CN104950135A - LM_BP-algorithm-based temperature compensation method and system of micro-silicon accelerometer - Google Patents
LM_BP-algorithm-based temperature compensation method and system of micro-silicon accelerometer Download PDFInfo
- Publication number
- CN104950135A CN104950135A CN201510138084.XA CN201510138084A CN104950135A CN 104950135 A CN104950135 A CN 104950135A CN 201510138084 A CN201510138084 A CN 201510138084A CN 104950135 A CN104950135 A CN 104950135A
- Authority
- CN
- China
- Prior art keywords
- temperature
- accelerometer
- temperature compensation
- silicon micro
- algorithm
- 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.)
- Pending
Links
Abstract
The invention discloses an LM_BP-algorithm-based temperature compensation method and system of a micro-silicon accelerometer. The system comprises an acceleration sensor, a data acquisition unit, a microprocessor unit, a data interface unit, a voltage stabilizing unit, and a reference source circuit. The data acquisition unit obtains acceleration output data and temperature data of the micro-silicon accelerometer, wherein the data are used as training samples of an LM_BP neural network; a temperature compensation model of the accelerometer is established based on a self learning way; model parameters are stored into a storage device of the microprocessor of the compensation system; and the model is operated by invoking the model parameters by using an MCU, thereby obtaining the real-time output data of the accelerometer after compensation. According to the temperature compensation method, the self-adaptive and learning capabilities are high; the measurement precision of the micro-silicon accelerometer is effectively improved; and the anti-interference capability is high and the interface becomes convenient.
Description
Technical field
The present invention relates to a kind of compensation technique of silicon micro-acceleration sensor, be specifically related to a kind of silicon micro accerometer temperature compensation based on LM_BP algorithm and system.
Background technology
Silicon micro accerometer is one of key element of inertia measurement and navigational system, has important application in fields such as space flight and aviation, inertial navigation and automotive safeties.Therefore, the precision of silicon micro accerometer also just seems particularly important, and environment temperature factor is one of principal element affecting silicon micro accerometer precision.This is because silicon micro accerometer itself causes the susceptibility of temperature and residing temperature profile effect.When the environmental temperature is changed, the mathematical model of silicon micro accerometer also will change.Therefore, carrying out temperature compensation to silicon micro accerometer is the important means improving silicon micro accerometer precision.
At present, conventional compensation method has following several.Wherein, hardware compensating method mainly contains: 1. take into full account thermal design when Micro-Accelerometer Design, deals carefully with the heat interference occurred in design, reduces the impact on silicon micro accerometer as far as possible; 2. in silicon micro accerometer, increase the design of temperature compensation structure, adopt the material of negative temperature coefficient or element to offset the change of the Material Physics parameter caused by temperature variation, thus reach the object of temperature compensation; 3. by increasing hardware measure, improving the environment temperature of accelerometer test and work, taking necessary temperature control measures in a test system.It is high that above method mostly realizes cost, and process is complicated.So more employing software compensation schemes in engineering, the method for software compensation is then that main method has based on polynomial surface fitting, vector machine, wavelet network etc. to obtain premised on silicon microaccelerometer temperature model.
In prior art, as Chinese invention patent CN 102323448A discloses the self-compensating linear accelerometer of a kind of zero-bit, by increasing capacitance detecting pole plate and compensating circuit, the change of the accelerometer zero size that auto-compensation is passed in time or temperature variation causes on the basis of source accelerometer.The method is compensated the zero-bit of accelerometer separately by hardware circuit, but circuit is complicated, and can not compensate the nonlinearity of scaling factor simultaneously.
Chinese invention patent CN 103558415A discloses a kind of mems accelerometer with temperature compensation, passes through
zero is measured partially and one group of output valve of scaling factor fit to fitting surface and obtain a series of fitting coefficient and be arranged in matrix of coefficients under multiple temperature spot;
the output signal of degree of will speed up sensor and the output signal modeling of temperature sensor are also expressed as model matrix;
matrix of coefficients and model matrix are done dot product and obtain the output formula of acceleration transducer after temperature compensation.Namely by compensating after the three-dimensional matching of curved surface and calculating, this method effectively can reduce accelerometer in the error of assembling and produce in installation process on the impact of system, but calculated amount is large, and matching is complicated, does not have learning ability.
In sum, existing patent, to the compensation of silicon micro accerometer temperature, all has castering action to the precision of silicon micro accerometer, but or there is circuit complexity, compensation performance is single, or backoff algorithm does not possess the problem of learning ability.
