CN-121984629-A - Time synchronization method and device of positioning system

CN121984629ACN 121984629 ACN121984629 ACN 121984629ACN-121984629-A

Abstract

The invention discloses a time synchronization method and device of a positioning system, wherein the positioning system comprises a processor, a satellite module and at least one sensing module, the method comprises the steps of initializing an internal time source, continuously capturing PPS pulse signals, recording a time stamp each time when the rising edge of the PPS pulse signals is captured, establishing an initial conversion relation according to the time stamp, constructing a Kalman filtering model, dynamically updating the Kalman filtering model according to the time stamp, carrying out real-time updating on the initial conversion relation according to the dynamically updated Kalman filtering model to obtain an accurate conversion relation, evaluating the health degree of the internal time source, determining a synchronization mode according to the health degree, carrying out time synchronization on the current internal time according to the synchronization mode and the accurate conversion relation, and outputting the synchronization time to the sensing module. The invention can improve the accuracy and the robustness of time synchronization.

Inventors

WANG JIEKAI
ZHOU GUANGHAI
HUANG ZHIZHOU
CHEN HANLONG

Assignees

广州南方测绘科技股份有限公司

Dates

Publication Date: 20260505
Application Date: 20251229

Claims (10)

1. A method of time synchronization of a positioning system comprising a processor, a satellite module and at least one sensor module, the method comprising the steps of: Initializing an internal time source in the processor, continuously capturing a PPS pulse signal sent by the satellite module through the internal time source, recording a time stamp each time the rising edge of the PPS pulse signal is captured, and establishing an initial conversion relation according to the time stamp, wherein the time stamp comprises UTC time and internal time of the processor, and the initial conversion relation is used for describing the conversion relation between the UTC time and the internal time; Establishing a Kalman filtering model, dynamically updating the Kalman filtering model according to the timestamp recorded at the latest time, and updating the initial conversion relation in real time according to the Kalman filtering model obtained by the latest dynamic update to obtain an accurate conversion relation, wherein the updating period of the real-time update is independently set; And evaluating the health degree of the internal time source, determining a synchronous mode according to the health degree, performing time synchronization on the current internal time according to the synchronous mode and the accurate conversion relation, and outputting the synchronous time to the sensing module.
2. The method for time synchronization of a positioning system according to claim 1, wherein each time a rising edge of the PPS pulse signal is captured, the processor records a time stamp, and establishes an initial conversion relationship according to the time stamp, and specifically comprises: receiving GNSS signals sent by the satellite module through the internal time source, and acquiring the PPS pulse signals and the UTC time; Each time a rising edge of the PPS pulse signal is captured, the processor records the current UTC time and the internal time, and generates one time stamp; Establishing a linear conversion model according to a timestamp corresponding to the rising edge of the first PPS pulse signal; And performing parameter fitting on the linear conversion model through a least square method according to a plurality of time stamps obtained later to obtain the initial conversion relation.
3. The method for time synchronization of a positioning system according to claim 2, wherein the linear transformation model is specifically: T_utc = a * T_raw + b Wherein t_ UTC is the UTC time, t_raw is the internal time, a is the clock rate ratio, and b is the time offset.
4. The method for time synchronization of a positioning system of claim 1 wherein the Kalman filtering model comprises a state vector, a state equation, and an observation equation; the state vector is specifically represented by X= [ a, b, da/dt ] (a is a clock rate ratio, b is time offset, and da/dt is the rate of change of the rate ratio); the state equation is specifically as follows: a k = a k-1 + (da/dt) k-1 *Δt + w a b k = b k-1 + a k-1 *Δt + w b (da/dt) k = (da/dt) k-1 + w (da/dt) wherein k is the current time step, Δt is the preset prediction step length, and w a , w b , w (da/dt) is the process noise; The observation equation is specifically as follows: z k = T_utc n - (a k * T_raw n + b k ) = 0 + v k Where z k is the observation residual, v k is the observation noise, and t_ UTC n and t_raw n are the UTC time and the internal time obtained by the nth recording, respectively.
