다 큰 거리가 있는 경우 해당 거리에 대응하는 데이터를 아웃리어로 판단하고 데이터 세트에서 제외하고 데이터 세트에 남아 있는 데이터의 수를 기록한다. 이런 식으로 임의의 데이터(2개의 데이터)를 선택하고 함수를 구한 후 아웃라이어가 아닌 데이터의 수를 기록하는 것을 N회 반복 수행하며 - N은 반복 횟수임 -, 반복 수행 후 아웃라이어가 가장 적은 함수를 찾아내어 이 때의 인라이어들(inliers)을 최종 데이터로 확정한다.The RANSAC filter 410 may be configured to determine outliers in the data set based on the RANSAC algorithm. Here, each data included in the data set includes N _ran first sample average values

And N _ran second sample average values

May be a pair of second sample mean values associated with any one of the first sample mean values. The data set includes a total of N _ran data. The RANSAC filter 410 may perform the following operation: Select the minimum number of data needed to obtain the desired function in the data set. If the desired function is a straight line, as in an embodiment of the present invention, two data can be selected in the data set. The function is obtained using the selected data. If there is a distance greater than the distance (T) set by the user among these distances, the data corresponding to the distance is determined as an outlier and excluded from the data set, The number of data remaining in the recording medium is recorded. After selecting arbitrary data (2 data) in this way and finding the function, it is repeated N times to record the number of data, not the outlier. - N is the number of repetitions - Function is found and the inliers at this time are determined as final data.

It is necessary to set the number of iterations N and the distance T when implementing the RANSAC filter 410. [ The number N of repetitions can be expressed by Equation (9).

Where N, p, w, and n represent the number of repetitions, the probability that all data other than the data used to determine the function are in-liers, the probability that the given data is in-liers and the number of data .

Therefore, for example, when assuming that p = 0.99, w = 0.5, and n = 2, the number N of repetitions can be set to 16. [

The distance T can be calculated by a frequency error limit set by the user. The frequency error threshold may be set based on a range of clock frequency errors that must be satisfied in implementing the clock frequency error estimation apparatus 200 of the present invention. For example, if the apparatus 200 of the present invention is based on an LTE system, the frequency error threshold may be 250 ppb (parts per billion). Since the clock frequency error is the magnitude of the delay (tm2 - tm1 or - (tm4 - tm3)) that changes over a specific time, the following equation (10) must be satisfied.

이고

이다.Here, the Frequency Error Limit represents a frequency error limit value, DELTA delay is a size of a first sample mean value (insert data) that varies with time,

ego

to be.

이 선택된 인라이어이고

이 최소값이며

이 최대값이며 두 데이터만 가지고도 구할 수 있는 1차원 함수가 RANSAC 필터(410) 후단의 최소자승 곡선 맞춤 산출부(420)에서 구한 1차원 함수와 같을 때이다. 이 경우에서 실제 클럭 주파수 오차를 0 ppb라 가정하면

과

의 차이는 2 * 거리(T)이다. 따라서, RANSAC 필터(410)를 구현할 때 거리(T)를 아래의 수학식 11 및 수학식 12와 같이 정리할 수 있다.DELTA DELTA DELTA DELTA DELTA DELTA DELTA DELTA DELTA DELTA DELTA DELTA DELTA DELTA DELTA DELTA DELTA DELTA DELTA DELTA DELTA DELTA DELTA DELTA DELTA DELTA DELTA DELTA DELTA DELTA DELTA DELTA DELTA DELTA DELTA DELTA DELTA DELTA DELTA DELTA DELTA If Δ delay has the maximum value

Is the chosen inlier

Is the minimum

Dimensional function obtained by the minimum square curve fitting calculator 420 at the rear end of the RANSAC filter 410 is the same as the one-dimensional function obtained by using only the two data. Assuming that the actual clock frequency error is 0 ppb in this case

and

Is 2 * distance (T). Therefore, when implementing the RANSAC filter 410, the distance T can be expressed by Equation (11) and Equation (12) below.

