Pulse per second time synchronization method of optical fiber strapdown inertial navigation and GNSS integrated navigation system

Abstract

The invention provides a second pulse time synchronization method of an optical fiber strapdown inertial navigation and GNSS integrated navigation system. The pulse per second time synchronization method of the optical fiber strapdown inertial navigation and GNSS integrated navigation system comprises the following steps: S1, signal monitoring and time sequence initialization: starting PPS signal monitoring by an FPGA depending on a 100MHz internal working clock, scanning signals by utilizing a double-edge trigger detection circuit, filtering burrs by combining a digital filtering window, detecting PPS signal period jitter and duty ratio conditions, and performing time sequence initialization; and if the period jitter exceeds + / -10% or the duty ratio is abnormal, automatically switching to an internal crystal oscillator time keeping mode. According to the second pulse time synchronization method of the optical fiber strapdown inertial navigation and GNSS integrated navigation system, the problem that the period of GNSS information with a whole second time scale sent by a satellite receiver is irregular or does not exist for a long time can be avoided, the fixity of the integrated navigation period is ensured, GNSS data is extrapolated and fitted to an inertial navigation data time point, and the precision of the integrated navigation system is improved. The time synchronization method of two kinds of sensor data is realized.

Description

Technical Field

[0001] This invention relates to the field of FPGA high-precision timing and cross-processor collaborative technology, and in particular to a second pulse time synchronization method for fiber optic strapdown inertial navigation and GNSS integrated navigation systems. Background Technology

[0002] FPGA is a product of further development based on devices such as programmable array logic, general-purpose array logic, and erasable programmable logic devices. It is a programmable logic chip capable of performing general-purpose functions; that is, it can be programmed to implement specific logic processing functions.

[0003] Fiber optic strapdown inertial navigation GNSS integrated navigation technology uses GNSS velocity and position as observation information. It utilizes Kalman filtering to address the system's measurement noise and system noise. Based on the noise characteristics, it uses statistical methods to estimate and compensate for the system's attitude, velocity, and position errors. This process can fully leverage the strengths of both, prevent error accumulation, ensure the system's anti-interference capability, and enable the system to provide high-precision navigation over long periods.

[0004] However, due to the satellite receiver's own calculation delay, the different time starting points of fiber optic strapdown inertial navigation and GNSS, and communication transmission delay, there is a time lag error between the inertial and satellite sensor information.

[0005] Therefore, it is necessary to provide a second pulse time synchronization method for fiber optic strapdown inertial navigation and GNSS integrated navigation systems to solve the above-mentioned technical problems. Summary of the Invention

[0006] This invention provides a second pulse time synchronization method for fiber optic strapdown inertial navigation and GNSS integrated navigation systems, solving the time asynchrony error caused by the relative time lag between inertial and satellite sensor information. To address the aforementioned technical problem, the second pulse time synchronization method for fiber optic strapdown inertial navigation and GNSS integrated navigation systems provided by this invention includes the following methods:

[0007] S1: Signal Monitoring and Timing Initialization: The FPGA relies on its 100MHz internal clock to enable PPS signal monitoring. It uses a dual-edge triggered detection circuit to scan the signal and combines a digital filter window to filter glitches. It detects the PPS signal period jitter and duty cycle. If the period jitter exceeds ±10% or the duty cycle is abnormal, it automatically switches to the internal crystal oscillator timekeeping mode, records the abnormality, and maintains timing according to the TCXO reference with a timekeeping accuracy of <1ms / min, waiting for PPS to recover. If normal, it determines that the PPS has a valid transition and enters the time reference layer operation. If no valid transition is detected, it continues to monitor.

[0008] S2: Time base construction and timing start: The TCXO is used as the basic clock source. The FPGA internal PLL multiplies the clock to generate a 100MHz working clock as the time base. When the PPS effective transition is captured, the FPGA immediately clears the internal microsecond counter to zero, and then starts the Time1 accumulation timing to enter the data interaction preparation stage.

[0009] S3: Inertial Navigation Data Storage and Cycle Control: Storage Cycle Setting: Based on the fiber optic strapdown inertial navigation calculation frequency, such as 200Hz, the storage time window is planned to specify how long to store the data. After the DSP responds to the FPGA hardware interrupt and starts the real-time storage task, it marks the stored inertial navigation data with a "timestamp" in real time, associates it with the FPGA's Time1 timing value, and records the duration of each data packet after PPS triggering. When the fixed cycle of the integrated navigation is reached, the storage of the current round of inertial navigation data is stopped, and the inertial acceleration and velocity data sequence within the storage cycle is sealed, clearly defining "what to store and when to stop".

[0010] S4: Cross-clock domain synchronous trigger: After the FPGA detects a valid PPS transition and starts Time1 timing, it immediately sends a trigger signal to the DSP through the hardware interrupt line. The DSP responds to the hardware interrupt and starts the real-time storage task according to the rules of inertial navigation data storage and cycle control. According to the fiber optic strapdown inertial navigation solution cycle, it continuously stores the acceleration and velocity data of the inertial navigation solution, which is recorded as the inertial navigation data sequence, i = 1, 2, 3...n, and enters the data temporary storage stage.

[0011] S5: Timing Value Transmission and Data Preprocessing: When the fixed period of fiber optic strapdown inertial navigation and GNSS integrated navigation arrives, the DSP actively reads the Time1 timing value from the FPGA. After receiving Time1, the DSP stops the real-time storage of inertial navigation acceleration and velocity data according to the rules of inertial navigation data storage and period control. At the same time, it retrieves the latest GNSS data packet from the communication interface, starts CRC-32 verification to ensure data integrity, and handles anomalies and ensures robustness - Data Interaction Anomaly Handling: If the CRC verification of GNSS_data0 fails, the historical data prediction queue is activated, and the current time stamp information is completed using linear regression. After marking the data as abnormal, the subsequent process is entered; if the verification passes, the millisecond-level fragmented time Time2 of the time stamp is extracted normally.