5. The method for time synchronization of a positioning system according to claim 4, wherein the dynamically updating the kalman filter model according to the timestamp recorded last time, and updating the initial conversion relationship in real time according to the kalman filter model obtained by each update, to obtain an accurate conversion relationship, specifically comprises: predicting the state vector at the current moment according to the state vector and the state equation at the previous moment to obtain a predicted vector; according to the timestamp recorded at the latest time, calculating to obtain the observation residual error and the Kalman gain at the current moment; updating the prediction vector according to the observation residual error and the Kalman gain to obtain an accurate state vector at the current moment; Substituting the clock rate ratio and the time offset in the accurate state vector into the initial conversion relation to obtain the accurate conversion relation.
6. The time synchronization method of a positioning system according to claim 1, wherein the evaluating the health of the internal time source comprises: the internal time source comprises a GNSS source, an NTP/PTP source and an OCXO source, wherein the GNSS source is a main time source, and the NTP/PTP source and the OCXO source are auxiliary time sources; obtaining the health degree of the GNSS source by acquiring the signal intensity, the number of visible satellites and the PDOP value in a module report output by the GNSS source; The health degree of the NTP/PTP source is evaluated and obtained according to network delay, jitter and server synchronization state; and evaluating and obtaining the health degree of the OCXO source according to the factory nominal parameters and the temperature characteristics of the OCXO source.
7. The method for time synchronization of a positioning system according to claim 6, wherein said determining a synchronization pattern according to said health degree comprises: when the health degree of the GNSS source is higher than a preset health threshold, determining that the synchronous mode is a normal mode; when the health degree of the GNSS source is lower than the preset health threshold value but the signals of the GNSS source are not received, determining that the synchronous mode is an early warning mode; when the signal of the GNSS source is completely lost, the synchronization mode is determined to be a hold mode.
8. The method for time synchronization of a positioning system according to claim 7, wherein said time synchronizing the current internal time according to the synchronization pattern and the accurate conversion relation, and outputting a synchronization time to the sensor module, comprises: When the synchronous mode is the normal mode, setting the updating period of the real-time updating to be a preset normal period, carrying out time synchronization on the internal time according to the accurate conversion relation obtained by the latest real-time updating to obtain the synchronous time, and finally outputting the synchronous time to the sensing module; When the synchronous mode is the early warning mode, stopping dynamic update of the Kalman filtering model, shortening the update period of the real-time update, performing time synchronization on the internal time according to the accurate conversion relation obtained by the latest real-time update to obtain the synchronous time, and finally outputting the synchronous time to the sensing module; and when the synchronous mode is the holding mode, stopping the real-time updating of the initial conversion relation, carrying out time synchronization on the internal time according to the accurate conversion relation obtained by the last real-time updating to obtain the synchronous time, and finally outputting the synchronous time to the sensing module.
9. The method of time synchronization of a positioning system of claim 8, further comprising, when the synchronization mode is the hold mode: and predicting the frequency offset of the OCXO source in the current environment according to the temperature-frequency characteristic curve of the OCXO source, and fine-tuning the accurate conversion relation obtained by the last real-time update according to the frequency offset.
10. The time synchronization method device of the positioning system comprises a processor, a satellite module and at least one sensing module, and is characterized by comprising a capturing module, a construction module and a synchronization module; The acquisition module is used for initializing an internal time source in the processor, continuously acquiring a PPS pulse signal sent by the satellite module through the internal time source, recording a time stamp by the processor each time the rising edge of the PPS pulse signal is acquired, and establishing an initial conversion relation according to the time stamp, wherein the time stamp comprises UTC time and the internal time of the processor, and the initial conversion relation is used for describing the conversion relation between the UTC time and the internal time; The construction module is used for constructing a Kalman filtering model, dynamically updating the Kalman filtering model according to the timestamp recorded at the latest time, and updating the initial conversion relation in real time according to the Kalman filtering model obtained by each update to obtain an accurate conversion relation, wherein the updating period of the real-time update is independently set; The synchronization module is used for evaluating the health degree of the internal time source, determining a synchronization mode according to the health degree, performing time synchronization on the current internal time according to the synchronization mode and the accurate conversion relation, and outputting the synchronization time to the sensing module.

Description

Time synchronization method and device of positioning system Technical Field The present invention relates to the field of time synchronization, and in particular, to a time synchronization method and apparatus for a positioning system. Background In applications such as automatic driving, unmanned aerial vehicle, precision mapping, and the like, multi-sensor data fusion is a key for realizing high-precision positioning and sensing. In order to guarantee the performance of the fusion algorithm, the data acquired by the different sensors must be aligned precisely in time, i.e. a high-precision time synchronization is achieved. In the prior art, a common method is to use a second pulse (PPS) signal output by a satellite module (such as GNSS) and universal coordinated time (UTC) information, establish a conversion relationship between an internal clock of a processor and universal time through a software algorithm, and perform time synchronization on other sensors (such as IMU and camera) according to the relationship. The method avoids complex hardware synchronous circuits and reduces the cost. However, the prior art still has the following significant drawbacks: (1) Clock drift causes degradation in accuracy existing schemes typically assume that the conversion relationship (e.g., time difference scale) of the processor's internal clock to the world clock is constant over a period of time. In practice, however, the frequency of the internal clock of the processor may drift due to temperature changes, voltage fluctuations, device aging, and the like. Such drift can cause changes in parameters of the conversion relationship (e.g., slope and offset), which if not updated in time, can introduce significant synchronization errors, especially if the system is operated for a long period of time or the environment is severely changed. (2) The satellite signal has strong dependence and poor robustness, and the existing scheme is seriously dependent on the continuous availability of the satellite signal. In the situations that satellite signals such as urban canyons, tunnels, underground parking lots, under overhead bridges and the like are blocked or completely lost, the system cannot acquire new PPS signals for calibration. At this time, the accumulated error of the internal clock increases rapidly, resulting in the failure of the time synchronization function, which seriously affects the reliability and safety of the system. (3) The existing scheme is mostly static or semi-static, and cannot be dynamically adjusted according to application scenes, environmental conditions or system states (such as vehicle speed and signal quality), so that the accuracy is ensured, and meanwhile, the energy efficiency of the system is difficult to consider. Therefore, a multi-sensor time synchronization method capable of overcoming the influence of clock drift, maintaining high accuracy when satellite signals are lost and having a certain intelligent adaptability is needed. Disclosure of Invention The invention aims to overcome the defects of the prior art, and provides a time synchronization method and a time synchronization device for a positioning system, which can improve the accuracy and the robustness of the internal time of equipment in time synchronization. An embodiment of the present invention provides a time synchronization method for a positioning system, including the following steps: initializing an internal time source in the processor, continuously capturing a PPS pulse signal through the satellite module, recording a time stamp each time a rising edge of the PPS pulse signal is captured, and establishing an initial conversion relation according to the time stamp, wherein the time stamp comprises UTC time and the internal time of the processor, and the initial conversion relation is used for describing the conversion relation between the UTC time and the internal time; Establishing a Kalman filtering model, dynamically updating the Kalman filtering model according to the timestamp recorded at the latest time, and updating the initial conversion relation in real time according to the Kalman filtering model obtained by the latest dynamic update to obtain an accurate conversion relation, wherein the updating period of the real-time update is independently set; and evaluating the health degree of the internal time source, determining a synchronous mode according to the health degree, and performing time synchronization on the current internal time according to the synchronous mode and the accurate conversion relation. Further, each time the rising edge of the PPS pulse signal is captured, the processor records a time stamp, and establishes an initial conversion relationship according to the time stamp, which specifically includes: receiving GNSS signals sent by the satellite module through the internal time source, and acquiring the PPS pulse signals and the UTC time; Each time a rising edge of the PPS pulse signal is ca