CN114459455B - LSTM-based fiber-optic gyroscope scale factor error compensation method - Google Patents
LSTM-based fiber-optic gyroscope scale factor error compensation method Download PDFInfo
- Publication number
- CN114459455B CN114459455B CN202111598526.0A CN202111598526A CN114459455B CN 114459455 B CN114459455 B CN 114459455B CN 202111598526 A CN202111598526 A CN 202111598526A CN 114459455 B CN114459455 B CN 114459455B
- Authority
- CN
- China
- Prior art keywords
- fiber
- optic gyroscope
- scale factor
- temperature
- lstm
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01C—MEASURING DISTANCES, LEVELS OR BEARINGS; SURVEYING; NAVIGATION; GYROSCOPIC INSTRUMENTS; PHOTOGRAMMETRY OR VIDEOGRAMMETRY
- G01C25/00—Manufacturing, calibrating, cleaning, or repairing instruments or devices referred to in the other groups of this subclass
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01C—MEASURING DISTANCES, LEVELS OR BEARINGS; SURVEYING; NAVIGATION; GYROSCOPIC INSTRUMENTS; PHOTOGRAMMETRY OR VIDEOGRAMMETRY
- G01C19/00—Gyroscopes; Turn-sensitive devices using vibrating masses; Turn-sensitive devices without moving masses; Measuring angular rate using gyroscopic effects
- G01C19/58—Turn-sensitive devices without moving masses
- G01C19/64—Gyrometers using the Sagnac effect, i.e. rotation-induced shifts between counter-rotating electromagnetic beams
- G01C19/72—Gyrometers using the Sagnac effect, i.e. rotation-induced shifts between counter-rotating electromagnetic beams with counter-rotating light beams in a passive ring, e.g. fibre laser gyrometers
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/044—Recurrent networks, e.g. Hopfield networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
Abstract
The invention discloses a fiber-optic gyroscope scale factor error compensation method based on LSTM. The invention considers the coupling influence of the input angular rate and the temperature on the scale factor, designs the calibration experiment under the continuously changed temperature and the input angular rate, obtains the scale factor nonlinear error model of the fiber-optic gyroscope under different temperature points and input angular rate points, and compensates the fiber-optic gyroscope scale factor nonlinear error by using the LSTM. The experimental result shows that compared with the traditional discrete calibration scheme, the calibration time is greatly shortened, the calibration precision is improved, and the performance of the fiber-optic gyroscope at different temperatures and input angular rates is improved.
Description
Technical Field
The invention relates to the technical field of sensor testing, in particular to an LSTM-based fiber-optic gyroscope scale factor error compensation method.
Technical Field
The fiber optic gyroscope is an angular velocity sensor based on the Sagnac effect, has the advantages of high reliability, short starting time, large dynamic range, full solid state, low cost and the like, and is widely applied to the fields of airplanes, missiles, automobiles, robots and the like. In most cases, the fiber optic gyroscope scale factor is considered to be a constant value. However, in the actual use process, the scale factor of the fiber-optic gyroscope is easily influenced by many factors, and the scale factor is different from the numerical value in the field calibration, so that the performance in the actual use process is influenced. At present, it is known that the scale factor of the fiber optic gyroscope has a difference in the input angular rate, and especially, in the case of a small input angular rate, a measurement output has a significant error, and in addition, since the optoelectronic device inside the fiber optic gyroscope is easily affected by temperature, the scale factor also has a significant difference in different temperatures.
Conventional scale factor calibration methods calibrate the temperature and input angular rate, respectively. The method comprises the steps of selecting different temperature points, measuring the output of the fiber-optic gyroscope under the condition of different input angular rates, carrying out segmentation method or polynomial fitting to obtain an error model of the scale factor of the fiber-optic gyroscope, and realizing the compensation of the scale factor without considering the coupling influence of the angular rate and the temperature on the scale factor. Moreover, discrete calibration schemes require longer temperature hold at each temperature point, longer calibration times and less effective temperature information.
