Underwater node time synchronization method for distributed ocean current acoustic measurement

Abstract

This invention discloses an underwater node time synchronization method for distributed ocean current field acoustic measurement. The method first initializes parameters, broadcasts time synchronization messages from beacon nodes, and performs clock skew rate estimation and clock offset estimation on ordinary nodes. Ordinary nodes then correct their local time based on these estimates. This invention addresses the assumption in common underwater node time synchronization algorithms that the round-trip propagation delay of bidirectional time synchronization messages is the same. In distributed ocean current field acoustic measurement scenarios, time synchronization estimation performance can be reduced due to errors caused by ocean current velocity. Therefore, the method analyzes and compensates for the current velocity error term. Furthermore, by combining underwater node time synchronization with the ocean current field measurement process, the method calculates and corrects the ocean current velocity between nodes after obtaining the node time synchronization estimation results, achieving joint estimation of ocean current velocity and node time synchronization parameters, thus effectively improving time synchronization accuracy.

Description

Technical Field

[0001] This invention belongs to the field of underwater node time synchronization technology, and relates to an underwater node time synchronization method for distributed ocean current field acoustic measurement. Specifically, it organically combines ocean current field measurement with node time synchronization process to jointly estimate ocean current velocity and time synchronization parameters. Background Technology

[0002] Acoustic measurement methods are one of the main methods for large-scale marine environmental observation. Sound waves are the best information carrier for underwater medium- and long-distance communication and are currently the main means of parameter measurement and data transmission in the underwater environment. Similarly, distributed underwater acoustic sensing systems are a novel means of measuring ocean current fields.

[0003] Acoustic-based ocean current field measurement methods invert current velocity based on the time difference of sound wave propagation between nodes, placing high demands on the synchronization of underwater node clocks and inter-node time. Due to hardware resource limitations, it is difficult for the local time of ordinary nodes to be synchronized with the reference clock, and long-term operation can lead to the accumulation of local time errors. If time synchronization correction is not performed, it will affect the time accuracy of data acquisition and measurement, resulting in errors in current velocity calculation. Therefore, time synchronization plays a crucial role in current field measurement. Due to factors such as the large propagation delay of underwater sound waves and the limited energy of underwater equipment, existing mature time synchronization methods, such as Network Time Protocol (NTP) and Flooding Time Synchronization Protocol (FTSP), are not suitable for underwater acoustic sensing systems.

[0004] To address the large propagation delay problem in underwater acoustic sensor networks, most underwater time synchronization algorithms employ bidirectional message exchange, such as the Time Synchronization for High Latency Networks (TSHL) algorithm, to compensate for clock frequency and phase offsets. These methods require underwater nodes to communicate with each other to compare and calibrate their clocks, achieving node clock synchronization. Simultaneously, through multiple information exchanges, propagation delays are calculated and compensated for, resulting in a better estimation of clock offsets. Compared to unidirectional underwater node time synchronization protocols, bidirectional interactive underwater node time synchronization protocols can estimate delays more accurately and also offer higher reliability and stability.

[0005] Patent No. 201910491113.9 discloses a joint estimation method for underwater node time synchronization and positioning based on pseudorange. Reference nodes periodically broadcast data packets with timestamps and location information, allowing the node to be located to synchronize its time and position itself based on the data packet information, without needing to send any data packets. Its limitation is that at least three reference nodes are required for underwater node auxiliary positioning and synchronization, and these nodes must be located within the bounded area of ​​the reference nodes to ensure positioning accuracy. Due to the distance limitations of underwater acoustic communication, there are certain requirements on the number of reference nodes. Patent Application No. 202010906496.4 discloses an underwater time synchronization method based on bidirectional time comparison. The time of an atomic clock is encoded and modulated onto a laser signal for bidirectional transmission between two time synchronization nodes. The time delay between the two nodes is calculated using the bidirectionally decoded time signal, and then their respective time delays are adjusted. This method uses lasers for underwater communication, has a short effective communication distance, requires a high-performance working environment, and the atomic clocks carried in the equipment are expensive, making it unsuitable for large-scale underwater communication network applications. Patent No. 202210938343.7 discloses a Doppler-assisted time synchronization method. This method utilizes a clustering model to achieve inter-node time synchronization through phased grouping. It considers the impact of clock frequency skew when estimating the Doppler scale factor, obtaining the clock frequency offset through multiple one-way interactions and the clock phase offset through one two-way interaction, thus completing the time synchronization process. This method requires multiple interaction processes to calculate the average and obtain estimates of the Doppler factor and clock frequency offset. Furthermore, when a time synchronization error occurs at the cluster head node, the error is propagated to all nodes within that cluster. In addition, the methods disclosed in the aforementioned patents all share a common problem: they assume equal round-trip propagation times during two-way message interaction, which does not reflect reality.

[0006] Under the condition of large propagation delay of underwater acoustic waves, a node time synchronization algorithm is designed. The basic working mechanism is to estimate the local clock deviation by exchanging time synchronization messages with a reference node. However, considering the limited energy of underwater nodes, it is necessary to minimize their participation in message sending and receiving to extend their working time; and given the limited storage and computing resources of underwater nodes, the time synchronization algorithm should not be too complex. Furthermore, existing underwater node time synchronization algorithms based on bidirectional message exchange assume the reciprocity of the underwater acoustic channel, ignoring the difference in propagation delay between forward and reverse directions caused by current velocity. Therefore, combining ocean current velocity with node clock parameters in a distributed ocean current acoustic measurement scenario is a problem with significant application value. Summary of the Invention

[0007] The purpose of this invention is to address the problems of limited energy of underwater nodes, large underwater acoustic signal propagation delay, and the influence of current velocity on signal propagation time in distributed ocean current acoustic measurement scenarios. By combining ocean current acoustic measurement with underwater node time synchronization, this invention provides an underwater node time synchronization method for distributed ocean current acoustic measurement, enabling accurate estimation of ocean current velocity and node clock offset between nodes.

[0008] The scenario in which this invention is applied: In a distributed underwater acoustic sensing system, the node with a standard reference clock (such as a GPS module or a high-precision atomic clock included in the system) is the beacon node B, and the nodes that need to be synchronized in time constitute the ordinary node set. N k This represents the kth ordinary node, where k = 1, 2, ..., K, and K is the number of ordinary nodes; the principle and process of time synchronization between each ordinary node and the beacon node are the same.

[0009] In a distributed underwater acoustic sensing system, underwater acoustic sensing nodes need to use the same clock unit, and their respective times are determined by the crystal oscillator inside each ordinary node. Due to differences in manufacturing processes, threshold design, and other factors, the actual frequency generated by the crystal oscillator is not consistent with the designed frequency; moreover, the crystal oscillator does not output a signal of the same frequency constantly during operation, but will change due to external conditions such as temperature and humidity or internal factors such as its own aging.

[0010] The deviation caused by crystal oscillator frequency error is described by the clock skew rate, where a is the clock skew rate of the k-th ordinary node at time t. k (t)=f k (t) / f0, f k (t) represents the actual operating frequency, and f0 represents the nominal frequency; the difference between the local time of each ordinary node and the standard time at reference time t=0 is called the clock offset, and the clock offset b of the kth ordinary node is... k =τ k (0), τ k This represents the local time of the k-th ordinary node.

[0011] If the underwater node operates in an environment with relatively stable temperature and humidity, it can be assumed that its clock frequency offset will not change over a period of time, i.e., a k (t)=a k The node's internal hardware time is linearly related to the standard reference time. A linear clock model is used to describe the clock offset of the underwater node, τ. k (t)=a k t+b k .

[0012] Each ordinary node N k The specific time synchronization methods are as follows:

[0013] Step (1) Parameter initialization:

[0014] Initialize the ocean current velocity v between the beacon node and ordinary nodes, set the spatial distance d between nodes, the underwater equivalent speed of sound c, the beacon node's broadcast time synchronization message sending interval T and the number of messages N, and the beacon node's time synchronization message processing time Δt. b,p Time synchronization message processing time Δt between ordinary nodes n,p .

[0015] Step (2) Beacon node time synchronization message broadcast:

[0016] The beacon node sends N time synchronization messages to ordinary nodes at intervals T. Each time synchronization message carries the beacon node's local timestamp at the time of message sending. i represents the number of time synchronization messages sent, 1≤i≤N, and the subscript b represents the beacon node.

[0017] Step (3) Clock skew rate estimation for ordinary nodes:

[0018] (3-1) After receiving the time synchronization message broadcast by the beacon node, the ordinary node records the local timestamp of each message when it arrives. The index n represents a regular node, and the local timestamp of the beacon node when sending the time synchronization message is parsed from the time synchronization message.

[0019] (3-2) Based on the clock model τ k (t)=a k t+b k Receive timestamp With sending timestamp The relationship between them is Δt d- This refers to the message propagation delay between beacon nodes and ordinary nodes. This refers to the time jitter error in receiving and synchronizing messages for ordinary nodes.

[0020] (3-3) Select the received number The time synchronization message is used as the main message. This indicates rounding down; the other (N-1) time synchronization messages are compared with the main message to obtain an estimate of the clock skew rate 'a' for the i-th time. because To avoid random errors in a single estimation process, the results of (N-1) estimations are averaged to obtain the final clock skew rate estimate.

[0021] Step (4) Clock skew estimation for ordinary nodes:

[0022] (4-1) Ordinary nodes send synchronization request messages to beacon nodes, recording the local timestamp of the ordinary node when the message is sent.

[0023] (4-2) After propagation delay Δt d+ Afterwards, the beacon node receives a synchronization request message from a regular node and records the beacon node's local timestamp at the time of message reception. It also replies with a synchronization response message, which includes the beacon node's local timestamp when it received the synchronization request message. Local timestamp of the beacon node when the response message was sent

[0024] (4-3) After propagation delay Δt d- Afterwards, the ordinary node receives the synchronization response message from the beacon node and records the local timestamp of the ordinary node at the time of message reception. And parse the two timestamps recorded at the beacon node from the synchronization response message. and

[0025] (4-4) The correspondence between the four timestamps above is as follows: This refers to the time jitter error in the beacon node's reception of synchronization request messages. The time jitter error for receiving synchronization response messages by ordinary nodes is Δt. d+ The estimated value of clock offset b is obtained for the message propagation delay between ordinary nodes and beacon nodes. Errors caused by ocean currents, and Then the estimated value of clock offset because Clock offset estimate

[0026] (4-5) The standard time corresponding to the timestamp of the message exchange process is... Calculate the ocean current velocity between nodes based on the corrected timestamp data.

[0027] (4-6) Recalculate the clock offset estimate for the normal node to obtain the final clock offset estimate.

[0028] Step (5) Output the estimated clock skew rate for node time synchronization. and clock offset estimates

[0029] Step (6) Ordinary node N k Based on the estimated value and Correct the local time.

[0030] The method of this invention takes into account that common underwater node time synchronization algorithms assume that the round-trip propagation delay of bidirectional time synchronization messages is the same. In the distributed ocean current acoustic measurement scenario, the time synchronization estimation performance will be reduced due to errors caused by ocean current velocity. Therefore, the flow velocity error term is analyzed and compensated. Furthermore, by combining the underwater node time synchronization and ocean current measurement process, the ocean current velocity between nodes is calculated and corrected after obtaining the node time synchronization estimation result, so as to realize the joint estimation of ocean current velocity and node time synchronization parameters, and the time synchronization accuracy is effectively improved. Attached Figure Description

[0031] Figure 1 This is a flowchart of the method of the present invention;

[0032] Figure 2 This is a schematic diagram of a clock model;

[0033] Figure 3 This is a timing diagram of the time synchronization message transmission process in the clock frequency offset estimation process of the joint estimation method;

[0034] Figure 4 This is a timing diagram for sending time synchronization messages during the clock phase offset estimation process in the joint estimation method. Detailed Implementation

[0035] The present invention will be further described in detail below with reference to the accompanying drawings and embodiments.