[0012] S6: Time Delay Compensation and Fusion Solution Closed Loop: Combining Time1 obtained from FPGA and Time2 of GNSS_data0, ΔT is calculated by substituting them into the time delay model; the validity of ΔT is determined. If ΔT ≥ 1s, an abnormal response is triggered; otherwise, the process continues. Based on ΔT and the stored inertial navigation data, an extrapolation model with angular velocity compensation is used to extrapolate the original GNSS velocity and position to the current inertial navigation time point, aligning the GNSS data with the inertial navigation data. The extrapolated GNSS data is used to construct a Kalman filter measurement vector. Combined with the inertial navigation state equation, Kalman filter time update and measurement update are performed to correct the inertial navigation system error. The fused integrated navigation information is output, completing the "time delay compensation-fusion solution" closed loop.

[0013] S7: Anomaly Handling and Robustness Guarantee: Embedded signal monitoring to handle PPS signal anomalies, embedded data preprocessing to handle data interaction anomalies, and embedded fusion solution to handle solution anomalies;

[0014] The FPGA utilizes an internal high-frequency operating clock, multiplied to 100MHz by a temperature-compensated crystal oscillator (TCXO) via a phase-locked loop (PLL), achieving a frequency stability of ±0.1ppm. It employs a dual-edge trigger detection circuit with rising and falling edge triggering, coupled with a configurable digital filter window of 200ns-1μs, to accurately identify valid PPS transitions, laying the foundation for subsequent time synchronization. From precise PPS signal identification to 10ns resolution timing, microsecond-level cross-processor triggering, standardized storage of inertial navigation data, and end-to-end anomaly protection, the entire process, including delay compensation and fusion calculation, outputs high-precision navigation information. This entire process allows fiber-optic strapdown inertial navigation and GNSS integrated navigation systems to break free from the traditional reliance on whole-second timescales.

[0015] Preferably, the PPS signal anomaly handling is as follows: when the signal monitoring layer detects that the PPS signal period jitter exceeds ±10% or the duty cycle is abnormal, it automatically switches to the internal crystal oscillator timekeeping mode, maintains timing according to the TCXO reference, and waits for the PPS to recover; the timekeeping accuracy is <1ms / min, which is better than the traditional 5ms / min, avoiding the PPS anomaly from destroying the time reference and ensuring that the system can maintain basic time synchronization capability when the GNSS signal is poor; the data interaction anomaly handling is as follows: when the CRC check of GNSS_data0 fails, the historical data prediction queue is activated, the current time stamp information is completed with linear regression, and after marking the data anomaly, it enters the subsequent process, intercepting transmission errors, data packet corruption and other problems, and using historical data as a "backup" to avoid the integrated navigation being interrupted due to single packet data anomalies, ensuring that the system "does not stop calculating and has fewer errors" in harsh communication environments; the calculation anomaly handling is as follows: when the time delay ΔT>300ms, the measurement noise variance matrix Rk of the Kalman filter is automatically increased, the weight of abnormal observations is reduced, and the abnormal observations are prevented from causing the filter to diverge, ensuring the long-term stable operation of the system.

[0016] Preferably, the time delay compensation and fusion solution closed loop includes time delay calculation and validity determination, dynamic extrapolation fitting of GNSS data and Kalman filter fusion solution;

[0017] By using "angular velocity coupling term compensation" and "high dynamic extrapolation model", the velocity extrapolation error is reduced from the traditional 0.15m / s to 0.02m / s when ΔT = 500ms and angular velocity ω = 10° / s, directly demonstrating the invention's adaptability to high dynamic scenarios.

[0018] Preferably, the delay calculation and validity determination includes multi-source time information integration and a two-layer validity determination;

[0019] Multi-source time information integration: Input data: FPGA side: timing value Time1, recording the time difference between "PPS transition - arrival of combined cycle", GNSS side: millisecond-level fragment time Time2 extracted from data packet GNSS_data0, reflecting the "deviation between GNSS data and the most recent whole second", calibration parameter: transmission delay compensation amount T_tr, validity double-layer judgment: basic judgment: if ΔT<1s (corresponding to the validity of PPS cycle, the delay is judged to be valid, and the compensation process is entered.

[0020] Preferably, the dynamic extrapolation fitting of GNSS data includes extrapolation model decomposition and position extrapolation simplification logic, and the Kalman filter fusion solution includes review of state equations and measurement equations, dynamic adjustment of the filtering process, and error feedback and correction.

[0021] By fusing raw GNSS data with inertial navigation stored acceleration, angular velocity, and other information, the velocity and position, aligned with the inertial navigation time, are obtained through dynamic extrapolation fitting with an angular velocity compensation term.

[0022] Preferably, the fiber optic strapdown inertial navigation system includes: a shock-absorbing base frame;

[0023] The outer casing is fixedly connected to the top of the shock-absorbing base frame. A partition is installed in the middle of the interior of the outer casing. Multiple mating strips are installed on both sides of the inner wall of the outer casing. A core mounting frame is installed at the bottom of the interior of the outer casing through a mating interface. A general mounting frame is installed near the top of the interior of the outer casing through a mating interface. Module interfaces are provided on the front of both the core mounting frame and the general mounting frame. Heat-conducting cover plates are installed on the top of both the core mounting frame and the general mounting frame.

[0024] The core mounting frame is located below the partition, and the ordinary mounting frame is located above the partition. The core mounting frame is made of zero-expansion microcrystalline glass, while the ordinary mounting frame is made of ceramic. The partition is made of titanium alloy sheet. The core mounting frame houses the core sensitive element layer, namely the fiber optic gyroscope and accelerometer. The ordinary mounting frame houses the signal conditioning and processing circuit layer. All components inside the core mounting frame and the ordinary mounting frame are connected to the module interface.

[0025] Preferably, a front cover is installed on the front of the housing, an external panel is installed on the front of the front cover, a rear cover is installed on the back of the housing, a circuit board with a flow port is installed on the front of the rear cover, and multiple module plugs are installed on the front of the circuit board.

[0026] There are gaps between the top and back of the circuit board and the top and back of the inner wall of the back cover. The air vents allow air to flow between the bottom and top of the partition, and the filter and heat dissipation components help to dissipate heat.

[0027] Preferably, a heat dissipation component with a protective plate is installed on the back of the rear cover near the top, and a filter component is installed on the back of the rear cover near the middle.

[0028] The filter element includes a frame and a structure for filtering dust and moisture.

[0029] Preferably, the heat-conducting cover plate includes a top cover and a heat-conducting substrate, and both the core mounting frame and the ordinary mounting frame have mounting openings on their back sides, with monitoring components installed inside each mounting opening;

[0030] Each top cover has multiple heat-conducting substrates installed at the bottom. These substrates are inserted between the components. The heat-conducting covers are made of highly thermally conductive materials, and the monitoring components can monitor the temperature in real time.