Summary of the invention
Goal of the invention of the present invention is to provide a kind of silicon micro accerometer temperature compensation based on LM_BP algorithm and system, there is good adaptivity, self-organization and very strong learning ability, effectively can improve the measuring accuracy of silicon micro accerometer, and system has antijamming capability by force, interface is feature easily.
To achieve the above object of the invention, the technical solution used in the present invention is: a kind of silicon micro accerometer temperature compensation based on LM_BP algorithm, comprises the steps:
(1) accelerometer of temperature sensor is selected silicon micro-acceleration to count to be integrated with, export and obtain multi-group data by being with circulate under the multiple temperature spot acceleration output of the accelerometer described in measuring and temperature of the dimensional turntable of temperature control, as the training sample of BP neural network;
(2) according to gained training sample, the model of temperature compensation of accelerometer is set up by self study
,
Wherein,
for the connection weights between input layer to hidden layer neuron,
for the connection weights between hidden layer neuron to output layer neuron,
for the threshold value of hidden neuron,
for the threshold value of output neuron;
(3) the model of temperature compensation parameter that step (2) obtains is left in the storer of microprocessor, coding, set up BP neural network model by call parameters, realize backoff algorithm and export in real time.
Preferred technical scheme, described BP neural network is three layers, hidden layer in the middle of comprising.
Preferred technical scheme, described middle hidden layer is provided with 15 neurons.
A kind of silicon micro accerometer temperature-compensated system based on LM_BP algorithm, for realizing the temperature compensation of claim 1, comprise acceleration transducer, data acquisition unit, microprocessor unit, digital interface unit, source of stable pressure circuit and reference source circuit, described acceleration transducer, data acquisition unit, microprocessor unit are connected successively with digital interface unit, and described digital interface unit adopts RS-422 digital interface.
Because technique scheme is used, the present invention compared with prior art has following advantages:
1. the present invention adopts LM_BP neural network algorithm, the temperature impact that the performance parameter improving accelerometer is subject to, the important performance characteristic such as the temperature coefficient of silicon micro accerometer constant multiplier, constant multiplier and bias instaility can be improved simultaneously, the measuring accuracy of further raising accelerometer, and there is adaptive learning ability due to LM_BP neural network algorithm, therefore practical.
2. the present invention adopts RS-422 digital interface transmission plan, the problem of analog sensor antijamming capability weakness can be solved, make it that there is structure simple, transmit reliable and stable, antijamming capability is strong, the feature that cost is low, can be widely used in severe environment for use and durability requires high field.
Accompanying drawing explanation
Fig. 1 is LM_BP neural network algorithm principle figure of the present invention in embodiment one.
Fig. 2 is microprocessor working procedure block diagram of the present invention in embodiment one.
Fig. 3 is the system chart of silicon micro accerometer compensation method of the present invention in embodiment one.
Fig. 4 is the power circuit diagram of silicon micro accerometer compensation method of the present invention in embodiment one.
Fig. 5 is the silicon micro accerometer circuit diagram of silicon micro accerometer compensation method of the present invention in embodiment one.
Fig. 6 is the circuit diagram of the data acquisition unit of silicon micro accerometer compensation method of the present invention in embodiment one.
Fig. 7 is the digital processing circuit figure based on ARM of silicon micro accerometer compensation method of the present invention in embodiment one.
Fig. 8 is the circuit diagram of the digital interface unit of silicon micro accerometer compensation method of the present invention in embodiment one.
Fig. 9 is the automatic test block diagram of silicon micro accerometer compensation method of the present invention in embodiment one.
Figure 10 is the data plot of silicon micro accerometer of the present invention lower nonlinearity of full temperature before and after compensating in embodiment one.
Figure 11 is the lower zero inclined data plot of full temperature before and after silicon micro accerometer of the present invention compensates.