및 M_ran 개의 제2 샘플

을 출력할 수 있다.RANSAC filter 410 by performing the RANSAC algorithm in the number of iterations (N) and distance (T) is set as described above M number of first samples _ran

And M _ran second samples

Can be output.

과 M_ran 개의 제2 샘플

을 이용하여 2차원 좌표계에서의 일 군의 점들을 정의하고, 이 일 군의 점들을 1차원 함수로 정의하고, 이 1차원 함수의 기울기를 제1 기울기 값(f_t2t1)으로서 결정하도록 구성될 수 있다. 일 실시예에서, 라인 피팅 알고리즘(line fitting algorithm)을 이용하여 일 군의 점들을 1차원 함수로 정의할 수 있다. 시간에 따른 딜레이의 변화량이 클럭 주파수 오차이므로 제1 기울기 값(f_t2t1)은 PTP 매스터(110)에서 PTP 슬레이브(120)까지의 딜레이 변화량에 의해 추정한 클럭 주파수 오차가 된다.Least-squares curve fit calculation section 420 _ran M number of first samples

And M _ran second samples

Dimensional coordinate system, defining points of the group in the two-dimensional coordinate system as a one-dimensional function, and determining the slope of the one-dimensional function as the first slope value (f _t2t1 ) have. In one embodiment, a group of points can be defined as a one-dimensional function using a line fitting algorithm. The first slope value f _t2t1 is a clock frequency error estimated by the amount of delay change from the PTP master 110 to the PTP slave 120 since the variation of the delay with time is a clock frequency error.

에 대한 표준 편차를 제1 표준 편차

로서 결정하고, M_ran 개의 제1 샘플

에 대해 평균을 취해 제1 평균값(D_t2t1)을 출력하고, M_ran 개의 제2 샘플

에 대해 평균을 취해 제2 평균값(T₁)을 출력하도록 구성될 수 있다.And delay time-average unit 430 _ran M number of first samples

The first standard deviation < RTI ID = 0.0 >

And M _ran of the first samples

And outputs a first average value (D _t2t1 ), and outputs M _ran second samples

And output a second average value T ₁ .

및 N_ran 개의 제4 샘플 평균값

을 입력 받아 제1 서브블록 선택 및 블록 평균부(252)에서와 동일한 연산 처리를 수행함으로써 제2 기울기값(f_t4t3), 제2 표준 편차

, 제3 평균값(D_t4t3) 및 제4 평균값(T₄)을 출력하도록 구성될 수 있다. 제2 기울기값(f_t4t3)은 PTP 슬레이브(120)에서 PTP 매스터(110)까지의 딜레이 변화량에 의해 추정한 클럭 주파수 오차가 된다. 제2 서브블록 선택 및 블록 평균부(254)의 구성은 제1 서브블록 선택 및 블록 평균부(252)의 구성과 동일하므로 그 상세한 설명은 생략한다.The second sub-block selection and block averaging unit 254 receives N _ran third sample average values

And N _ran fourth sample average values

And performs the same calculation processing as in the first sub-block selection and block averaging unit 252 to obtain a second slope value f _t4t3 , a second standard deviation

, A third average value (D _t4t3 ), and a fourth average value (T ₄ ). The second slope value f _t4t3 is a clock frequency error estimated by the delay variation amount from the PTP slave 120 to the PTP master 110. [ The configuration of the second sub-block selection and block averaging unit 254 is the same as that of the first sub-block selection and block averaging unit 252, and a detailed description thereof will be omitted.

5 is a block diagram of a frequency and time error estimator according to an embodiment of the present invention.