[0031] Preferably, the shock-absorbing base frame includes a base plate, a connecting block, and a shock-absorbing component, wherein the shock-absorbing component is installed between the top of the base plate and the bottom of the outer shell;

[0032] The mating blocks are located on both sides of the base plate.

[0033] Compared with related technologies, the second pulse time synchronization method for fiber optic strapdown inertial navigation and GNSS integrated navigation systems provided by this invention has the following advantages:

[0034] This invention provides a second pulse timing method for fiber optic strapdown inertial navigation and GNSS integrated navigation systems. To reduce errors in the use of this method, at the hardware level, the system constructs a dual-processor collaborative architecture of FPGA and DSP, establishing a high-precision time reference through a three-level timing control chain: the signal monitoring layer employs FPGA dual-edge triggered detection and configurable digital filtering (200ns-1μs), which can accurately capture effective PPS transitions even under strong electromagnetic interference; the time reference layer uses a ±0.1ppm temperature-compensated crystal oscillator (TCXO) as the source, which is frequency multiplied to 100MHz by an FPGA phase-locked loop to achieve 10ns resolution timing (far exceeding traditional methods). A 1μs counter is used; the data interaction layer transmits data via EMIF bus and DMA, writing the 32-bit timing value (high 16 bits for seconds count, low 16 bits for nanosecond count) to the DSP with a delay of less than 500ns. Combined with a hardware interrupt triggering mechanism, the delay from PPS transition to DSP storage of inertial navigation data is controlled within 2μs (traditional solutions > 50μs), eliminating cross-clock domain synchronization errors from the bottom layer. At the algorithm level, an innovative delay calculation and dynamic extrapolation model is designed: based on the FPGA timing value (Time1) and GNSS millisecond-level fragmented time (Time2), a delay model of ΔT = T1 + (T2 - LT2) - Ttr is constructed to decompose measurement error and transmission delay. To address quantization errors, and for high-dynamic scenarios, the extrapolation algorithm introduces an angular velocity compensation term [(ωibb×).(Cbn)T.vGNSSn]ΔT. Combined with the inertial navigation acceleration mean and rate of change, the GNSS velocity and position are extrapolated to the current inertial navigation time point, resolving the time asynchrony issue between the two types of sensors. Simultaneously, through PPS periodic verification (switching to a crystal oscillator with a speed of <1ms / min when jitter exceeds ±10%), GNSS data CRC-32 verification (using historical data prediction upon failure), and Kalman filter dynamic adjustment (increasing the measurement noise variance when ΔT>300ms), a full-link anti-interference mechanism is formed. This design effectively estimates and compensates for observation information. The second pulse timing method for delay error improves the accuracy of the integrated navigation system. It utilizes the high-frequency characteristics of the FPGA system clock to ensure the accuracy and real-time performance of the second pulse signal monitoring and the high precision of the timing function. At the same time, it reduces the workload of the application layer processor DSP, ensuring the continuous and reliable operation of all functions of the fiber optic strapdown inertial navigation system / GNSS integrated navigation system. It does not rely on GNSS information with whole-second time stamps, avoiding the problem of irregular periods or long periods of absence of GNSS information with whole-second time stamps sent by satellite receivers. It also ensures the fixity of the integrated navigation period. The GNSS data is extrapolated and fitted to the inertial navigation data time points to achieve time synchronization of the two sensor data. Attached Figure Description

[0035] Figure 1This is a schematic diagram of the first embodiment of the second pulse time synchronization method for the fiber optic strapdown inertial navigation and GNSS integrated navigation system provided by the present invention.

[0036] Figure 2 This invention provides a schematic diagram of the time delay compensation and fusion solution closed loop;

[0037] Figure 3 This invention provides a schematic diagram of multi-source time information integration;

[0038] Figure 4 This invention provides a schematic diagram of dynamic extrapolation and fitting of GNSS data;

[0039] Figure 5 This invention provides a schematic diagram of Kalman filter fusion solution;

[0040] Figure 6 This invention provides a handling process for PPS anomalies;

[0041] Figure 7 The timing flowchart provided for this invention;

[0042] Figure 8 The software flowchart provided for this invention;

[0043] Figure 9 This invention provides a comparison chart of speed errors between integrated navigation results;

[0044] Figure 10 This invention provides a comparison chart of positional errors between integrated navigation results;

[0045] Figure 11 This is a schematic diagram of the second embodiment of the second pulse time synchronization method for the fiber optic strapdown inertial navigation and GNSS integrated navigation system provided by the present invention.

[0046] Figure 12 A schematic diagram of the heat dissipation component is provided for this invention;

[0047] Figure 13 A schematic diagram of the partition structure is provided for this invention;

[0048] Figure 14 A schematic diagram of the structure of the thermally conductive substrate is provided for this invention;

[0049] Figure 15 Provided for the present invention Figure 13 An enlarged view of point A shown.

[0050] The diagram is labeled as follows: 1. Vibration damping base frame; 101. Base plate; 102. Connecting block; 103. Vibration damping component; 2. Outer shell; 3. Rear cover; 4. Front cover; 5. External panel; 6. Protective plate; 7. Heat dissipation component; 8. Core mounting frame; 9. Monitoring component; 10. Mounting port; 11. Module plug; 12. Circuit board; 13. Flow port; 14. Thermal conductive cover plate; 141. Top cover; 142. Thermal conductive substrate; 15. Connecting strip; 16. Partition; 17. Ordinary mounting frame; 18. Module interface; 19. Filter component; 20. Connecting interface. Detailed Implementation

[0051] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0052] Please refer to the following: Figure 1 , Figure 2 , Figure 3 , Figure 4 , Figure 5 , Figure 6 , Figure 7 , Figure 8 , Figure 9 and Figure 10 ,in, Figure 1 This is a schematic diagram of the first embodiment of the second pulse time synchronization method for the fiber optic strapdown inertial navigation and GNSS integrated navigation system provided by the present invention. Figure 2 This invention provides a schematic diagram of the time delay compensation and fusion solution closed loop; Figure 3 This invention provides a schematic diagram of multi-source time information integration; Figure 4 This invention provides a schematic diagram of dynamic extrapolation and fitting of GNSS data; Figure 5 This invention provides a schematic diagram of Kalman filter fusion solution; Figure 6 This invention provides a handling process for PPS anomalies; Figure 7 The timing flowchart provided for this invention; Figure 8 The software flowchart provided for this invention; Figure 9 This invention provides a comparison chart of speed errors between integrated navigation results; Figure 10 This invention provides a comparison chart of position errors between integrated navigation results. The second pulse time synchronization method for fiber optic strapdown inertial navigation and GNSS integrated navigation systems includes the following methods:

[0053] S1: Signal Monitoring and Timing Initialization: The FPGA relies on its 100MHz internal clock to enable PPS signal monitoring. It uses a dual-edge triggered detection circuit to scan the signal and combines a digital filter window to filter glitches. It detects the PPS signal period jitter and duty cycle. If the period jitter exceeds ±10% or the duty cycle is abnormal, it automatically switches to the internal crystal oscillator timekeeping mode, records the abnormality, and maintains timing according to the TCXO reference with a timekeeping accuracy of <1ms / min, waiting for PPS to recover. If normal, it determines that the PPS has a valid transition and enters the time reference layer operation. If no valid transition is detected, it continues to monitor.

[0054] S2: Time base construction and timing start: The TCXO is used as the basic clock source. The FPGA internal PLL multiplies the clock to generate a 100MHz working clock as the time base. When the PPS effective transition is captured, the FPGA immediately clears the internal microsecond counter to zero, and then starts the Time1 accumulation timing to enter the data interaction preparation stage.

[0055] S3: Inertial Navigation Data Storage and Cycle Control: Storage Cycle Setting: Based on the fiber optic strapdown inertial navigation calculation frequency, such as 200Hz, the storage time window is planned to specify how long to store the data. After the DSP responds to the FPGA hardware interrupt and starts the real-time storage task, it marks the stored inertial navigation data with a "timestamp" in real time, associates it with the FPGA's Time1 timing value, and records the duration of each data packet after PPS triggering. When the fixed cycle of the integrated navigation is reached, the storage of the current round of inertial navigation data is stopped, and the inertial acceleration and velocity data sequence within the storage cycle is sealed, clearly defining "what to store and when to stop".

[0056] S4: Cross-clock domain synchronous trigger: After the FPGA detects a valid PPS transition and starts Time1 timing, it immediately sends a trigger signal to the DSP through the hardware interrupt line. The DSP responds to the hardware interrupt and starts the real-time storage task according to the rules of inertial navigation data storage and cycle control. According to the fiber optic strapdown inertial navigation solution cycle, it continuously stores the acceleration and velocity data of the inertial navigation solution, which is recorded as the inertial navigation data sequence, i = 1, 2, 3...n, and enters the data temporary storage stage.

[0057] S5: Timing Value Transmission and Data Preprocessing: When the fixed period of fiber optic strapdown inertial navigation and GNSS integrated navigation arrives, the DSP actively reads the Time1 timing value from the FPGA. After receiving Time1, the DSP stops the real-time storage of inertial navigation acceleration and velocity data according to the rules of inertial navigation data storage and period control. At the same time, it retrieves the latest GNSS data packet from the communication interface, starts CRC-32 verification to ensure data integrity, and handles anomalies and ensures robustness - Data Interaction Anomaly Handling: If the CRC verification of GNSS_data0 fails, the historical data prediction queue is activated, and the current time stamp information is completed using linear regression. After marking the data as abnormal, the subsequent process is entered; if the verification passes, the millisecond-level fragmented time Time2 of the time stamp is extracted normally.

[0058] S6: Time Delay Compensation and Fusion Solution Closed Loop: Combining Time1 obtained from FPGA and Time2 of GNSS_data0, ΔT is calculated by substituting them into the time delay model; the validity of ΔT is determined. If ΔT ≥ 1s, an abnormal response is triggered; otherwise, the process continues. Based on ΔT and the stored inertial navigation data, an extrapolation model with angular velocity compensation is used to extrapolate the original GNSS velocity and position to the current inertial navigation time point, aligning the GNSS data with the inertial navigation data. The extrapolated GNSS data is used to construct a Kalman filter measurement vector. Combined with the inertial navigation state equation, Kalman filter time update and measurement update are performed to correct the inertial navigation system error. The fused integrated navigation information is output, completing the "time delay compensation-fusion solution" closed loop.

[0059] S7: Anomaly Handling and Robustness Guarantee: Embedded signal monitoring to handle PPS signal anomalies, embedded data preprocessing to handle data interaction anomalies, and embedded fusion solution to handle solution anomalies;

[0060] The FPGA utilizes an internal high-frequency operating clock, multiplied to 100MHz by a temperature-compensated crystal oscillator (TCXO) via a phase-locked loop (PLL), achieving a frequency stability of ±0.1ppm. It employs a dual-edge trigger detection circuit with rising and falling edge triggering, coupled with a configurable digital filter window of 200ns-1μs, to accurately identify valid PPS transitions, laying the foundation for subsequent time synchronization. From precise PPS signal identification to 10ns resolution timing, microsecond-level cross-processor triggering, standardized storage of inertial navigation data, and end-to-end anomaly protection, the entire process, including delay compensation and fusion calculation, outputs high-precision navigation information. This entire process allows fiber-optic strapdown inertial navigation and GNSS integrated navigation systems to break free from the traditional reliance on whole-second timescales.