Embodiment
Below in conjunction with drawings and Examples, the invention will be further described:
Embodiment one: shown in Figure 1, a kind of silicon micro accerometer temperature compensation based on LM_BP algorithm, comprises the steps:
(1) accelerometer of temperature sensor is selected silicon micro-acceleration to count to be integrated with, 10 temperature spots are evenly got at-30 DEG C ~ 60 DEG C by being with the dimensional turntable of temperature control, the acceleration of the silicon micro accerometer described in circulation measurement exports and temperature exports and obtains multi-group data, as the training sample of BP neural network;
(2) according to gained training sample, optimization BP neural network is three layers, and hidden neuron is 15.Wherein, the method applying LM Algorithm for Training amendment weights is:
In formula: J is the Jacobi matrix of network error to weights derivative, and e is error vector, and u is self-adaptative adjustment scalar, and present networks target error function is set to:
When setting network training error be less than preset error amount or e-learning number of times reach preset value time, by self-adaptation, obtain the connection weights between each layer neuron and hidden layer, output layer neuron threshold value:
Wherein,
be 15 × 2 matrixes,
be 1 × 15 matrix,
be 1 × 15 matrix,
be 1 × 1 matrix,
Through to train for 5000 times the connection weights that obtain between each layer neuron and neuron threshold value as follows:
(3) the model of temperature compensation parameter that step (2) obtains is left in the storer of microprocessor, coding, set up BP neural network model by call parameters, realize backoff algorithm and export in real time.
Shown in Figure 2, be compensation program flow process in processor.First program carries out initialization, and Microprocessor S3C44B0X AD carries out sampling and reading original acceleration signal
and temperature signal
, will
input LM_BP neural network model, train the neuron weights and network parameter that obtain above then calling, run BP network model as follows:
。
Calculate offset, finally carry out numeral and export, arranging output speed is 200Hz; Simultaneously, for convenience of the comparison of compensation front and back data and the accuracy of data transmission, design digital interface exports (3 bytes)+verification (1 byte) after sending output (3 bytes)+temperature (3 bytes) before data structure is frame head (2 bytes)+compensation+compensation.
Shown in Figure 3, a kind of silicon micro accerometer temperature-compensated system based on LM_BP algorithm, for realizing the temperature compensation of claim 1, comprise silicon micro-acceleration sensor, high-precision AD data acquisition unit, microprocessor unit, digital interface unit, source of stable pressure circuit and reference source circuit, adopt single power supply, described acceleration transducer, data acquisition unit, microprocessor unit are connected successively with digital interface unit, and described digital interface unit adopts RS-422 digital interface.
Native system gathers the acceleration of silicon micro accerometer and the output voltage values of temperature by Microprocessor S3C44B0X AD unit, and the data obtained are carried out LM_BP computing in the microprocessor, exports acceleration compensation value in real time finally by digital interface.
System of the present invention adopts the STM32F405RG of ST company as microprocessor chip, and the AD7190 of AD company is as data acquisition unit, and the MAX491 of MAXIN company is as the digital output interface chip of RS_422.Its component part is described below:
Shown in Figure 4, be the power circuit of system, comprise LT1962EMS8-5 and ADP3338 power supply chip and AD780 superfinishing accurate 2.5V reference voltage chip and ADR4550 5V reference source chip.Meanwhile, to connect a 1N4148 diode D1 at the power input of whole circuit, power protection is provided, and the magnetic bead L1 that connects thereafter, the AC ripple in filter out power.In whole system, core circuit is powered and to be provided by source of stable pressure and reference source, brings impact to reduce power supply noise to measuring accuracy.
Shown in Figure 5, be silicon micro accerometer circuit diagram, install and measure for convenience, the sensitive axes center line of silicon micro-acceleration sensor should be placed in the midline position of circuit board.
Shown in Figure 6, be the two-way voltage collection circuit of AD7190, adopt LT1962EMS8-5 to simulate power supply as 5V, adopt ADR4550 as 5V reference source, and adopt ADP338 to provide 3.3V digital power, thus without the need to external level conversion, simplify the interface of ADC and microcontroller.The electric capacity composition RC filtering circuit of 100 Ω resistance and 0.1uF is connected, with the suppression in enhanced modulation device sample frequency at each analog input end of AD.Each power end and reference source termination enter filter capacitor, remove the impact of ripple.
Shown in Figure 7, be the digital processing circuit based on ARM, adopt STM32F405 as process chip, as the center cell of whole device collection control, data processing, output in real time.
Shown in Figure 8, for system interface circuit, comprise the digital output interface chip MAX491 of RS_422, and for convenience of carrying out measuring and installing with aerial lug, the circuit board power supply provided in the entire system, ARM circuit are downloaded and RS_422 numeral output connection port.
See shown in Fig. 9 to 11, through test, the present invention all has the nonlinearity under the bias instaility of accelerometer constant multiplier temperature coefficient, full temperature and full temperature and significantly promotes effect.
Claims (4)
1., based on a silicon micro accerometer temperature compensation for LM_BP algorithm, it is characterized in that, comprise the steps:
(1) accelerometer of temperature sensor is selected silicon micro-acceleration to count to be integrated with, export and obtain multi-group data by being with circulate under the multiple temperature spot acceleration output of the accelerometer described in measuring and temperature of the dimensional turntable of temperature control, as the training sample of BP neural network;
(2) according to gained training sample, the model of temperature compensation of accelerometer is set up by self study;
(3) the model of temperature compensation parameter that step (2) obtains is left in the storer of microprocessor, coding, set up BP neural network model by call parameters, realize backoff algorithm and export in real time.
2. a kind of silicon micro accerometer temperature compensation based on LM_BP algorithm according to claim 1, is characterized in that: described BP neural network is three layers, hidden layer in the middle of comprising.
3. a kind of silicon micro accerometer temperature compensation based on LM_BP algorithm according to claim 2, is characterized in that: described middle hidden layer is provided with 15 neurons.
4. the silicon micro accerometer temperature-compensated system based on LM_BP algorithm, for realizing the temperature compensation of claim 1, comprise acceleration transducer, data acquisition unit, microprocessor unit, digital interface unit, source of stable pressure circuit and reference source circuit, it is characterized in that: described acceleration transducer, data acquisition unit, microprocessor unit are connected successively with digital interface unit, described digital interface unit adopts RS-422 digital interface.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510138084.XA CN104950135A (en) | 2015-03-26 | 2015-03-26 | LM_BP-algorithm-based temperature compensation method and system of micro-silicon accelerometer |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510138084.XA CN104950135A (en) | 2015-03-26 | 2015-03-26 | LM_BP-algorithm-based temperature compensation method and system of micro-silicon accelerometer |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN104950135A true CN104950135A (en) | 2015-09-30 |
Family
ID=54164944
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510138084.XA Pending CN104950135A (en) | 2015-03-26 | 2015-03-26 | LM_BP-algorithm-based temperature compensation method and system of micro-silicon accelerometer |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104950135A (en) |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108073075A (en) * | 2017-12-21 | 2018-05-25 | 苏州大学 | Silicon micro accerometer temperature-compensation method, system based on GA Optimized BP Neural Networks |
| CN108120451A (en) * | 2017-12-21 | 2018-06-05 | 苏州大学 | Based on silicon micro accerometer temperature-compensation method, the system for improving PSO optimization neural networks |
| CN109738669A (en) * | 2019-01-11 | 2019-05-10 | 北京麦斯泰克科技有限公司 | A kind of temperature drift compensation method of piezoelectric acceleration transducer |
| CN109816094A (en) * | 2019-01-03 | 2019-05-28 | 山东省科学院海洋仪器仪表研究所 | Optical dissolved oxygen sensor non-linear temperature compensation method based on neural network L-M algorithm |
| CN109855623A (en) * | 2019-01-09 | 2019-06-07 | 东南大学 | Geomagnetic model online approximating method based on Legendre multinomial and BP neural network |
| CN113514665A (en) * | 2021-05-28 | 2021-10-19 | 成都工业职业技术学院 | Acceleration monitoring system and method for spent fuel assembly transport container |
| CN115047213A (en) * | 2021-03-09 | 2022-09-13 | 北京大学 | Method for improving long-term stability of MEMS accelerometer |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1865852A (en) * | 2006-06-27 | 2006-11-22 | 北京航空航天大学 | Output error compensating process for open-loop optic fiber gyroscope based on neural network |
| WO2013023068A1 (en) * | 2011-08-11 | 2013-02-14 | Greenray Industries, Inc. | Neural network frequency control |
-
2015
- 2015-03-26 CN CN201510138084.XA patent/CN104950135A/en active Pending
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1865852A (en) * | 2006-06-27 | 2006-11-22 | 北京航空航天大学 | Output error compensating process for open-loop optic fiber gyroscope based on neural network |
| WO2013023068A1 (en) * | 2011-08-11 | 2013-02-14 | Greenray Industries, Inc. | Neural network frequency control |