를 입력 받고 제2 서브블록 선택 및 블록 평균부(254)로부터 제2 기울기 값(f_t4t3) 및 제2 표준 편차

를 입력 받아 최종 클럭 주파수 오차(E_Freq)를 아래의 수학식 13에 의해 추정할 수 있다.As shown in FIG. 5, the frequency and time error estimator 270 may include a frequency error estimator 515 and a time error estimator 517. The frequency error estimator 515 receives the first slope value f _t2t1 and the first standard deviation f _t2t1 from the first sub-block selection and block averaging unit 252,

And outputs the second slope value f _t4t3 and the second standard deviation 254 from the second sub-block selection and block averaging unit 254,

And the final clock frequency error E _Freq can be estimated by Equation (13) below.

The fact that the standard deviation is small means that the set of data is more uniform and reliable. Therefore, the weighting factor according to the standard deviation is given to the first slope value f _t2t1 and the second slope value f _t4t3 to obtain the clock frequency error. Since the second slope value f _t4t3 is a frequency error estimated by the amount of delay change from the PTP slave 120 to the PTP master 110, the second slope value f _t4t3 is a value obtained by _subtracting the negative slope value f _t2t1 Lt; / RTI >

The time error estimator 517 receives the clock frequency error E _Freq from the frequency error estimator 515 and receives the first average value D _t2t1 and the first average value D _t2t1 from the first _sub- second average value (T ₁₎ the input received, the second sub-block enable and block average unit 254 may be configured receives from the input of the third average value (D _t4t3) and the fourth average value (T ₄₎ to estimate the final time offset Since the values input from the first and second sub-block selection and block averaging units 252 and 254 are average values, time drift is generated according to the frequency error when the time error is estimated only from these values. It is necessary to correct the time drift due to the frequency error due to the difference between the time T _current at which the calculation is performed in the time error estimator 517 and the time averaged ((T ₁ + T ₄ ) / 2). Therefore, the time error estimating unit 517 calculates the time error according to the following equation (14) The car can be estimated.

Although each of the components / circuits of the device 200 has been described as a separate separate component, all of the components / circuits of the device 200, or a portion thereof, may be referred to as an Application Specific Integrated Circuit (ASIC), a PLD Programmable Logic Device) or FPGA (Field-Programmable Gate Array) design method. In various embodiments, it is also possible to implement the device 200 as a system on chip (SoC) chip for an embedded system.

6 is a flow chart illustrating an embodiment of a method for estimating clock frequency error and clock time error based on PTP in accordance with the present invention.

The method of the present invention is similar to the method of the first embodiment except that the size (N _sub ) of the first and second sub-block filters 222 and 224, the alignment of the first and second sub-block filters 222 and 224, (N _sub ), the size of the first and second sub-block selection and block averaging units 252 and 254 (N _ran ), the maximum number of rollover delays (R _MAX ), the time error threshold and the frequency error threshold (S610). In step S615, the first sub-block filter 222 receives N _sub first delay time samples d _t2t1 (i) and N _sub transmission time samples t1 (i) from the delay detector 210 as Block filter 224 receives N _sub second delay time samples d _t4t3 (i) and N _sub reception time samples t4 (i) from the delay detector 210. The second _sub- In step S620, it is checked whether the count value _Rcnt of the count register is 0 in each of the first sub-block filter 222 and the second sub-block filter 224. If step tests the count value of the count register at (S620) (R _cnt) it is judged to be 0, and each of the first sub-block filter 222 and a second sub-block filter 224 in step (S625) And selects M _sub delay time samples having the smallest delay among N _sub delay time samples. If it is determined in step S620 that the count value _Rcnt of the count register is not 0, it is determined in step S630 whether the first sub-block filter 222 and the second sub-block filter 224 the N selects the M _sub of the sample having the minimum delay among the delays of delay _{sub sub} + M samples.

In step S635, it is checked whether the standard deviation of M _sub delay time samples selected in each of the first sub-block filter 222 and the second sub-block filter 224 is smaller than the time error threshold. If it is determined in step S635 that the standard deviation of the test result is not smaller than the time error threshold value, it is checked in step S640 whether the count value _Rcnt is greater than R _MAX . Step (S640) the count value (R _cnt) is R when it is determined to be greater than _MAX, the count value M _sub of delay samples in the (R _cnt) initialize to zero, and the step (S650) in the step (S645) in Stored in the rollover delay storage unit 350, and the process returns to step S615. Meanwhile, stored in the step (S640) the count value when (R _cnt) is determined that is not greater than R _MAX, rollover de delay storage unit 350, the M _sub of delay samples in step (S650) in the step (S615) Lt; / RTI >

에 대한 제1 샘플 평균값

및 M_sub 개의 제1 지연 시간 샘플(d'_t2t1(i))과 각각 연관된 M_sub 개의 송신 시간 샘플(t1'(j))에 대한 제2 샘플 평균값

을 결정하고 카운트 값(R_cnt)을 0으로 초기화한다. 또한 본 단계에서는 제2 서브블록 필터(224)가 M_sub 개의 제2 지연 시간 샘플(d'_t4t3(i))에 대한 제3 샘플 평균값

및 M_sub 개의 제2 지연 시간 샘플(d'_t4t3(i))과 각각 연관된 M_sub 개의 수신 시간 샘플(t4'(j))에 대한 제4 샘플 평균값

을 결정하고 카운트 값(R_cnt)을 0으로 초기화한다. 단계(S660)에서는 제1 서브블록 필터(222)에서 제1 샘플 평균값

및 제2 샘플 평균값

이 각각 N_ran 개 만큼 결정되었는지의 여부가 검사된다. 또한 본 단계에서는 제2 서브블록 필터(224)에서 제3 샘플 평균값

및 제4 샘플 평균값

이 각각 N_ran 개 만큼 결정되었는지의 여부가 검사된다. 단계(S660)에서의 검사 결과 각 샘플 평균값이 N_ran 개 만큼 결정되지 않은 것으로 판단되는 경우, 프로세스는 단계(S615)로 복귀한다. 한편 단계(S660)에서의 검사 결과 각 샘플 평균값이 N_ran 개 만큼 결정된 것으로 판단되는 경우, 단계(S665)에서 제1 및 제2 서브블록 선택 및 블록 평균부(252, 254)에 의한 처리가 수행되고 단계(S670)에서 주파수 오차 및 시간 오차 추정부(270)에 의해 최종 클럭 주파수 오차 및 최종 클럭 시간 오차가 추정된다.If it is determined in step S635 that the standard deviation is smaller than the time error threshold, in step S655, the first sub-block filter 222 generates M _sub first delay time samples

Lt; RTI ID = 0.0 >

And M _sub transmission time samples t1 '(j) associated respectively with M _sub first delay time samples d' _t2t1 (i)

And the count value _Rcnt is initialized to zero. Also, in this step, the second sub-block filter 224 calculates the third sample average value M d for the M _sub second delay time samples d ' _t4t3 (i)

And M number of second _sub delay time samples (d _'t4t3 (i)) and the M _sub reception sample time (t4 associated with each of the "Fourth-sample average value of the (j))

And the count value _Rcnt is initialized to zero. In step S660, the first sub-block filter 222 samples the first sample average value

And the second sample average value

Are each determined to be N _ran . In this step, in the second sub-block filter 224,

And the fourth sample average value

Are each determined to be N _ran . If it is determined in step S660 that the average value of each sample is not determined by N _ran , the process returns to step S615. On the other hand, if it is determined in step S660 that the average value of each sample is determined to be N _ran , processing by the first and second sub-block selection and block averaging units 252 and 254 is performed in step S665 And the final clock frequency error and the final clock time error are estimated by the frequency error and time error estimator 270 in step S670.

In the above description, the meaning that an element is connected to or coupled to another element means that the element is directly connected or coupled to the other element, as well as the meaning of one or more other elements interposed therebetween Quot; element " may < / RTI > In addition, terms (e.g., 'between', 'between', etc.) for describing the relationship between components should be interpreted in a similar sense.

In the embodiments disclosed herein, the arrangement of the components shown may vary depending on the environment or requirements in which the invention is implemented. For example, some components may be omitted or some components may be integrated into one. In addition, the arrangement order and connection of some components may be changed.

While the present invention has been particularly shown and described with reference to exemplary embodiments thereof, it is to be understood that the invention is not limited to the disclosed exemplary embodiments, but, on the contrary, is intended to cover various modifications and equivalent arrangements included within the spirit and scope of the appended claims. It will be understood by those skilled in the art that various changes in form and details may be made therein without departing from the spirit and scope of the invention. Accordingly, the technical scope of the present invention should be determined only by the appended claims.

110: PTP master
120: PTP slave
200: Clock frequency error and clock time error estimation device
210: delay detector
222: first sub-block filter
224: second sub-block filter
252: First sub-block selection and block averaging
254: second sub-block selection and block averaging
270: frequency error and time error estimating unit
310: alignment and selection unit
330: Standard deviation checker
350: Rollover delay storage unit
410: RANSAC filter
420: least square curve fitting calculation unit
430: delay and time averaging
515: Frequency error estimator
517: time error estimating unit

Claims (22)

A method for estimating a clock frequency error in a PTP slave (Precision Time Protocol slave)
Configuring a plurality of first delay time samples by determining a plurality of first delay times respectively associated with a plurality of sync messages sequentially transmitted from a PTP master to reach the PTP slave,
Identifying a plurality of transmission times of a plurality of sink messages sequentially transmitted from the PTP master to construct a plurality of transmission time samples, each of the plurality of transmission time samples being associated with the plurality of first delay time samples, ,
Configuring a plurality of second delay time samples by determining a plurality of second delay times each time a plurality of delay request messages sequentially transmitted from the PTP slave arrive at the PTP master,
Identifying a plurality of reception times in which a plurality of delay request messages sequentially transmitted from the PTP slave arrive at the PTP master to construct a plurality of reception time samples, Each associated with a delay time sample -, and
Based on a selected portion of the plurality of first delay time samples, a selected portion of the plurality of transmission time samples, a selected portion of the plurality of second delay time samples, and a selected portion of the plurality of receive time samples, And estimating a clock frequency error. The method according to claim 1,
Based on a selected portion of the plurality of first delay time samples, a selected portion of the plurality of transmission time samples, a selected portion of the plurality of second delay time samples, and a selected portion of the plurality of receive time samples, The step of estimating
Averaging each of a plurality of first delay time sample groups selected from the plurality of first delay time samples to determine a plurality of first sample average values,
Averaging each of a plurality of transmission time sample groups selected from the plurality of transmission time samples and each associated with the plurality of first delay time sample groups to determine a plurality of second sample average values,
Averaging each of a plurality of second delay time sample groups selected from the plurality of second delay time samples to determine a plurality of third sample average values,
Determining a plurality of fourth sample mean values by averaging each of a plurality of receive time sample groups selected from the plurality of receive time samples and associated with each of the plurality of second delay time sample groups,
Constructing a plurality of first samples and a plurality of second samples from the plurality of first sample mean values and the plurality of second sample mean values, wherein configuring the plurality of first samples and the plurality of second samples comprises: A first group of points in a two-dimensional coordinate system is defined using the plurality of first sample average values and the plurality of second sample average values, and if there is at least one outlier among the first group of points, A first sample average value and a second sample average value corresponding to at least one outlier are respectively removed from the plurality of first sample average values and the plurality of second sample average values to obtain the plurality of first samples and the plurality of second samples Comprising the steps of:
Constructing a plurality of third samples and a plurality of fourth samples from the plurality of third sample mean values and the plurality of fourth sample mean values, wherein configuring the plurality of third samples and the plurality of fourth samples comprises: A second group of points in a two-dimensional coordinate system is defined using the plurality of third sample average values and the plurality of fourth sample average values, and if there is at least one outlier among the points of the second group, The third sample average value and the fourth sample average value corresponding to the outlier are respectively removed from the plurality of third sample average values and the plurality of fourth sample average values to construct the plurality of third samples and the plurality of fourth samples Comprising steps < RTI ID = 0.0 >
And estimating the clock frequency error using the plurality of first samples to the plurality of fourth samples. 3. The method of claim 2,
And estimating the clock frequency error using the plurality of first samples to the plurality of fourth samples
Defining points of a third group in a two-dimensional coordinate system using the plurality of first samples and the plurality of second samples and defining the points of the third group as a first one-dimensional function,
Determining a slope of the first one-dimensional function as a first slope value,
Defining points of a fourth group in a two-dimensional coordinate system using the plurality of third samples and the plurality of fourth samples and defining the points of the fourth group as a second one-dimensional function,
Determining a slope of the second one-dimensional function as a second slope value,
Calculating a standard deviation for the plurality of first samples as a first standard deviation,
Calculating a standard deviation for the plurality of third samples as a second standard deviation, and
Estimating the clock frequency error using the first slope value, the second slope value, the first standard deviation, and the second standard deviation. The method of claim 3,
Calculating a mean value for the plurality of first samples as a first mean value,
Calculating an average value for the plurality of second samples as a second average value,
Calculating a mean value for the plurality of third samples as a third mean value,
Calculating a mean value for the plurality of fourth samples as a fourth mean value,
Estimating a clock time error using a value indicating the first mean value to the fourth mean value, the estimated clock frequency error, and a current time. 3. The method of claim 2,
Determining a plurality of first sample mean values by averaging each of a plurality of first delay time sample groups selected from the plurality of first delay time samples,
Dividing the plurality of first delay time samples into a plurality of first subblocks each comprising N _sub first delay time samples, where N _sub is a natural number,
A second sub-block consisting of M _sub first delay time samples from each of the plurality of first sub-blocks, wherein M _sub is a natural number smaller than N _sub ,
Selecting valid N _ran second sub-blocks among the plurality of second sub-blocks, where N _ran is a natural number, and
And determining an average value of M _sub first delay time samples included in each of the selected N _ran second sub-blocks as the corresponding first sample average value. 3. The method of claim 2,
Determining a plurality of third sample mean values by averaging each of a plurality of second delay time sample groups selected from the plurality of second delay time samples,
Dividing the plurality of second delay time samples into a plurality of first subblocks each having N _sub second delay time samples, where N _sub is a natural number,
A second sub-block consisting of M _sub second delay time samples from each of the plurality of first sub-blocks, wherein M _sub is a natural number smaller than N _sub ,
Selecting valid N _ran second sub-blocks among the plurality of second sub-blocks, where N _ran is a natural number, and
And determining an average value of M _sub second delay time samples included in each of the selected N _ran second sub-blocks as the corresponding third sample average value. 6. The method of claim 5,
_Wherein the step of selecting valid N _ran second sub-blocks among the plurality of second sub-blocks comprises the steps of: determining a standard deviation of M _sub first delay time samples included in each of the plurality of second sub- Block, determining the second sub-block as a valid second sub-block if the second sub-block is smaller than the second sub-block. The method according to claim 6,
_Wherein the step of selecting valid N _ran second sub-blocks among the plurality of second sub-blocks comprises: _comparing a standard deviation of M _sub second delay time samples included in each of the plurality of second sub- Block, determining the second sub-block as a valid second sub-block if the second sub-block is smaller than the second sub-block. 3. The method of claim 2,
Wherein constructing a plurality of first samples and a plurality of second samples from the plurality of first sample mean values and the plurality of second sample mean values comprises:
Setting abscissa coordinate values of the first group of points to values of the plurality of second samples and setting ordinate coordinate values of the first group of points to values of the plurality of first samples,
Determining whether the at least one outlier among the first group of points is present using a random sample consensus (RANSAC) algorithm. 3. The method of claim 2,
Wherein configuring the plurality of third samples and the plurality of fourth samples from the plurality of third sample mean values and the plurality of fourth sample mean values comprises:
Setting the abscissa coordinate values of the second group of points to the values of the plurality of fourth samples and the ordinate coordinate values of the second group of points to the values of the plurality of third samples,
And determining whether there is at least one outlier among the second group of points using the RANSAC algorithm. The method of claim 3,
Defining the third group of points in the two-dimensional coordinate system using the plurality of first samples and the plurality of second samples, and defining the points of the third group as a first one-dimensional function, And applying a curve fitting algorithm to the points of the clock frequency error. The method of claim 3,
Defining the fourth group of points in the two-dimensional coordinate system using the plurality of third samples and the plurality of fourth samples, and defining the points of the fourth group as a second one-dimensional function, And applying a curve fitting algorithm to the points of the clock frequency error. 5. The method of claim 4,
The step of estimating the clock frequency error using the first slope value, the second slope value, the first standard deviation, and the second standard deviation