[0061] Attitude error equation:

[0062]

[0063] Velocity error equation:

[0064]

[0065] Position error equation:

[0066] in, The misalignment angle errors are respectively ε n This represents the measurement error of the gyroscope in the navigation coordinate system. and These are the Earth's rotation speed and position rate, respectively. and These are the errors in Earth's rotation speed and position rate, respectively, V n The velocities in the navigation coordinate system are V N V W V U , The velocity errors in the navigation coordinate system are respectively f n This indicates the measurement performed in the navigation coordinate system. This represents the measurement error added in the navigation coordinate system, where L represents the local latitude, δλ, δL, and δh represent the longitude, latitude, and altitude errors, respectively, and R represents the measurement error added in the navigation coordinate system. Mh and R Nh Given the sum of Earth's radius and altitude, and based on the above systematic error equations, a 15th-order state equation is established with misalignment angle, velocity, position, accelerometer constant bias, and gyroscope constant drift as state variables X, along with measurement equations with velocity and position as observations. This serves as the Kalman filter state-space model for the fiber optic strapdown inertial navigation / GNSS combination:

[0067] State equation: x = FX + W Observation equation: Z = HX + V

[0068] The transition matrix F for each consecutive step is obtained from the system error equation, and the state vector X is as follows: Taking the velocity error and position error calculated from fiber optic strapdown inertial navigation and GNSS as the observed quantities, the measurement vector Z is: Z = [δV N δV W δv U δ L δλ δ h ] T The observation matrix is: H = [0 6×3 1 6×6 0 6×6 W is the system noise vector, and v is the observation noise vector; a hardware interrupt triggers the DSP real-time storage task, which begins calculating and storing the inertial velocity of each fiber strapdown inertial navigation system, with a calculation cycle T. INS acceleration and speed MpvV i , where i = 1, 2, 3…n;

[0069] Among them, f i b For accelerometer output; C i The attitude transformation matrix for this navigation cycle ΔV rotm_i This is the compensation amount for the rotational effect; ΔV sculm_i Paddle effect compensation amount; ΔV glcorm_i This is the velocity compensation amount caused by harmful acceleration;

[0070] MpvV i =Mpv i *V i =[v Mpv_Ni v Mpv_wi v Mpv_Ui ]

[0071]

[0072] Where L represents the local latitude; R Mh and R Nh It is the sum of the Earth's radius and altitude; The north-west-sky velocity calculated in the previous navigation solution cycle; V i n The north-west-sky speed calculated for this navigation solution cycle;

[0073] Speed ​​delay error:

[0074]

[0075] Location delay error:

[0076]

[0077] in, The north-west-sky velocity in the received GNSS_data0 data, and the time corresponding to this velocity is the time stamp in the GNSS_data0 data; V N_GNSS V W_GNSS V U_GNSS The estimated GNSS North-West-Sky velocity corresponding to the current moment in the inertial navigation information; The longitude, latitude, and altitude of the received GNSS_data0 data are represented by the time stamp in the GNSS_data0 data; λ GNSS L GNSS H GNSS This is an estimate of the GNSS longitude, latitude, and altitude corresponding to the current moment in the inertial navigation information; North-West-Sky Acceleration for storing inertial navigation solution cycles The mean; To store the north-west-sky velocity MpvV of the inertial navigation solution cycle i The mean.

[0078] The PPS signal anomaly handling mechanism is as follows: When the signal monitoring layer detects a PPS signal period jitter exceeding ±10% or an abnormal duty cycle, it automatically switches to the internal crystal oscillator timekeeping mode, maintaining timing according to the TCXO reference until the PPS recovers; the timekeeping accuracy is <1ms / min, which is better than the traditional 5ms / min, avoiding PPS anomalies from disrupting the time reference and ensuring that the system can maintain basic time synchronization capability even when the GNSS signal is poor. The data interaction anomaly handling mechanism is as follows: When the CRC check of GNSS_data0 fails, the historical data prediction queue is activated, and the current time stamp information is completed using linear regression. After marking the data anomaly, it enters the subsequent process, intercepting transmission errors, data packet corruption, and other problems, using historical data as a "backup" to avoid the integrated navigation being interrupted due to a single packet data anomaly, ensuring that the system can "calculate continuously and with fewer errors" in harsh communication environments. The calculation anomaly handling mechanism is as follows: When the time delay ΔT > 300ms, the measurement noise variance matrix Rk of the Kalman filter is automatically increased, and the weight of abnormal observations is reduced to prevent abnormal observations from causing the filter to diverge, ensuring the system's long-term stable operation.

[0079] Anomaly response and robustness assurance are core mechanisms in the design and operation of systems, including software, hardware, industrial processes, and AI models. They complement each other and work together to ensure that the system functions stably in complex, dynamic, or uncontrollable environments.

[0080] The time delay compensation and fusion solution closed loop includes time delay calculation and validity determination, dynamic extrapolation and fitting of GNSS data, and Kalman filter fusion solution;

[0081] By using "angular velocity coupling term compensation" and "high dynamic extrapolation model", the velocity extrapolation error is reduced from the traditional 0.15m / s to 0.02m / s when ΔT = 500ms and angular velocity ω = 10° / s, directly demonstrating the invention's adaptability to high dynamic scenarios.

[0082] The delay calculation and validity determination include multi-source time information integration and a two-layer validity determination.

[0083] Multi-source time information integration: Input data: FPGA side: timing value Time1, recording the time difference between "PPS transition - arrival of combined cycle"; GNSS side: millisecond-level fragmented time Time2 extracted from data packet GNSS_data0, reflecting the "deviation between GNSS data and the most recent whole second"; calibration parameters: transmission delay compensation amount T_tr; validity dual-layer judgment: basic judgment: if ΔT < 1s (corresponding to PPS cycle validity, the delay is judged to be valid, and the compensation process is entered; otherwise, an abnormal response is triggered; dynamic judgment: if 300ms < ΔT ≤ 1s, "delay is too large" is marked, preparing for subsequent Kalman filter adjustment of measurement noise; if ΔT ≤ 300ms, compensation is performed according to the normal process.

[0084] The dynamic extrapolation and fitting of GNSS data includes extrapolation model decomposition and position extrapolation simplification logic, and the Kalman filter fusion solution includes review of state equations and measurement equations, dynamic adjustment of the filtering process, and error feedback and correction.

[0085] By fusing raw GNSS data with inertial navigation stored acceleration, angular velocity, and other information, and then using dynamic extrapolation fitting with angular velocity compensation terms to obtain velocity and position aligned with inertial navigation time, the data is then fused and solved using a 15th-order Kalman filter for the state vector. The error terms are then fed back to correct the fiber optic strapdown inertial navigation output, thereby improving accuracy.

