WO2018098791A1

WO2018098791A1 - Clock synchronization frequency deviation estimation method applicable to multi-hop wireless sensor network

Info

Publication number: WO2018098791A1
Application number: PCT/CN2016/108291
Authority: WO
Inventors: 王恒; 王平; 邵伦; 王保国
Original assignee: 重庆邮电大学
Priority date: 2016-12-01
Filing date: 2016-12-01
Publication date: 2018-06-07

Abstract

The present invention relates to a clock synchronization frequency deviation estimation method applicable to a multi-hop wireless sensor network, and belongs to the technical field of wireless sensor networks. In the method, two methods comprising a forwarding method and a transparent transmission method, are provided based on demands of a node for synchronization precision: A, in the forwarding method, a node in an intermediate layer initiates synchronous interaction with a node in the next layer at the same time interval after every correction time; and B, in the transparent transmission method, the processing time of the node in the intermediate layer is regarded as an additionally introduced fixed time delay and random time delay, and the node directly establishes a synchronization relationship with a root node. According to the estimation method provided in the present invention, the node can correct the local clock in each synchronous information interaction process, the synchronization precision error of the node is always kept at a relatively low level, and the problem of synchronization precision uncontrollability in a synchronization parameter estimation process is accordingly overcome.

Description

Clock synchronization frequency offset estimation method suitable for multi-hop wireless sensor networks

Technical field

The invention belongs to the technical field of wireless sensor networks, and relates to a clock synchronization frequency offset estimation method suitable for a multi-hop wireless sensor network.

Background technique

Time synchronization technology is an indispensable part of wireless sensor networks. It can provide a standard time reference for network-wide devices. It is the basic support for network data fusion, energy management, transmission scheduling and other services. In practical wireless sensor network applications, the network is usually composed of a large number of battery-powered nodes deployed in different environments, so energy consumption is a critical issue to consider. In addition, communication is the main source of energy consumption in wireless sensor networks, and an effective clock synchronization parameter estimation algorithm can reduce the number of resynchronizations in the network, prolong the network synchronization period, thereby reducing the communication overhead of the network and saving the energy of the nodes.

In the prior art, some clock synchronization parameter estimation algorithms have been proposed. By analyzing and modeling the process of clock synchronization, statistical signal processing techniques are used to estimate clock offset and frequency offset. However, the synchronization parameter estimation of the nodes in these algorithms is performed after multiple synchronization interaction periods, and the node cannot adjust its own local clock during the operation of the algorithm, so the clock synchronization accuracy cannot be guaranteed within the running time of the algorithm. This is unacceptable in some applications that require high synchronization accuracy, such as ISA100.11a. In addition, when the synchronization between two points is extended to the multi-hop wireless sensor network, it can be mainly divided into two types: hierarchical and proxy.

The invention mainly considers the whole network synchronization of a multi-hop wireless sensor network adopting a layered synchronization structure. In addition, the synchronization between two points cannot be directly extended to the multi-hop wireless sensor network, because the delay error of the intermediate node is introduced into the whole network, which affects the clock synchronization performance of the entire network.

Summary of the invention

In view of this, an object of the present invention is to provide a clock synchronization frequency offset estimation method suitable for a multi-hop wireless sensor network, in which the node can correct the local clock and synchronize the synchronization of the node during each synchronization message interaction. The error is always maintained at a low level.

In order to achieve the above object, the present invention provides the following technical solutions:

A clock synchronization frequency offset estimation method suitable for multi-hop wireless sensor networks. In this method, according to the requirements of the node for synchronization accuracy, it is divided into two methods: a forwarding method and a transparent transmission method. When the forwarding method is adopted and the node requires synchronization accuracy, the transparent transmission method is adopted.

Forwarding method: The time message in the multi-hop hierarchical synchronization network is propagated down by the root node step by step. The intermediate layer node synchronizes with the node to be synchronized as the node to be synchronized, and uses two timestamps after each interaction. Difference

As the adjustment time of the node time, the local clock is corrected. After the interval is fixed, the source node participates in the synchronization of the next-level node until it is spread to the node to be synchronized, repeats the above process, and uses the statistical signal after multiple synchronization cycles. The method of processing estimates the relative frequency offset of the node;

Transparent transmission method: in the multi-hop hierarchical synchronization network, the processing time of the intermediate layer node is regarded as the additional introduced fixed delay

And random delay

The message passing process of the intermediate node is regarded as transparent until the node to be synchronized obtains the resynchronization time of the root node and utilizes the difference between the two timestamps.

The calibration is completed, and the above process is repeated. After a plurality of cycles, the relative frequency offset is estimated.

Further, in the forwarding method adopted by the present invention, when the L-th node A _L in the network needs to implement time synchronization with the source node, the source node A serves as a clock source reference node, and the L-th node A _L is a node to be synchronized, and the synchronization step thereof include:

1) The time message in the network is propagated step by step by the source node, and the child node receives the data frame sent from its parent node, and obtains the sending timestamp and the local receiving timestamp, which are respectively

versus

2) Node A _L uses the difference between two timestamps

Get the adjustment

And at a later time

Adjust your local time, ie subtract the adjustment from the current local time

New time as node A _L ;

3) Starting from the Nth (N>1) cycle, each child node in the multi-hop network updates the local clock in the new cycle, so a compensation amount μ _i needs to be added, assuming that the node starts from receiving the data frame of the parent node. When the data frame is sent to the next-level child node, X is a fixed value, so that it is t', then there is

4) Repeat steps 1) and 2), use the array to store the timestamp, and after N cycles, estimate the relative frequency offset. The formula is as follows:

Estimated node A ₂ A ₁ with respect to the frequency shift of the node

Where N is the number of times of time synchronization.

Further, the transparent transmission method adopted by the present invention, the synchronization step specifically includes:

1) Treat the processing time of the intermediate node level as an additional introduced fixed delay in a multi-hop hierarchical synchronization network

And random delay

2) Node A _L utilizes the difference between the two timestamps

Get the adjustment

And at a later time

Adjust your local time, ie subtract the adjustment from the current local time

New time as node A _L ;

3) Repeat steps 1) and 2), use the array to store the timestamp, and after N synchronization interaction cycles, estimate the clock frequency offset. The formula is as follows:

among them,

The beneficial effects of the invention are:

1. The clock synchronization frequency offset estimation method applicable to a multi-hop wireless sensor network provided by the present invention, after a plurality of synchronization messages are exchanged by a node, the relative frequency offset is estimated at one time, and the synchronization precision is maintained when the crystal oscillator is unchanged. In a certain range, there is no need to perform synchronous message interaction for a period of time, thereby saving energy.

2. Compared with the existing multi-hop wireless sensor network time synchronization scheme, the present invention provides a clock synchronization frequency offset estimation method suitable for a multi-hop wireless sensor network, in which each node can correct its local state during each clock synchronization period. The clock keeps its synchronization accuracy error at a low level, thereby overcoming the problem that the synchronization accuracy is uncontrollable during the operation of the synchronization algorithm.

3. According to the requirement of clock synchronization precision of the multi-hop wireless sensor network, the present invention provides two methods for estimating the clock synchronization frequency offset of the forwarding method and the transparent transmission method. The former is suitable for multi-hop wireless sensor networks with high synchronization accuracy requirements, and the latter is suitable for multi-hop wireless sensor networks with general accuracy requirements.

DRAWINGS

In order to make the objects, technical solutions and advantageous effects of the present invention more clear, the present invention provides the following drawings for explanation:

1 is a schematic diagram of synchronization of a multi-hop node provided by the present invention;

2 is a comparison diagram of a corrected and non-corrected type in a multi-hop network provided by the present invention;

3 is a flowchart of a method for estimating a frequency offset of a multi-hop network according to the present invention;

4 is a performance comparison diagram of a frequency offset estimation result of a forwarding method according to the present invention;

FIG. 5 is a performance comparison diagram of the frequency shift estimation result of the transparent transmission method provided by the present invention.

detailed description

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Hereinafter, preferred embodiments of the present invention will be described in detail with reference to the accompanying drawings.

FIG. 1 is a schematic diagram of synchronization of a multi-hop node provided by the present invention. As shown in the figure, in a multi-hop hierarchical synchronous network, a root node A serves as a clock source reference node, and an L-th node A _L is a node to be synchronized.

They are the send timestamp and the receive timestamp respectively.