Wherein E _Freq represents the clock frequency error, f _t2t1 represents the first slope value, f _t4t3 represents the second slope value,

Represents the first standard deviation,

And wherein the step of estimating the clock frequency error comprises: estimating the clock frequency error by: 14. The method of claim 13,
The step of estimating the clock time error using the first mean value to the fourth mean value, the estimated clock frequency error and the current time value

Wherein E _Time represents the clock time error, E _Freq represents the clock frequency error, D _t2t1 represents the first average value, D _t4t3 represents the third average value, T ₁ represents the second average value, T ₄ represents the fourth average value, T _current represents the current time, And estimating the clock time error by the clock frequency error. An apparatus for estimating a clock frequency error in a PTP slave,
A delay configured to construct a plurality of first delay time samples and a plurality of transmission time samples each associated with a plurality of sink messages and to configure a plurality of second delay time samples and a plurality of reception time samples, Wherein the plurality of transmission time samples are each associated with the plurality of first delay time samples and the plurality of reception time samples are each with the plurality of second delay time samples,
A selected portion of the plurality of first delay time samples, a portion of the plurality of transmission time samples associated with a selected portion of the plurality of first delay time samples, a selected portion of the plurality of second delay time samples, And a clock frequency error estimator configured to estimate the clock frequency error based on some received time samples of the plurality of received time samples associated with a selected portion of the plurality of second delay time samples, Device. An apparatus for estimating a clock frequency error in a PTP slave,
Determining a plurality of transmission time samples by identifying a plurality of transmission times in which a plurality of sync messages are transmitted from the PTP master and calculating delay times of the plurality of sync messages from the PTP master to the PTP slave, Determining a plurality of delay time samples by calculating delay times of a plurality of delay request messages from the PTP master to the PTP master and determining a plurality of second delay time samples, A delay detector configured to identify a plurality of reception times and to determine a plurality of reception time samples, the plurality of transmission time samples being each associated with the plurality of first delay time samples, Respectively, and the delay detector is associated with a second delay time sample, N number of first _sub delay time samples, N _sub transmit time samples, N of second _sub delay samples, and is configured to output the N _sub reception time samples, wherein N is a natural number _sub-, and
Each period of the selection by receiving the N _sub of first delay sample and the N _sub transmit time samples from the delay detection part configured to output the N _ran of first sample average value and the N _ran of second sample average first Sub-block filter, the N _ran is a natural number,
Each period of the selection by receiving the N _sub of second delay samples and the N _sub reception time samples from the delay detection unit 2 configured to output the N _ran of third sample average value and the N _ran of the fourth sample mean value Sub-block filter,
The first sub-block N _ran of first samples the average value from the filter and N _ran of second samples by receiving the mean value M _ran of the first sample and the first RANSAC filter, the M _ran configured to provide an M _ran of the second sample Dimensional coordinate system using the first sample of the first one-dimensional function and the second sample of the M _ran , defining the points of the first group as a first one-dimensional function, A first minimum square curve fitting calculator configured to determine a slope as a first slope value and a first delay and a time average unit configured to determine a standard deviation for the first sample of M _ran as a first standard deviation, Sub-block selection and block averaging section, wherein M _ran is a natural number equal to or less than N _ran ,