[0086] The working principle of the second pulse time synchronization method for fiber optic strapdown inertial navigation and GNSS integrated navigation systems provided by this invention is as follows:

[0087] At the hardware level, the system employs a dual-processor architecture of FPGA and DSP, establishing a high-precision time reference through a three-level timing control chain: the signal monitoring layer utilizes FPGA dual-edge triggered detection and configurable digital filtering (200ns-1μs), enabling accurate capture of effective PPS transitions even under strong electromagnetic interference; the time reference layer uses a ±0.1ppm temperature-compensated crystal oscillator (TCXO) as its source, multiplied to 100MHz by an FPGA phase-locked loop, achieving 10ns resolution timing (far exceeding the traditional 1μs counter); the data interaction layer uses EMIF bus and DMA transmission to write 32-bit timing values ​​(high 16 bits for seconds count, low 16 bits for nanosecond count) to the DSP with a delay of less than 500ns. Combined with a hardware interrupt triggering mechanism, this keeps the delay from PPS transitions to the DSP storing inertial navigation data within 2μs (traditional solutions >50μs), eliminating cross-clock domain synchronization errors from the ground up. At the algorithm level, innovative delay calculation and dynamic extrapolation models are designed. Model: Based on FPGA timing values ​​(Time1) and GNSS millisecond-level fragmented time (Time2), a time delay model of ΔT = T1 + (T2 - LT2) - Ttr is constructed to decompose measurement error, transmission delay, and quantization error. For high dynamic scenarios, an extrapolation algorithm introduces an angular velocity compensation term [(ωibb × ). (Cbn)T. vGNSSn]ΔT, combined with the inertial navigation acceleration mean and acceleration change rate, to extrapolate GNSS velocity and position to the current inertial navigation time point, solving the time asynchrony problem between the two types of sensors. At the same time, through PPS periodic verification (switching to crystal oscillator timekeeping of <1ms / min when jitter exceeds ±10%), GNSS data CRC-32 verification (using historical data prediction when failure), and Kalman filter dynamic adjustment (increasing the measurement noise variance when ΔT > 300ms), a full-link anti-interference mechanism is formed. In actual use, in high dynamic scenarios (such as UAV 3g acceleration climb, 8 / s 2In angular velocity hovering, the position error was reduced from 15.2m (RMS) in the traditional scheme to 2.3m (RMS), and the attitude convergence time was shortened by 60%. After the GNSS signal was interrupted for 10 seconds and then restored, it could converge to 1m accuracy within 3 seconds (the traditional scheme required more than 8 seconds). During long-term operation, the FPGA timing drift was <1μs / h, and the standard deviation of the horizontal position error was stable within 0.8m (the standard deviation of the traditional scheme increased to 5m+ over time). Moreover, the DSP load was only 12.4% and the FPGA resource utilization rate was 35.2%, which balanced accuracy and stability. In extreme environments, the time delay deviation is <±0.5μs / ℃ during temperature cycling from -40℃ to +70℃, and the PPS false trigger rate is <0.1 times / hour under electromagnetic interference of 200V / m, fully meeting the high-precision requirements of integrated navigation in complex scenarios. In addition, the system breaks through the dependence of traditional solutions on GNSS whole-second time scales and uses millisecond-level fragmented time to achieve time delay compensation of >1s (measured upper limit of 1.5s), avoiding the calculation interruption caused by irregular satellite information cycles and ensuring the fixity and continuity of the integrated navigation cycle.

[0088] Compared with related technologies, the second pulse time synchronization method for fiber optic strapdown inertial navigation and GNSS integrated navigation systems provided by this invention has the following advantages:

[0089] To reduce errors in the use of the second pulse timing method in fiber optic strapdown inertial navigation and GNSS integrated navigation systems, the system employs a dual-processor architecture of FPGA and DSP at the hardware level. A high-precision time reference is established through a three-level timing control chain: the signal monitoring layer uses FPGA dual-edge triggered detection and configurable digital filtering (200ns-1μs), enabling accurate capture of effective PPS transitions even under strong electromagnetic interference; the time reference layer uses a ±0.1ppm temperature-compensated crystal oscillator (TCXO) as its source, multiplied to 100MHz by an FPGA phase-locked loop to achieve 10ns resolution timing (far exceeding the traditional 1μs counter); the data interaction layer communicates with... via an EMIF bus... DMA transfer writes 32-bit timing values ​​(high 16 bits for seconds and low 16 bits for nanoseconds) to the DSP with a delay of less than 500ns. Combined with a hardware interrupt triggering mechanism, this keeps the delay from PPS transition to DSP-stored inertial navigation data within 2μs (traditional solutions >50μs), eliminating cross-clock domain synchronization errors at the underlying level. At the algorithm level, an innovative delay calculation and dynamic extrapolation model is designed: based on the FPGA timing value (Time1) and GNSS millisecond-level fragmented time (Time2), a delay model of ΔT = T1 + (T2 - LT2) - Ttr is constructed to decompose measurement error, transmission delay, and quantization error. For high-dynamic scenarios, the extrapolation... The extrapolation algorithm introduces an angular velocity compensation term [(ωibb×).(Cbn)T.vGNSSn]ΔT, combining the inertial navigation acceleration mean and acceleration change rate to extrapolate GNSS velocity and position to the current inertial navigation time point, thus solving the time asynchrony problem between the two types of sensors. Simultaneously, through PPS period verification (switching to crystal oscillator timing <1ms / min when jitter exceeds ±10%), GNSS data CRC-32 verification (using historical data prediction when failure occurs), and Kalman filter dynamic adjustment (increasing the measurement noise variance when ΔT>300ms), a full-link anti-interference mechanism is formed. This design effectively estimates and compensates for the second pulse pair of observation information delay errors. This time synchronization method improves the accuracy of the integrated navigation system's calculations. It utilizes the high-frequency characteristics of the FPGA system clock to ensure the accuracy and real-time performance of the second pulse signal monitoring and the high precision of the timing function. At the same time, it reduces the workload of the application layer processor DSP, ensuring the continuous and reliable operation of all functions of the fiber optic strapdown inertial navigation system / GNSS integrated navigation system. It does not rely on GNSS information with whole-second time stamps, avoiding the problem of irregular periods or long periods of absence of GNSS information with whole-second time stamps sent by the satellite receiver. It also ensures the fixedness of the integrated navigation period. The GNSS data is extrapolated and fitted to the inertial navigation data time points to achieve time synchronization of the two sensor data.