As the adjustment time of the node time, the local clock is corrected. After the interval is fixed, the source node participates in the synchronization of the next-level node until it is spread to the node to be synchronized, repeats the above process, and uses the statistical signal after multiple synchronization cycles. The processing method estimates the relative frequency offset of the node. The specific steps are as follows:

For the first synchronization cycle,

Implementation model can be expressed as

among them,

with

The initial clock offset and the relative frequency offset at time t ₀ are respectively indicated, and the fixed delay and the random delay in the transmission process are respectively

The next moment

Node A ₁ uses the difference

Adjust local time, and

Indicates the time after node A _{1 is} adjusted.

Actually, t ₀ to

Real clock offset of node A ₁ and node A in time

(4)-(2) Obtain a new clock offset after node A ₁ adjustment

Since the node in the multi-hop network adopts a step-by-step synchronization mechanism, then node A ₁ sends a time message to its child node A ₂ , similar to the above process, there is

among them

A ₁ represents a transmission time stamp of the node, the node A ₂ after the new clock offset adjustment

for

Starting from the second cycle, each child node in the multi-hop network updates the local clock in the new cycle, so a compensation amount μ _i needs to be added, that is, in the second cycle,

Have

For a two-hop node, the new clock offset after the second cycle adjustment can be expressed as

From N (N> 1) cycle starts, the clock source node single-hop scenario different, since each child node in a multi-hop network will correct their time in each cycle, it is necessary to add a compensation amount μ _i . Let the node start from the data frame receiving the parent node to the next-level child node to send a data frame. The time X is a fixed value, so that it is t'.

Repeat steps A0 and A1, and use the array to store the timestamp. After N cycles, the relative frequency offset can be estimated. The formula is as follows:

To derive the estimated node A ₂ A ₁ with respect to frequency offset node

Where N is the number of times of time synchronization.

In order to verify the validity of the results, the Cramer-Rao Lower Bound (CRLB) is found.

Transparent transmission method: in the multi-hop hierarchical synchronization network, the processing time of the intermediate node level is regarded as the additional introduced fixed delay

And random delay

The message passing process of the intermediate node is regarded as transparent until the node to be synchronized obtains the current time of the root node and utilizes the difference between the two timestamps.

The calibration is completed, and the above process is repeated. After a plurality of cycles, the relative frequency offset can be estimated.

For a two-hop node, node A ₂ is the node to be synchronized, and A is the root node. The specific synchronization steps are as follows:

Let t ₀ be the synchronization start time, and the transmission timestamp and the reception timestamp are respectively

For the first cycle,

Can be expressed as:

Node A _{2 is} then

Use the difference between the two timestamps to adjust the local time,

Indicates the time after node A _{2 is} adjusted, and

Can get

Two-type subtraction results in a new clock offset adjusted by node A ₂

For the second cycle, the time adjustment of node A ₂

will

with

Substituting the above formula

Repeat the above steps, the time adjustment of the ith cycle is

among them

Similarly, for the L-th layer node, referring to the above derivation, the synchronization expression of the i-th cycle is obtained as

Estimation formula based on frequency offset

Estimate the frequency offset of node A relative to node A _L

among them,

In order to verify the validity of the results, the Cramer-Rao Lower Bound (CRLB) is obtained.

2 is a comparison diagram of correction and non-correction in the multi-hop sensor network provided by the present invention, wherein FIG. 2(a) is a correction synchronization process, and FIG. 2(b) is a non-correction synchronization process.

Example:

FIG. 3 is a flowchart of a method for estimating a frequency offset of a multi-hop sensor network according to the present invention. This embodiment provides a clock offset estimation method suitable for a multi-hop wireless sensor network. As shown in the figure, the method includes the following steps:

C1: The synchronization process begins.

C2~C4: The node sends a local timestamp to the next-level child node to determine whether the node is the root node. If it is the root node, it enters the next stage C5. Otherwise, it enters C4 to use the adjustment amount to correct the local time.

C5~C6: The child node receives and records the timestamp, and uses the difference between the transmission timestamp and the reception timestamp to find the compensation amount ΔT.

C7~C8: Determine whether the node is the node to be synchronized. If yes, proceed to the next stage C9. Otherwise, proceed to process C3 to prepare the local timestamp for the next level node.