Disclosure of Invention
Aiming at the coupling influence of angular rate and temperature on the scale factor, calibration experiments under continuous temperature points and rate points are designed, only one group of experiments are needed to obtain the output data of the fiber optic gyroscope under the continuous temperature points and rate points, the calibration time is greatly shortened, and the calibration precision is improved. The technical problem to be solved by the invention is realized by the following technical scheme:
a fiber-optic gyroscope scale factor error compensation method based on LSTM comprises the following steps:
1) The optical fiber gyroscope is arranged on a rotary table in the incubator, so that a sensitive axis of the optical fiber gyroscope is parallel to a rotary shaft of the rotary table;
2) Setting the initial temperature of the incubator, electrifying the fiber-optic gyroscope and stabilizing the incubator for a period of time, then controlling the incubator to heat up or cool down at a certain variable temperature rate, and simultaneously controlling the turntable to rotate at a variable speed to provide a variable input angular rate for the fiber-optic gyroscope; the range of temperature rise or temperature drop is the full temperature range of the normal work of the fiber-optic gyroscope;
3) Acquiring the output F of the fiber-optic gyroscope, corresponding real-time temperature T and input angular rate omega, and calculating scale factor values K (T, omega) at different input angular rates and temperatures:
in the formula, F 0 (T) is a zero-bias model of the temperature function, and K (is) is a fiber optic gyroscope scale factor;
after the experiment is finished, obtaining a group of sample data { F, T, omega, K (T, omega) };
4) Establishing an LSTM neural network, taking { F, T } in sample data as an input sequence of the LSTM neural network, taking { omega, K (T, omega) } in the sample data as an output sequence, training the LSTM neural network, and taking the trained LSTM neural network as a scale factor model of the fiber-optic gyroscope;
5) In the actual working process of the fiber-optic gyroscope, the real-time temperature T and the fiber-optic gyroscope output F in the working process are collected and used as the input of a scale factor model of the fiber-optic gyroscope, and the scale factors of the fiber-optic gyroscope under different input angular rates omega are obtained.
Further, the temperature range tested in the step 2) is 60 ℃ to-40 ℃, the initial temperature is 60 ℃, and the temperature change rate is-1 ℃/min.
Further, the change rule of the angular rate of the turntable input is as follows:
the angular velocity of the turntable input is a/s from the angular velocity of reversal-M/s 2 The angular acceleration is changed to the forward rotation angular speed M DEG/s, and the speed is kept between 1 and 20s; then at-a/s 2 The angular acceleration of the rotor is changed to a reversal angular rate-M DEG/s, and the speed is kept between 1 and 20s; this process is repeated until the temperature is raised or lowered from the initial temperature to the final temperature.
Further, the angular acceleration is 10 °/s 2 Or-10 °/s 2 。
Further, the input angular rate ranges from-300 °/s to 300 °/s.
Furthermore, the LSTM neural network adopts a double-layer LSTM structure.
Compared with the prior art, the invention has the advantages that: the invention provides an LSTM-based fiber-optic gyroscope scale factor error compensation method, which researches the coupling influence of angular rate and temperature on scale factors, designs calibration experiments under continuous temperature points and rate points, and can obtain fiber-optic gyroscope output data under the continuous temperature points and the rate points only by carrying out a group of experiments. The experimental result shows that compared with the traditional discrete calibration scheme, the calibration time is greatly shortened, the calibration precision is improved, and the performance of the fiber-optic gyroscope at different temperatures and input angular rates is improved.
Drawings
FIG. 1 is a flow chart of the scale factor error compensation of the present embodiment;
FIG. 2 is a graph of the temperature and input angular rate for the calibration scheme of the present embodiment;
FIG. 3 is a plot of scale factor error versus temperature and input angular rate for the present embodiment;
FIG. 4 shows the result of the error compensation of the scale factor in the present embodiment;
Detailed Description
The technical solution of the present invention is further described in detail below with reference to the accompanying drawings.
As shown in fig. 1, the invention designs calibration experiments under continuous temperature points and rate points, and only needs to perform a group of experiments to obtain the output data of the fiber-optic gyroscope under the continuous temperature points and rate points, which specifically comprises the following steps:
(1) The optical fiber gyroscope is arranged on a rotary table in the incubator, so that a sensitive shaft of the optical fiber gyroscope is parallel to a rotary shaft of the rotary table;
(2) The fiber-optic gyroscope scale factor calibration experiment under continuous temperature points and rate points is carried out, and the method comprises the following specific steps:
(2-1) keeping the initial temperature of the incubator at 60 ℃, and electrifying the gyroscope;
and (2-2) after the gyroscope is electrified for half an hour, the interior of the gyroscope is fully insulated. Then, the temperature of the incubator is controlled to be reduced to-40 ℃ at a specified temperature change rate, in the embodiment, the temperature change rate is set to-1 ℃/min, and the temperature change curve is shown as a solid line in fig. 2; at the same time, the turret provides a varying input angular rate, the input angular rate curve being shown in dashed lines in fig. 2; the gyroscope operates at a dynamically changing temperature and input angular velocity, and its output is denoted as F.