The second from the sub-block filter N _ran of third sample average value and the N _ran of the fourth sample receives the average value M _ran of third sample and the second RANSAC filter, the M _ran configured to provide an M _ran of the fourth sample Dimensional coordinate system using the third sample and the fourth sample of M _ran , defines the points of the second group as a second one-dimensional function, and defines the points of the second one-dimensional function A second least squares curve fitting calculator configured to determine a slope as a second slope value and a second delay and time average unit configured to determine a standard deviation for the third sample of M _ran as a second standard deviation, Sub-block selection and block averaging, and
And a frequency error and a time error estimator configured to estimate the clock frequency error using the first slope value, the second slope value, the first standard deviation, and the second standard deviation. 17. The method of claim 16,
The first sub-block filter
And an M selection unit configured to select M _sub first delay time samples of the N _sub first delay time samples, wherein M _sub is a natural number smaller than N _sub ,
A rollover delay storage unit,
Count value (R _cnt) a count register for storing initial value of the count value (R _cnt) being stored in the count register is 0 -, and
The M _sub of the first delay to validate the time samples the M _sub of claim 1 when it is determined that the delay time the sample is valid the M _sub of claim taking the average for the first delay sample determining the first sample mean value and andago not the effective that the determined M _sub of first delay taking the average of the M _sub transmit time samples associated with the time samples and determining the second sample mean value, the M _sub of first delay sample validity determination when containing the standard deviation checking unit configured to store the M number of first _sub delay samples in the delay rollover de storage unit and increases the count value (R _cnt) of the count register by one, and
The sorting and selecting unit, in the next cycle N _sub of first delay sample in response to an input by checking the count value (R _cnt) of the count register, the count value of the count register (R _cnt) is not 0 The first delay time sample of N _sub that is input in the next cycle and the first delay time sample of M _{sub that} is the smallest among the first delay time samples of M _sub stored in the rollover delay storage unit are selected Wherein the clock frequency error estimator further comprises: 18. The method of claim 17,
The standard deviation inspection unit, the M _sub of the first delay time is less than the threshold standard deviation is selected for the sample estimate the M _sub of first delay sample clock frequency error, is further configured to determine that the effective device . 17. The method of claim 16,
The first RANSAC filter, the N _ran of first sample average value and the N _ran of the at least one from using the two-sample mean value, and defining the third group points in a two-dimensional coordinate system, the points of the third group of If the outlier removing the first sample average value and the second sample average value corresponding to the at least one outlier the N _ran of claim, respectively from the first sample average value and the N _ran of second sample average the M _ran the 1 samples and the M _ran second samples. 17. The method of claim 16,
_Wherein the first delay and time averaging unit is further configured to determine an average value for the first sample of M _ran as a first mean value and determine an average value for the second sample of M _ran as a second mean value,
_Wherein the second delay and time averaging unit are further configured to determine an average value for the third sample of M _ran as a third mean value and determine an average value for the fourth sample of M _ran as a fourth mean value,
Wherein the frequency error and time error estimator are further configured to estimate a clock time error using a value indicating the first mean value to the fourth mean value, the estimated clock frequency error, and a current time. 21. The method of claim 20,
The frequency error and time error estimator may include:

Wherein E _Freq represents the clock frequency error, f _t2t1 represents the first slope value, f _t4t3 represents the second slope value,

Represents the first standard deviation,

Wherein the second frequency offset estimator is further configured to estimate the clock frequency error by: 22. The method of claim 21,
The frequency error and time error estimator may include:

Wherein E _Time represents the clock time error, E _Freq represents the clock frequency error, D _t2t1 represents the first average value, D _t4t3 represents the third average value, T ₁ represents the second average value, T ₄ represents the fourth average value, T _current represents the current time, And to estimate said clock time error by said clock frequency offset estimator.

Images