C9~C11: It is determined whether the number of synchronization cycles reaches the set value N. If it has been reached, the relative frequency offset of the node is estimated. Otherwise, the process proceeds to the process C3 to continue the message transmission process.

C12~C13: The local clock is compensated by the estimated relative frequency offset, and the synchronization process ends.

Figure 4 shows a comparison of the performance of the forwarding method frequency offset estimation with its CRLB. The figure shows the mean square error curve and the curve is very close CRLB proved node ₂ A frequency offset estimate is valid, and the estimated near optimum estimate.

FIG. 5 is a comparison diagram of the results of the transparent transmission frequency offset estimation and the corresponding CRLB performance provided by the present invention. It can be seen from the figure that the two curves are very close, and the mean square error curve is always above the CRLB curve, indicating that the estimation result is approximated as the least variance unbiased estimate.

It is to be understood that the above-described preferred embodiments are only illustrative of the technical solutions of the present invention, and are not intended to be limiting, although the present invention has been described in detail by the foregoing preferred embodiments, those skilled in the art Various changes are made in the details without departing from the scope of the invention as defined by the appended claims.

Claims

A clock synchronization frequency offset estimation method suitable for a multi-hop wireless sensor network, characterized in that: in the method, according to the requirement of the synchronization precision of the node, the forwarding method and the transparent transmission method are divided into two types, and the node pair synchronization precision When the requirement is high, the forwarding method is adopted, and when the node requires the synchronization accuracy in general, the transparent transmission method is adopted; the details are as follows:

Forwarding method: The time message in the multi-hop hierarchical synchronization network is propagated down by the root node step by step. The intermediate layer node synchronizes with the node to be synchronized as the node to be synchronized, and uses two timestamps after each interaction. Difference
Correcting your own local clock as an adjustment to the node time, where
Respectively, and transmit timestamp node A L A L-1 of the adjacent receiving time stamp, a fixed time interval, but also as a source node involvement synchronous at a node until the node to be synchronized to the diffusion, the above process is repeated, Estimating the relative frequency offset of the node by using a statistical signal processing method after multiple synchronization periods;

Transparent transmission method: in the multi-hop hierarchical synchronization network, the processing time of the intermediate layer node is regarded as the additional introduced fixed delay
And random delay
The message passing process of the intermediate node is regarded as transparent until the node to be synchronized obtains the resynchronization time of the root node and utilizes the difference between the two timestamps.
The calibration is completed, and the above process is repeated. After a plurality of cycles, the relative frequency offset is estimated.
The method for estimating a clock synchronization frequency offset suitable for a multi-hop wireless sensor network according to claim 1, wherein the forwarding method is used, when the L-th node A L in the network needs to implement time synchronization with the source node, the source Node A serves as the clock source reference node, and the Lth node A L is the node to be synchronized. The synchronization steps include:

1) The time message in the network is propagated step by step by the source node, and the child node receives the data frame sent from its parent node, and obtains the sending timestamp and the local receiving timestamp, which are respectively
versus

2) Node A L according to the formula
Get the adjustment
And at a later time
Adjust your local time, ie subtract the adjustment from the current local time
New time as node A L ;

3) Starting from the Nth (N>1) cycle, each child node in the multi-hop network updates the local clock in the new cycle, so a compensation amount μ i needs to be added, assuming that the node starts from receiving the data frame of the parent node. When the data frame is sent to the next-level child node, X is a fixed value, so that it is t', then there is

4) Repeat steps 1) and 2), use the array to store the timestamp, and after N cycles, estimate the relative frequency offset. The formula is as follows:

Estimating the frequency offset of node A 2 relative to node A 1
Where N is the number of periods of time synchronization.
The method for estimating a clock synchronization frequency offset for a multi-hop wireless sensor network according to claim 1, wherein the step of synchronizing, the level L node is to be synchronized, and the step of synchronizing comprises:

1) Treat the processing time of the intermediate node level as an additional introduced fixed delay in a multi-hop hierarchical synchronization network
And random delay

2) Node A L uses the difference between two timestamps
Get the adjustment
And at a later time
Adjust your local time, ie subtract the adjustment from the current local time
New time as node A L ;

3) Repeat steps 1) and 2), use the array to store the timestamp, and after N synchronous interaction cycles, estimate the clock frequency offset. The formula is as follows:

among them,