[0090] Second Embodiment

[0091] Please refer to the following: Figures 11-12 - Figures 13-14 - Figure 15 , Figure 11 This is a schematic diagram of the second embodiment of the second pulse time synchronization method for the fiber optic strapdown inertial navigation and GNSS integrated navigation system provided by the present invention. Figure 12 A schematic diagram of the heat dissipation component is provided for this invention; Figure 13 A schematic diagram of the partition structure is provided for this invention; Figure 14 A schematic diagram of the structure of the thermally conductive substrate is provided for this invention; Figure 15 Provided for the present invention Figure 13 The enlarged view at point A is shown. Based on the second pulse time synchronization method for fiber optic strapdown inertial navigation and GNSS integrated navigation systems provided in the first embodiment of this application, the second embodiment of this application proposes another second pulse time synchronization method for fiber optic strapdown inertial navigation and GNSS integrated navigation systems. The second embodiment is merely a preferred embodiment of the first embodiment, and the implementation of the second embodiment will not affect the separate implementation of the first embodiment.

[0092] Specifically, the difference between the fiber optic strapdown inertial navigation and GNSS integrated navigation system second pulse time synchronization method provided in the second embodiment of this application is that the fiber optic strapdown inertial navigation system includes: a shock-absorbing base frame 1;

[0093] The outer shell 2 is fixedly connected to the top of the shock-absorbing base frame 1. A partition 16 is installed in the middle of the interior of the outer shell 2. Multiple mating strips 15 are installed on both sides of the inner wall of the outer shell 2. A core mounting frame 8 is installed at the bottom of the interior of the outer shell 2 through a mating interface 20. A general mounting frame 17 is installed near the top of the interior of the outer shell 2 through a mating interface 20. Module interfaces 18 are provided on the front of both the core mounting frame 8 and the general mounting frame 17. A heat-conducting cover plate 14 is installed on the top of both the core mounting frame 8 and the general mounting frame 17.

[0094] The core mounting frame 8 is located below the partition 16, and the ordinary mounting frame 17 is located above the partition 16. The core mounting frame 8 is made of zero-expansion microcrystalline glass, and the ordinary mounting frame 17 is made of ceramic. The partition 16 is a titanium alloy sheet. The core mounting frame 8 is equipped with a core sensitive element layer, namely a fiber optic gyroscope and an accelerometer. The ordinary mounting frame 17 is equipped with a signal conditioning and processing circuit layer. The internal components of the core mounting frame 8 and the ordinary mounting frame 17 are connected to the module interface 18. There is a certain gap between the bottom of the partition 16 and the top of the core mounting frame 8 heat-conducting cover plate 14. There is also a certain gap between the top of the ordinary mounting frame 17 heat-conducting cover plate 14 and the top of the inner wall of the outer shell 2. The mating strip 15 is inserted into the inside of the mating interface 20.

[0095] The front cover 4 is installed on the front of the housing 2, the external panel 5 is installed on the front of the front cover 4, the back cover 3 is installed on the back of the housing 2, the circuit board 12 with a flow port 13 is installed on the front of the back cover 3, and multiple module plugs 11 are installed on the front of the circuit board 12.

[0096] The top and back of the circuit board 12 have a certain gap with the top and back of the inner wall of the back cover 3. The airflow port 13 allows air to flow under and above the partition 16. The filter component 19 and heat dissipation component 7 assist in heat dissipation.

[0097] A heat dissipation component 7 with a protective plate 6 is installed on the back of the rear cover 3 near the top, and a filter component 19 is installed on the back of the rear cover 3 near the middle.

[0098] The filter component 19 includes a frame and a structure for filtering dust and moisture, the heat dissipation component 7 includes a frame and a cooling fan, and the protective plate 6 has holes for protection.

[0099] The heat-conducting cover plate 14 includes a top cover 141 and a heat-conducting substrate 142. The back of both the core mounting frame 8 and the ordinary mounting frame 17 are provided with mounting ports 10, and monitoring components 9 are installed inside the mounting ports 10.

[0100] Each top cover 141 has multiple heat-conducting substrates 142 installed at its bottom. The heat-conducting substrates 142 are inserted between the components. The heat-conducting cover plates 14 are all made of high thermal conductivity materials. The monitoring component 9 can monitor the temperature in real time.

[0101] The shock-absorbing base frame 1 includes a base plate 101, a docking block 102, and a shock-absorbing component 103, wherein the shock-absorbing component 103 is installed between the top of the base plate 101 and the bottom of the outer shell 2;

[0102] The mating blocks 102 are located on both sides of the base plate 101, making it easy for bolts to pass through and fix them in the corresponding positions.

[0103] Compared with related technologies, the second pulse time synchronization method for fiber optic strapdown inertial navigation and GNSS integrated navigation systems provided by this invention has the following advantages:

[0104] To improve the integration of the fiber optic strapdown inertial navigation system and enhance its environmental adaptability and overall performance, the core sensitive components of the system are installed inside the core mounting frame 8, while the signal processing and circuitry are installed inside the ordinary mounting frame 17. The core mounting frame 8 is made of zero-expansion microcrystalline glass, while the ordinary mounting frame 17 is made of ceramic. This combination reduces interference between the core components and the signal processing circuitry. A titanium alloy sheet 16 separates the ordinary mounting frame 17 from the core mounting frame 8, increasing stability and independence. The outer shell 2 is made of silver-plated aluminum, ensuring strength while increasing anti-interference performance. All components inside the ordinary mounting frame 17 and the core mounting frame 8 are connected via module interfaces 18 and module connectors 11 on the circuit board 12, fulfilling the requirement for unified management of multi-level installations. This design improves the integration of the fiber optic strapdown inertial navigation system, reduces interference between internal components during operation, and facilitates later management and maintenance through modular layout.

[0105] The above description is merely an embodiment of the present invention and does not limit the patent scope of the present invention. Any equivalent structural or procedural transformations made based on the content of the present invention specification and drawings, or direct or indirect applications in other related technical fields, are similarly included within the patent protection scope of the present invention.

Claims

1. A method for second pulse time transfer in a fiber-optic strapdown inertial navigation, GNSS integrated navigation system, characterized in that, The method comprises the following steps: S1: signal monitoring and timing initialization: FPGA relies on 100MHz internal working clock, starts PPS signal monitoring, scans signals by using double-edge trigger detection circuit, filters burr by combining digital filter window, detects PPS signal period jitter and duty cycle, if period jitter is more than ±10% or duty cycle is abnormal, automatically switches to internal crystal oscillator time keeping mode, records abnormal mark, maintains timing according to TCXO reference, time keeping precision is less than 1ms / min, and waits for PPS recovery; If normal, it is judged that PPS is valid jump, and time reference layer operation is entered; If valid jump is not detected, continuous monitoring is carried out; S2: time reference construction and timing start: taking TCXO as basic clock source, generating 100MHz working clock through internal PLL frequency multiplication of FPGA, as time reference, when PPS valid jump is captured, FPGA immediately clears internal microsecond counter, and then starts Time1 cumulative timing, and enters data interaction preparation stage; S3: inertial navigation data storage and period control: storage period setting: according to fiber-optic SINS calculation frequency such as 200Hz, 5ms / time, planning SINS data storage time length window, clearly storing for how long, after real-time storage task is started in response to FPGA hardware interrupt, real-time marking "time stamp" is given to stored SINS data, Time1 timing value of FPGA is associated, and time length of each data packet after PPS trigger is recorded, when combined navigation fixed period arrives, stopping current round SINS data storage, storing SINS acceleration and speed data sequence in period, and clearly storing what and when to stop; S4: cross-clock domain synchronous trigger: after FPGA detects PPS valid jump and starts Time1 timing, immediately sending a trigger signal to DSP through hardware interrupt line, in response to hardware interrupt, starting real-time storage task according to rules of SINS data storage and period control, continuously storing acceleration and speed data of inertial navigation calculation according to SINS calculation period, recording as SINS data sequence, i=1, 2, 3…n, and entering data temporary storage stage; S5: timing value transmission and data preprocessing: when fiber-optic SINS and GNSS combined navigation fixed period arrives, DSP actively reads Time1 timing value from FPGA, after Time1 is received, stopping real-time storage of SINS acceleration and speed data according to rules of SINS data storage and period control, and simultaneously taking the latest GNSS data packet from a communication interface, starting CRC-32 check, guaranteeing data integrity, and abnormality coping and robustness guarantee- data interaction abnormality processing: if CRC check of GNSS_data0 fails, enabling historical data prediction queue, using linear regression to complete current time mark information, marking data abnormality, and then entering subsequent process; if check is passed, the millisecond-level fragment time Time2 of the time mark is extracted normally; S6: Time delay compensation and fusion calculation closed loop: combine Time1 obtained from FPGA and Time2 of GNSS_data0, substitute into time delay model to calculate AT; determine AT validity, if AT >= 1s, trigger abnormality response, otherwise continue, based on AT and stored inertial navigation data, use extrapolation model with angular velocity compensation term to extrapolate GNSS raw speed and position to current time point of inertial navigation, align GNSS data time with inertial navigation data, construct Kalman filter measurement vector based on extrapolated GNSS data, combine inertial navigation state equation, execute Kalman filter time update and measurement update, correct inertial navigation system error, output fused integrated navigation information, complete "time delay compensation-fusion calculation" closed loop; S7: Abnormality response and robustness guarantee: embed signal monitoring response PPS signal abnormality, embed data preprocessing response data interaction abnormality, embed fusion calculation response calculation abnormality.

2. The method of claim 1, wherein the method further comprises: The PPS signal abnormality response: when the signal monitoring layer detects that the PPS signal cycle jitter exceeds ±10% or the duty cycle is abnormal, automatically switch to the internal crystal oscillator timekeeping mode, maintain timing according to the TCXO reference, and wait for the PPS to recover; the timekeeping accuracy is <1ms / min, which is better than the traditional 5ms / min, avoids the destruction of the time reference due to PPS abnormality, and guarantees the basic time synchronization capability of the system in poor GNSS signal conditions; the data interaction abnormality response: when the CRC check of GNSS_data0 fails, enable the historical data prediction queue, complete the current time information using linear regression, mark the data as abnormal and enter the subsequent process, intercept transmission errors, data packet damage, etc., and use historical data to "bottom out", avoid interrupting integrated navigation due to single packet data abnormality, and guarantee "continuous calculation, less error" of the system in harsh communication environment; the calculation abnormality response: when the time delay AT > 300ms, automatically increase the measurement noise variance matrix Rk of the Kalman filter, reduce the weight of abnormal observation, avoid abnormal observation from causing filter divergence, and guarantee the stability of the system during long-time operation.

3. The method of claim 1, wherein the method further comprises: The time delay compensation and fusion calculation closed loop includes time delay calculation and validity determination, GNSS data dynamic extrapolation fitting, and Kalman filter fusion calculation.

4. The method of claim 3, wherein the method further comprises: The time delay calculation and validity determination includes multi-source time information integration and validity double-layer determination.

5. The method of claim 3, wherein, The GNSS data dynamic extrapolation fitting includes extrapolation model disassembly and position extrapolation simplification logic, and the Kalman filter fusion calculation includes state equation and measurement equation review, filter process dynamic adjustment, and error feedback and correction.

6. The method of claim 1, wherein, The fiber-optic strapdown inertial navigation system comprises a shock-absorbing chassis; A shell is fixedly connected to the top of the shock-absorbing chassis, a partition plate is installed at the middle position in the interior of the shell, a plurality of docking strips are installed on the two sides of the inner wall of the shell, a core mounting frame is installed at the bottom position in the interior of the shell through a docking interface, a general mounting frame is installed at the position close to the top of the interior of the shell through a docking interface, module interfaces are arranged on the front surfaces of the core mounting frame and the general mounting frame, and heat-conducting cover plates are installed on the top of the core mounting frame and the general mounting frame.

7. The method of claim 6, wherein the method further comprises: The front surface of the shell is provided with a front cover, the front surface of the front cover is provided with an external panel, the back surface of the shell is provided with a rear cover, the front surface of the rear cover is provided with a circuit board with a flow port, and the front surface of the circuit board is provided with a plurality of module plugs.

8. The method of claim 7, wherein the method further comprises: The back surface of the rear cover is provided, near the top, with a heat dissipation component with a protective plate, and the back surface of the rear cover is provided, near the middle, with a filter component.

9. The method of claim 6, wherein the method further comprises: The heat-conducting cover plate comprises a top cover and a heat-conducting base sheet, the back surfaces of the core mounting frame and the general mounting frame are provided with mounting openings, and the monitoring components are mounted in the mounting openings.

10. The method of claim 6, wherein the method further comprises: The shock-absorbing chassis comprises a bottom plate, a butt joint block and a shock-absorbing component, and the shock-absorbing component is mounted between the top of the bottom plate and the bottom of the shell.

Images