In this embodiment, the change rule of the turntable input angular rate is as follows:
(2-2-1) turntable input angular velocity from reverse angular velocity 300 °/s at 10 °/s 2 The angular acceleration of the rotor is changed to the forward rotation angular rate of 300 DEG/s;
(2-2-2) the turn table is maintained at the forward rotation angular rate of 300 °/s for 10 seconds;
(2-2-3) turntable input angular rate from the forward rotation angular rate of 300 DEG/s to-10 DEG/s 2 To a reversal angular rate of 300 °/s;
(2-2-4) the turntable was kept at a reversal angle rate of 300 °/s for 10 seconds;
(2-2-5) repeating the steps until the temperature is reduced to-40 ℃.
(3) Synchronously acquiring the temperature T of the fiber-optic gyroscope, the output F of the fiber-optic gyroscope and the input angular rate omega of the turntable, and calculating scale factor values K (T, omega) at different input angular rates and temperatures;
the formula for calculating the scale factor of the fiber-optic gyroscope is as follows:
in the formula, F 0 (T) is a zero-bias model of the temperature function, and K (is) is a fiber optic gyroscope scale factor.
The scale factor values at different input angular rates and temperatures are shown in fig. 3, and it can be clearly seen that the relationship between the fiber optic gyroscope scale factor and the angular rate shows a remarkable hyperbolic trend, and when the angular rate input is small, the scale factor has a remarkable error and increases with the increase of the temperature.
(4) Training an LSTM neural network by using scale factor values at different input angular rates and temperatures to determine a hyper-parameter, wherein in the training process, the input of the LSTM neural network is the temperature T of a fiber optic gyroscope and the output F of the fiber optic gyroscope, and the output of the LSTM neural network is the scale factor of the fiber optic gyroscope at different input angular rates omega; and obtaining a scale factor model of the fiber-optic gyroscope after the training is finished.
The LSTM neural network consists of a forgetting gate, an input gate and an output gate, and the specific calculation process is as follows:
f t ＝σ g (W f x t +U f h t-1 +b f )
i t ＝σ g (W i x t +U i h t-1 +b i )
o t ＝σ g (W o x t +U o h t-1 +b o )
in the formula, x t The method comprises the following steps of inputting a sequence, namely a fiber optic gyroscope temperature T and a fiber optic gyroscope output F; h is t Is an output sequence, namely the scale factors of the fiber-optic gyroscope under different input angular rates omega; f. of t To forget the door; i.e. i t And
forming an input gate; o. o t Is an output gate; c. C t Is a memory cell state; w and U are weight matrixes; b is a deviation vector; sigma g Activating a function for Sigmod; sigma h Is a hyperbolic tangent function.
In this embodiment, the relevant parameters of the LSTM neural network are: two layers of LSTM structures, 1 fully connected layer; the two layers of LSTM respectively have 50 LSTMCells, the initial learning rate is 0.03, and the optimizer adopts an adaptive moment estimation optimizer Adam which can dynamically adjust the learning rate according to the gradient in the training process; the Dropout layer is added on the full-connection layer, and zero weight is randomly assigned to the neurons in the network according to a set ratio value, so that overfitting is avoided.
(5) The model obtained in the step 4 is used for compensating the scale factor of the fiber-optic gyroscope, the compensation effect is shown in fig. 4, the error of the scale factor of the fiber-optic gyroscope is reduced from 280ppm to 13ppm, and the compensation effect is obvious.
Therefore, the scope of the present invention should not be limited by the disclosure of the embodiments, and the actual scope of the protection should be determined by the claims.