Non-Patent Citations (5)
| Title |
|---|
| 任春华 等: ""基于神经网络的石英加速度计的二维时、温漂移补偿"", 《中国惯性技术学报》 * |
| 倪建丽 等: ""BP神经网络在加速度计误差补偿中的应用"", 《测控技术》 * |
| 吴晓主编: "《计算机工业控制技术》", 31 August 2005 * |
| 耿小庆 等: ""几种改进BP算法及其在应用中的比较分析"", 《计算机工程与应用》 * |
| 郭慧斌 等: ""基于BP神经网络的加速度计温度补偿方法研究"", 《电子世界》 * |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108073075A (en) * | 2017-12-21 | 2018-05-25 | 苏州大学 | Silicon micro accerometer temperature-compensation method, system based on GA Optimized BP Neural Networks |
| CN108120451A (en) * | 2017-12-21 | 2018-06-05 | 苏州大学 | Based on silicon micro accerometer temperature-compensation method, the system for improving PSO optimization neural networks |
| CN109816094A (en) * | 2019-01-03 | 2019-05-28 | 山东省科学院海洋仪器仪表研究所 | Optical dissolved oxygen sensor non-linear temperature compensation method based on neural network L-M algorithm |
| CN109855623A (en) * | 2019-01-09 | 2019-06-07 | 东南大学 | Geomagnetic model online approximating method based on Legendre multinomial and BP neural network |
| CN109738669A (en) * | 2019-01-11 | 2019-05-10 | 北京麦斯泰克科技有限公司 | A kind of temperature drift compensation method of piezoelectric acceleration transducer |
| CN109738669B (en) * | 2019-01-11 | 2020-12-25 | 北京麦斯泰克科技有限公司 | Temperature drift compensation method of piezoelectric type acceleration sensor |
| CN115047213A (en) * | 2021-03-09 | 2022-09-13 | 北京大学 | Method for improving long-term stability of MEMS accelerometer |
| CN113514665A (en) * | 2021-05-28 | 2021-10-19 | 成都工业职业技术学院 | Acceleration monitoring system and method for spent fuel assembly transport container |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104950135A (en) | LM_BP-algorithm-based temperature compensation method and system of micro-silicon accelerometer | |
| CN108120451A (en) | Based on silicon micro accerometer temperature-compensation method, the system for improving PSO optimization neural networks | |
| CN103335814B (en) | Correction method for inclination angle measurement error data of experimental model in wind tunnel | |
| CN108955727B (en) | Optical fiber coil performance evaluation method | |
| CA2499272A1 (en) | Simultaneous resolution enhancement and dip correction of resistivity logs through nonlinear iterative deconvolution | |
| CN103487011B (en) | A kind of attitude angle detection method of data glove | |
| CN102243080A (en) | Signal detecting method and device with function of temperature compensation for high-accuracy optical fiber gyroscope | |
| CN104048676B (en) | MEMS (Micro Electro Mechanical System) gyroscope random error compensating method based on improved particle filter | |
| CN107966242B (en) | One-touch baroceptor field calibration system and method suitable for deep ocean buoy | |
| CN106909065A (en) | A kind of real-time clock clock correction bearing calibration of prospecting instrument | |
| CN109521384A (en) | A kind of vector magnetic compensation method based on atom magnetometer | |
| CN103592484A (en) | Electricity larceny prevention method for preventing strong magnet from generating error influences on electronic-type energy meter | |
| CN111308198B (en) | Harmonic measurement method of windowed interpolation DFT based on Hanning window | |
| CN104748714B (en) | A kind of spaceborne rope net deployable antenna surface accuracy measuring method | |
| CN109143942A (en) | A kind of attitude transducer control system | |
| CN102508503B (en) | Compensation method based on generalized inner module for eccentric torque of three-shaft inertially stabilized platform | |
| CN210666065U (en) | Calibration device of seismic intensity instrument | |
| CN103940453A (en) | Method for improving sensor measuring precision | |
| CN201335718Y (en) | Bubble-type water gauge | |
| CN202903814U (en) | Temperature compensation and time drift correction device for acceleration sensor | |
| CN109270455A (en) | Induction machine state monitoring method based on hyposensitiveness Ensemble Kalman Filter | |
| CN101598540A (en) | 3-D positioning method and 3 D positioning system | |
| CN103674120A (en) | Composite gas flow measuring device based on MEMS (micro-electromechanical system) sensitive elements and measuring method thereof | |
| CN102944688A (en) | Temperature compensation and time drift correction device and method for acceleration sensor | |
| CN201876464U (en) | Intelligent multifunctional speed and displacement calibrating system |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20150930 |
|
| RJ01 | Rejection of invention patent application after publication |