[0036] A method for underwater node time synchronization for distributed ocean current field acoustic measurement, the process of which is as follows: Figure 1 As shown, the clock model used is as follows: Figure 2 As shown. Specifically, it is achieved through the following steps:

[0037] Step (1) Parameter initialization:

[0038] Initialize the ocean current velocity between the beacon node and ordinary nodes to v = 0.6 m / s, set the spatial distance between nodes to d = 1 km, the underwater equivalent speed of sound to c = 1496 m / s, set the beacon node's broadcast time synchronization message sending interval T = 1 s and the number of messages sent to N = 10, and set the beacon node's time synchronization message processing time Δt. b,p =80ms, time synchronization message processing time Δt for ordinary nodes n,p =80ms.

[0039] Step (2) Beacon node time synchronization message broadcast:

[0040] The beacon node sends 10 time synchronization messages to ordinary nodes at 1-second intervals. Each time synchronization message carries the beacon node's local timestamp at the time of message sending. i represents the number of time synchronization messages sent, 1 ≤ i ≤ 10, and the subscript b represents the beacon node, such as... Figure 3 As shown.

[0041] Step (3) Clock skew rate estimation for ordinary nodes:

[0042] (3-1) After receiving the time synchronization message broadcast by the beacon node, the ordinary node records the local timestamp of each message when it arrives. The index n represents a regular node, and the local timestamp of the beacon node when sending the time synchronization message is parsed from the time synchronization message.

[0043] (3-2) Based on the clock model τ k (t)=a k t+b k Receive timestamp With sending timestamp The relationship between them is Δt d- This refers to the message propagation delay between beacon nodes and ordinary nodes. This refers to the time jitter error in receiving and synchronizing messages for ordinary nodes.

[0044] (3-3) Select the fifth received time synchronization message as the main message, compare the other nine time synchronization messages with the main message, and obtain the estimated value of the clock skew rate a for the i-th time. because To avoid random errors in a single estimation process, the results of nine estimations are averaged to obtain the final clock skew rate estimate.

[0045] Step (4) Clock skew estimation for ordinary nodes:

[0046] (4-1) Ordinary nodes send synchronization request messages to beacon nodes, recording the local timestamp of the ordinary node when the message is sent.

[0047] (4-2) After propagation delay Δt d+ Afterwards, the beacon node receives a synchronization request message from a regular node and records the beacon node's local timestamp at the time of message reception. It also replies with a synchronization response message, which includes the beacon node's local timestamp when it received the synchronization request message. Local timestamp of the beacon node when the response message was sent

[0048] (4-3) After propagation delay Δt d- Afterwards, the ordinary node receives the synchronization response message from the beacon node and records the local timestamp of the ordinary node at the time of message reception. And parse the two timestamps recorded at the beacon node from the synchronization response message. and

[0049] (4-4) such as Figure 4 As shown, the correspondence between the four timestamps above is as follows: This refers to the time jitter error in the beacon node's reception of synchronization request messages. The time jitter error for receiving synchronization response messages by ordinary nodes is Δt. d+ The estimated value of clock offset b is obtained for the message propagation delay between ordinary nodes and beacon nodes. Errors caused by ocean currents, and Then the estimated value of clock offset b because Clock offset estimate

[0050] (4-5) The standard time corresponding to the timestamp of the message exchange process is... Calculate the ocean current velocity between nodes based on the corrected timestamp data.

[0051] (4-6) Recalculate the clock offset estimate for the normal node to obtain the final clock offset estimate.

[0052] Step (5) Output the estimated clock skew rate for node time synchronization. and clock offset estimates

[0053] Step (6) Ordinary node N k Based on the estimated value and Correct the local time.

[0054] The above description of the embodiments is merely an enumeration of implementations of the present invention. The scope of protection of the present invention should not be limited to the specific forms described in the embodiments. Any improvements and modifications made by those skilled in the art based on the disclosure of the present invention should be within the scope of protection of the present invention.

Claims

1. A time synchronization method for underwater nodes in distributed ocean current field acoustic measurement, wherein nodes with a standard reference clock in the distributed underwater acoustic sensing system are beacon nodes. The nodes that need to be synchronized in time constitute the set of ordinary nodes. , Indicates the first A regular node, , This represents the number of ordinary nodes; In a distributed underwater acoustic sensing system, underwater acoustic sensing nodes use the same clock unit, and their respective times are determined by a crystal oscillator within each ordinary node; the deviation caused by the crystal oscillator frequency error is described by the clock skew rate. A normal node Clock skew rate at time , Indicates the actual operating frequency. Indicates the nominal frequency; clock offset is the reference time. The difference between the local time and the standard time of each ordinary node, the first... Clock skew of a normal node , Indicates the first Local time of each ordinary node; setting the clock model ; Its characteristic is that each ordinary node The specific time synchronization methods are as follows: Step (1) Parameter initialization: Initialize the ocean current velocity between beacon nodes and ordinary nodes Set the spatial distance between nodes underwater equivalent speed of sound The interval for sending time synchronization messages broadcast by beacon nodes and number of messages sent Time synchronization message processing time of beacon nodes Time synchronization message processing time with ordinary nodes ; Step (2) Beacon node time synchronization message broadcast: beacon nodes transmit at intervals Send to ordinary nodes Each time synchronization message carries a local timestamp from the beacon node at the time the message was sent. , To determine the number of time synchronization messages sent, subscript Indicates a beacon node; Step (3) Clock skew rate estimation for ordinary nodes: (3-1) After receiving the time synchronization message broadcast by the beacon node, the ordinary node records the local timestamp of each message when it arrives. subscript This represents a regular node, and it extracts the beacon node's local timestamp when the beacon node sent the time synchronization message from the time synchronization message. ; (3-2) Based on the clock model Receive timestamp With sending timestamp The relationship between them is , This refers to the message propagation delay between beacon nodes and ordinary nodes. This refers to the time jitter error in receiving and synchronizing messages for ordinary nodes. (3-3) Select the received number The time synchronization message is used as the main message. Indicates rounding down; rounds other values ​​down. The first time synchronization message is compared with the main message to obtain the first time synchronization message. Secondary clock offset The estimated value ;right The final clock skew rate estimate is obtained by averaging the results of the multiple estimates. ; Step (4) Clock skew estimation for ordinary nodes: (4-1) Ordinary nodes send synchronization request messages to beacon nodes, recording the local timestamp of the ordinary node when the message is sent. ; (4-2) Propagation delay Afterwards, the beacon node receives a synchronization request message from a regular node and records the beacon node's local timestamp at the time of message reception. It then replies with a synchronization response message, which also carries the beacon node's local timestamp when it received the synchronization request message. Local timestamp of the beacon node when the response message was sent ; (4-3) Propagation delay Afterwards, the ordinary node receives the synchronization response message from the beacon node and records the local timestamp of the ordinary node at the time of message reception. And the two timestamps recorded at the beacon node are parsed from the synchronization response message. and ; (4-4) The correspondence between the four timestamps is as follows: , This refers to the time jitter error in the beacon node's reception of synchronization request messages. This refers to the time jitter error in receiving synchronization response messages by ordinary nodes. This refers to the message propagation delay between ordinary nodes and beacon nodes; The estimated value of the clock offset is obtained: Errors caused by ocean currents ; (4-5) The standard time corresponding to the timestamp of the message exchange process is... Ocean current velocities between nodes are calculated based on the corrected timestamp data. ; (4-6) Recalculate the clock skew estimate for the normal nodes to obtain the final output clock skew estimate. ; Step (5) Output the estimated clock skew rate for node time synchronization. and clock offset estimates ; Step (6) Ordinary Nodes Based on the estimated value and Correct the local time.

Images