Claims (6)
1. A fiber-optic gyroscope scale factor error compensation method based on LSTM is characterized by comprising the following steps:
1) The optical fiber gyroscope is arranged on a rotary table in the incubator, so that a sensitive axis of the optical fiber gyroscope is parallel to a rotary shaft of the rotary table;
2) Setting the initial temperature of the incubator, electrifying the fiber-optic gyroscope and stabilizing for a period of time, then controlling the incubator to heat up or cool down at a certain variable temperature rate, and simultaneously controlling the rotary table to rotate at a variable speed to provide a variable input angular rate for the fiber-optic gyroscope; the range of temperature rise or temperature drop is the full temperature range of the normal work of the fiber-optic gyroscope;
3) Acquiring the output F of the fiber-optic gyroscope, corresponding real-time temperature T and input angular rate omega, and calculating scale factor values K (T, omega) at different input angular rates and temperatures:
in the formula, F 0 (T) is a zero-bias model of the temperature function, and K (.) is a fiber optic gyroscope scale factor;
after the experiment is finished, obtaining a group of sample data { F, T, omega, K (T, omega) };
4) Establishing an LSTM neural network, taking { F, T } in sample data as an input sequence of the LSTM neural network, taking { omega, K (T, omega) } in the sample data as an output sequence, training the LSTM neural network, and taking the trained LSTM neural network as a scale factor model of the fiber-optic gyroscope;
5) In the actual working process of the fiber-optic gyroscope, the real-time temperature T and the fiber-optic gyroscope output F in the working process are collected and used as the input of a scale factor model of the fiber-optic gyroscope, and the scale factors of the fiber-optic gyroscope under different input angular rates omega are obtained.
2. The LSTM-based fiber-optic gyroscope scale factor error compensation method of claim 1, wherein the temperature range tested in step 2) is 60 ℃ to-40 ℃, the initial temperature is 60 ℃, and the ramp rate is-1 ℃/min.
3. The LSTM-based fiber optic gyroscope scale factor error compensation method of claim 1, wherein the turntable input angular rate change rule is as follows:
the angular velocity of the turntable input is a/s from the angular velocity of reversal-M/s 2 The angular acceleration is changed to the forward rotation angular speed of M DEG/s, and the angular acceleration is kept for 1-20s; then at-a/s 2 The angular acceleration of the rotor is changed to a reversal angular rate-M DEG/s, and the speed is kept between 1 and 20s; this process is repeated until the temperature is raised or lowered from the initial temperature to the final temperature.
4. The LSTM-based fiber optic gyroscope scale factor error compensation method of claim 3, wherein the angular acceleration is 10 °/s 2 Or-10 °/s 2 。
5. The LSTM-based fiber optic gyroscope scale factor error compensation method of claim 3, wherein the input angular rate ranges from-300 °/s to 300 °/s.
6. The LSTM-based fiber optic gyroscope scale factor error compensation method of claim 1 where the LSTM neural network employs a two-layer LSTM architecture.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111598526.0A CN114459455B (en) | 2021-12-24 | 2021-12-24 | LSTM-based fiber-optic gyroscope scale factor error compensation method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111598526.0A CN114459455B (en) | 2021-12-24 | 2021-12-24 | LSTM-based fiber-optic gyroscope scale factor error compensation method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN114459455A CN114459455A (en) | 2022-05-10 |
| CN114459455B true CN114459455B (en) | 2023-02-14 |
Family
ID=81406832
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111598526.0A Active CN114459455B (en) | 2021-12-24 | 2021-12-24 | LSTM-based fiber-optic gyroscope scale factor error compensation method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114459455B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115014399A (en) * | 2022-08-04 | 2022-09-06 | 中国船舶重工集团公司第七0七研究所 | Scale factor error analysis method of fiber-optic gyroscope |
| CN115855016B (en) * | 2023-02-27 | 2023-06-16 | 南开大学 | Low-temperature impact error compensation method for optical fiber gyroscope |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102519489A (en) * | 2011-12-16 | 2012-06-27 | 东南大学 | Fiber optic gyro (FOG) scale factor modelling method based on temperatures and input angular rates |
| CN104713574A (en) * | 2013-12-11 | 2015-06-17 | 中国航空工业第六一八研究所 | Closed loop fiber optic gyroscope scale factor high precision calibrating method |
| CN111238462A (en) * | 2020-01-19 | 2020-06-05 | 湖北三江航天红峰控制有限公司 | LSTM fiber-optic gyroscope temperature compensation modeling method based on deep embedded clustering |
| CN112629560A (en) * | 2020-12-14 | 2021-04-09 | 北京航天时代光电科技有限公司 | Automatic calibration method for multiple temperature parameters of high-precision fiber-optic gyroscope |
| CN112729347A (en) * | 2021-01-19 | 2021-04-30 | 湖北三江航天万峰科技发展有限公司 | Temperature compensation method and device for fiber-optic gyroscope, electronic equipment and storage medium |
| CN113720357A (en) * | 2021-09-16 | 2021-11-30 | 北京控制工程研究所 | Gyro scale factor calibration and compensation method under vacuum full-temperature condition of 3S optical fiber IMU |
-
2021
- 2021-12-24 CN CN202111598526.0A patent/CN114459455B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102519489A (en) * | 2011-12-16 | 2012-06-27 | 东南大学 | Fiber optic gyro (FOG) scale factor modelling method based on temperatures and input angular rates |
| CN104713574A (en) * | 2013-12-11 | 2015-06-17 | 中国航空工业第六一八研究所 | Closed loop fiber optic gyroscope scale factor high precision calibrating method |
| CN111238462A (en) * | 2020-01-19 | 2020-06-05 | 湖北三江航天红峰控制有限公司 | LSTM fiber-optic gyroscope temperature compensation modeling method based on deep embedded clustering |
| CN112629560A (en) * | 2020-12-14 | 2021-04-09 | 北京航天时代光电科技有限公司 | Automatic calibration method for multiple temperature parameters of high-precision fiber-optic gyroscope |
| CN112729347A (en) * | 2021-01-19 | 2021-04-30 | 湖北三江航天万峰科技发展有限公司 | Temperature compensation method and device for fiber-optic gyroscope, electronic equipment and storage medium |
| CN113720357A (en) * | 2021-09-16 | 2021-11-30 | 北京控制工程研究所 | Gyro scale factor calibration and compensation method under vacuum full-temperature condition of 3S optical fiber IMU |
Non-Patent Citations (4)
| Title |
|---|
| 《光纤陀螺测量系统标度因数标定方法研究》;马知瑶，周一览;《电子技术与软件工程》;20191201(第23期);第73-74页 * |
| 光纤陀螺温度与标度因数非线性建模与补偿;王新龙，马闪;《北京航空航天大学学报》;20090115;第35卷(第1期);第28-31页 * |
| 光纤陀螺温度漂移自适应网络模糊推理补偿;赵曦晶，刘光斌，汪立新，何志昆，赵晗;《红外与激光工程》;20140325;第43卷(第3期);第790-793页 * |
| 基于高斯过程回归的光纤陀螺温度漂移补偿;何志昆，刘光斌，赵曦晶，刘冬;《高技术通讯》;20130615;第23卷(第6期);第642-647页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN114459455A (en) | 2022-05-10 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN114459455B (en) | LSTM-based fiber-optic gyroscope scale factor error compensation method | |
| CN108710303B (en) | Spacecraft relative attitude control method containing multi-source disturbance and actuator saturation | |
| CN110687787B (en) | Self-adaptive control method for mechanical arm system | |
| CN105607473B (en) | The attitude error Fast Convergent self-adaptation control method of small-sized depopulated helicopter | |
| CN112668104B (en) | Online identification method for pneumatic parameters of hypersonic aircraft | |
| CN112945225A (en) | Attitude calculation system and method based on extended Kalman filtering | |
| CN111716360B (en) | Fuzzy logic-based flexible joint mechanical arm sampling control method and device | |
| CN112077839B (en) | Motion control method and device for mechanical arm | |
| CN114061559B (en) | Compensation method, system and computer storage medium for zero offset drift of fiber optic gyroscope | |
| CN111273544B (en) | Radar pitching motion control method based on prediction RBF feedforward compensation type fuzzy PID | |
| CN103344257A (en) | Quick temperature calibrating method of inertia measuring unit | |
| CN112558468B (en) | Launching platform adaptive robust output feedback control method based on double observers | |
| CN113591240B (en) | Modeling method for thermal error model of tooth grinding machine based on bidirectional LSTM network | |
| CN110083062A (en) | A kind of optic central extract composite control method based on velocity disturbance observer and Fuzzy-PID | |
| CN112181002B (en) | Micro gyroscope dual-recursion disturbance fuzzy neural network fractional order sliding mode control method | |
| CN114310851B (en) | Dragging teaching method of robot moment-free sensor | |
| CN116105724A (en) | Full-temperature calibration method and device for strapdown inertial navigation system | |
| CN111427267B (en) | High-speed aircraft attack angle tracking method adopting force and moment adaptive estimation | |
| CN114815618B (en) | Adaptive neural network tracking control method based on dynamic gain | |
| CN110595434B (en) | Quaternion fusion attitude estimation method based on MEMS sensor | |
| CN114310911B (en) | Driving joint dynamic error prediction and compensation system and method based on neural network | |
| CN114740723A (en) | Software robot robust self-adaptive control method based on disturbance observer | |
| Song et al. | Adaptive dynamic programming: single and multiple controllers | |
| Liang et al. | Attitude control of quadrotor UAV based on LADRC method | |
| Liu et al. | Fuzzy PID tracking controller for two-axis airborne optoelectronic stabilized platform |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |