CN112946318A - Calibration algorithm of acceleration sensor - Google Patents
Calibration algorithm of acceleration sensor Download PDFInfo
- Publication number
- CN112946318A CN112946318A CN202110311605.2A CN202110311605A CN112946318A CN 112946318 A CN112946318 A CN 112946318A CN 202110311605 A CN202110311605 A CN 202110311605A CN 112946318 A CN112946318 A CN 112946318A
- Authority
- CN
- China
- Prior art keywords
- acceleration sensor
- data
- neural network
- input
- calibration
- 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
- 230000001133 acceleration Effects 0.000 title claims abstract description 36
- 238000013528 artificial neural network Methods 0.000 claims abstract description 30
- 238000000034 method Methods 0.000 claims abstract description 23
- 238000012549 training Methods 0.000 claims abstract description 11
- 238000012545 processing Methods 0.000 claims abstract description 8
- 230000008569 process Effects 0.000 claims abstract description 7
- 230000006870 function Effects 0.000 claims description 10
- 230000007613 environmental effect Effects 0.000 abstract description 4
- 230000005484 gravity Effects 0.000 description 4
- 230000008901 benefit Effects 0.000 description 3
- 230000004044 response Effects 0.000 description 3
- 230000003068 static effect Effects 0.000 description 3
- 238000004364 calculation method Methods 0.000 description 2
- 238000004519 manufacturing process Methods 0.000 description 2
- 238000005259 measurement Methods 0.000 description 2
- 230000004075 alteration Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 210000004556 brain Anatomy 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 230000010365 information processing Effects 0.000 description 1
- 238000002955 isolation Methods 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 238000013178 mathematical model Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000007230 neural mechanism Effects 0.000 description 1
- 230000035945 sensitivity Effects 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01P—MEASURING LINEAR OR ANGULAR SPEED, ACCELERATION, DECELERATION, OR SHOCK; INDICATING PRESENCE, ABSENCE, OR DIRECTION, OF MOVEMENT
- G01P21/00—Testing or calibrating of apparatus or devices covered by the preceding groups
-
- 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
- G06N3/084—Backpropagation, e.g. using gradient descent
Abstract
The invention discloses a calibration algorithm of an acceleration sensor, belonging to the technical field of calibration algorithms, and particularly relating to a calibration algorithm of an acceleration sensor, which comprises the following steps of measuring output data Yi of an unknown system by giving different input Xi, using the collected input and output data (Xi, Yi) for training a BP neural network, wherein in the training process of the BP neural network, firstly, randomly selecting some data as weight values and threshold values in a structure, transmitting the input data Xi to a hidden layer for processing through an input layer, the node action function of the hidden layer is usually an S-shaped nonlinear function, transmitting the data to an output layer after the data is processed by the nonlinearity of the hidden layer, finally generating output data Oi, and adopting the BP neural network calibration algorithm with a three-layer structure to calibrate the acceleration sensor, so that the limitation of the traditional calibration method can be solved, the calibration precision and efficiency are improved, and the problem that the calibration is often influenced by load quality and environmental factors is avoided.
Description
Technical Field
The invention relates to the technical field of calibration algorithms, in particular to a calibration algorithm of an acceleration sensor.
Background
The acceleration sensor is a necessary device for acquiring information in various equipment and information systems nowadays, is largely used in the fields of industrial control, automobiles, medical devices, instruments and meters, digital products and the like, plays a vital role in improving the technical level of production and economic benefit, and converts the information needing to be acquired in the processes of scientific experiments and production, particularly in an automatic detection and closed-loop control system, into an electric signal which can be easily processed and transmitted by the system through the acceleration sensor. The accuracy of any meters and devices is not significant any more if the acceleration sensor is not able to make an accurate and reliable measurement of the original signal or to accurately convert the sensed measurement into an electrical signal.
A calibration algorithm of an acceleration sensor adopts a BP neural network calibration algorithm with a three-layer structure to calibrate the acceleration sensor, can solve the limitation of the traditional calibration method and improve the calibration precision and efficiency, the BP neural network is an information processing mathematical model simulating the human brain neural mechanism and consists of an input layer, an output layer and a hidden layer, wherein the hidden layer can be one layer or a plurality of layers, the BP neural network has the characteristics of parallel processing, associative memory and strong anti-interference performance, and can theoretically process the nonlinear mapping problem of unclear inference rule n- > m dimensional space, the BP neural network with the three-layer structure can fit any nonlinear curve with any precision, and the BP neural network calibration algorithm with the three-layer structure can calibrate the acceleration sensor, can solve the limitation of the traditional calibration method and improve the calibration precision and efficiency, the calibration algorithm of the current acceleration sensor has low precision and efficiency in the use process, and is often influenced by load quality and environmental factors.
Disclosure of Invention
The invention aims to provide a calibration algorithm of an acceleration sensor, and aims to solve the problems that the existing BP neural network acceleration calibration algorithm in the background art is low in precision and efficiency and can be greatly influenced by load quality and other environmental factors.
In order to achieve the purpose, the invention provides the following technical scheme: a calibration algorithm of an acceleration sensor comprises the following specific steps:
(1) for an unknown system, measuring output data Yi of the unknown system by giving different input Xi, and then using the collected input and output data (Xi, Yi) for training a BP neural network;
(2) in the training process of the BP neural network, firstly, randomly selecting some data as weight values and threshold values in a structure, and transmitting input data Xi to a hidden layer through an input layer for processing;
(3) the node function of the hidden layer is usually an S-shaped nonlinear function, data is transmitted to the output layer after being subjected to nonlinear processing of the hidden layer, and finally output data Oi is generated;
(4) comparing the output result Oi with the expected output value Yi, and calculating an error Ei between the output result Oi and the expected output value Yi;
(5) and if the Ei cannot meet the requirement of the system, the BP neural network modifies the weight and the threshold value in the structure according to the value of the Ei and the learning algorithm until the Ei meets the requirement of the system.
Preferably, in the present system, the acceleration sensor is regarded as a "black box", the input/output data of the acceleration sensor is used as a sample of the BP neural network, and the trained BP neural network is made to be equivalent to the characteristic function of the input/output of the acceleration sensor by the training network.
Compared with the prior art, the invention has the beneficial effects that:
the calibration algorithm of the invention adopts a BP neural network calibration algorithm with a three-layer structure to calibrate the acceleration sensor, can solve the limitation of the traditional calibration method, improves the calibration precision and efficiency, and avoids the problem that the acceleration sensor is often influenced by load quality and environmental factors.
Drawings
Fig. 1 is a schematic flow chart of the calibration algorithm.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
In the description of the present invention, it is to be understood that the terms "upper", "lower", "front", "rear", "left", "right", "top", "bottom", "inner", "outer", and the like, indicate orientations or positional relationships based on the orientations or positional relationships shown in the drawings, are merely for convenience in describing the present invention and simplifying the description, and do not indicate or imply that the device or element being referred to must have a particular orientation, be constructed and operated in a particular orientation, and thus, should not be construed as limiting the present invention.
Example (b):
referring to fig. 1, the present invention provides a technical solution: a calibration algorithm of an acceleration sensor comprises the following specific steps:
(1) for an unknown system, measuring output data Yi of the unknown system by giving different input Xi, and then using the collected input and output data (Xi, Yi) for training a BP neural network;
(2) in the training process of the BP neural network, firstly, randomly selecting some data as weight values and threshold values in a structure, and transmitting input data Xi to a hidden layer through an input layer for processing;
(3) the node function of the hidden layer is usually an S-shaped nonlinear function, data is transmitted to the output layer after being subjected to nonlinear processing of the hidden layer, and finally output data Oi is generated;
(4) comparing the output result Oi with the expected output value Yi, and calculating an error Ei between the output result Oi and the expected output value Yi;
(5) and if the Ei cannot meet the requirement of the system, the BP neural network modifies the weight and the threshold value in the structure according to the value of the Ei and the learning algorithm until the Ei meets the requirement of the system.
Taking the calibration of a typical sensor, i.e. an acceleration sensor, as an example, several common calibration methods are introduced in detail.
1. A gravity field method: for the accelerometer with zero frequency response, the calibration can be performed by using the earth static gravity field method, as shown in the figure, the earth gravity field method is also an absolute calibration method and has higher precision, and the method needs to be noted in the use: firstly, the method belongs to static absolute calibration in principle and is only suitable for the calibration of accelerometers with zero-frequency response (such as servo type, piezoresistive type, strain type and tension type); the device must have good vibration isolation foundation, the rotating shaft should keep horizontal position strictly, otherwise the precision will be influenced;
2. turntable type gravity field calibration method: the zero-frequency calibration belongs to static calibration, a zero-frequency calibration device is properly improved, namely, a table top rotates at a constant speed at a certain frequency to form a low-frequency gravitational field calibration table, when the table top rotates, the sensitive element of the acceleration sensor is subjected to alternating acceleration of-1 to +1g, the output values of the acceleration sensor under different frequencies are recorded, and the dynamic sensitivity of the acceleration sensor and the acceleration sensor can be obtained, but the method is only suitable for low frequency;
the 3 comparison method is the most common method for calibrating the sensor, and has a series of advantages of simple principle, convenient operation, low requirement on equipment and the like, so the method is widely applied, but the high-frequency response of the standard sensor is influenced by the mass load.
In the system, the acceleration sensor is regarded as a black box, the input and output data of the acceleration sensor are used as samples of a BP neural network, and the trained BP neural network is equivalent to a characteristic function of the input and output of the acceleration sensor through a training network, wherein a specific calculation formula is as follows:
first, a connection weight value from the ith node in the k (k ═ 1,2, 3) layer to the jth node in the k +1 layer is defined as
Defining the threshold value of the ith node of the k layers as
Input is as
Output is as
Defining the input-output function of the layer
Assume that the system is capable of providing N pairs of sample data (X)i,Yi) (i ═ 1,2,3.. N), so X is input for in a BP neural networkiThe input value of the jth node of the output layer is
And the desired output is YjObtained using a squared error function
The BP neural network continuously reduces the output error E according to a gradient descent algorithm
Is changed by an amount of each change of
Wherein
According to formula (5,12), formula (5,13), formula (5,14), formula (5,15), formula (5,16) there are:
in BP neural network
Called error signal, which starts from the input layer, travels to the previous layer in the opposite direction of the data calculation, and finally to the output layer, in equation (5,17)
This feature can be provided, which is also the root cause of the BP neural network called error back propagation neural network.
While there have been shown and described the fundamental principles and essential features of the invention and advantages thereof, it will be apparent to those skilled in the art that the invention is not limited to the details of the foregoing exemplary embodiments, but is capable of other specific forms without departing from the spirit or essential characteristics thereof; the present embodiments are therefore to be considered in all respects as illustrative and not restrictive, the scope of the invention being indicated by the appended claims rather than by the foregoing description, and all changes which come within the meaning and range of equivalency of the claims are therefore intended to be embraced therein, and any reference signs in the claims are not intended to be construed as limiting the claim concerned.
Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that changes, modifications, substitutions and alterations can be made in these embodiments without departing from the principles and spirit of the invention, the scope of which is defined in the appended claims and their equivalents.
Claims (2)
1. The calibration algorithm of the acceleration sensor is characterized by comprising the following specific steps:
(1) for an unknown system, measuring output data Yi of the unknown system by giving different input Xi, and then using the collected input and output data (Xi, Yi) for training a BP neural network;
(2) in the training process of the BP neural network, firstly, randomly selecting some data as weight values and threshold values in a structure, and transmitting input data Xi to a hidden layer through an input layer for processing;
(3) the node function of the hidden layer is usually an S-shaped nonlinear function, data is transmitted to the output layer after being subjected to nonlinear processing of the hidden layer, and finally output data Oi is generated;
(4) comparing the output result Oi with the expected output value Yi, and calculating an error Ei between the output result Oi and the expected output value Yi;
(5) and if the Ei cannot meet the requirement of the system, the BP neural network modifies the weight and the threshold value in the structure according to the value of the Ei and the learning algorithm until the Ei meets the requirement of the system.
2. The calibration algorithm for an acceleration sensor according to claim 1, characterized in that: in the system, the acceleration sensor is regarded as a black box, the input and output data of the acceleration sensor are used as samples of the BP neural network, and the trained BP neural network is equivalent to the characteristic function of the input and output of the acceleration sensor through the training network.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110311605.2A CN112946318A (en) | 2021-03-24 | 2021-03-24 | Calibration algorithm of acceleration sensor |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110311605.2A CN112946318A (en) | 2021-03-24 | 2021-03-24 | Calibration algorithm of acceleration sensor |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN112946318A true CN112946318A (en) | 2021-06-11 |
Family
ID=76228000
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110311605.2A Pending CN112946318A (en) | 2021-03-24 | 2021-03-24 | Calibration algorithm of acceleration sensor |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112946318A (en) |
Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101158588A (en) * | 2007-11-16 | 2008-04-09 | 北京航空航天大学 | MEMS gyroscopes error compensation method for micro satellite based on integration nerval net |
| CN103400123A (en) * | 2013-08-21 | 2013-11-20 | 山东师范大学 | Gait type identification method based on three-axis acceleration sensor and neural network |
| US20180108440A1 (en) * | 2016-10-17 | 2018-04-19 | Jeffrey Stevens | Systems and methods for medical diagnosis and biomarker identification using physiological sensors and machine learning |
| CN108037317A (en) * | 2017-12-06 | 2018-05-15 | 中国地质大学（武汉） | The dynamic decoupling method and system of a kind of accelerometer |
| 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 |
| US20200111005A1 (en) * | 2018-10-05 | 2020-04-09 | Sri International | Trusted neural network system |
| CN111284497A (en) * | 2020-01-20 | 2020-06-16 | 北京津发科技股份有限公司 | Driving state recognition device |
| CN111458748A (en) * | 2020-03-30 | 2020-07-28 | 青岛理工大学 | Performance earthquake motion risk analysis method based on three-layer data set neural network |
| CN112305473A (en) * | 2020-10-23 | 2021-02-02 | 哈尔滨工程大学 | Calibration method of three-axis TMR sensor |
-
2021
- 2021-03-24 CN CN202110311605.2A patent/CN112946318A/en active Pending
Patent Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101158588A (en) * | 2007-11-16 | 2008-04-09 | 北京航空航天大学 | MEMS gyroscopes error compensation method for micro satellite based on integration nerval net |
| CN103400123A (en) * | 2013-08-21 | 2013-11-20 | 山东师范大学 | Gait type identification method based on three-axis acceleration sensor and neural network |
| US20180108440A1 (en) * | 2016-10-17 | 2018-04-19 | Jeffrey Stevens | Systems and methods for medical diagnosis and biomarker identification using physiological sensors and machine learning |
| CN108037317A (en) * | 2017-12-06 | 2018-05-15 | 中国地质大学（武汉） | The dynamic decoupling method and system of a kind of accelerometer |
| 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 |
| US20200111005A1 (en) * | 2018-10-05 | 2020-04-09 | Sri International | Trusted neural network system |
| CN111284497A (en) * | 2020-01-20 | 2020-06-16 | 北京津发科技股份有限公司 | Driving state recognition device |
| CN111458748A (en) * | 2020-03-30 | 2020-07-28 | 青岛理工大学 | Performance earthquake motion risk analysis method based on three-layer data set neural network |
| CN112305473A (en) * | 2020-10-23 | 2021-02-02 | 哈尔滨工程大学 | Calibration method of three-axis TMR sensor |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Liu et al. | Subspace network with shared representation learning for intelligent fault diagnosis of machine under speed transient conditions with few samples | |
| CN109188026B (en) | Automatic calibration deep learning method suitable for MEMS accelerometer | |
| CN108073075A (en) | Silicon micro accerometer temperature-compensation method, system based on GA Optimized BP Neural Networks | |
| CN114548368B (en) | Modeling method and prediction method of lithium battery temperature field prediction model based on multilayer nuclear overrun learning machine | |
| CN106908095B (en) | A kind of extraction and appraisal procedure of sensing data alignment features | |
| CN102023010A (en) | MEMS (micro-electromechanical system)-based wavelet field multisensor information fusion system and fusion method | |
| CN104921736A (en) | Continuous blood glucose monitoring device comprising parameter estimation function filtering module | |
| CN111307452B (en) | Intelligent fault diagnosis method for rotating machinery at time-varying rotating speed | |
| CN108844702A (en) | The measuring method of Bridge Influence Line when vehicle at the uniform velocity passes through | |
| CN103808349A (en) | Error correction method and device for vector sensors | |
| Liu et al. | A semi-convex function for both constant and time-varying moving force identification | |
| CN111076748A (en) | Horizontal inclinometer error compensation method and system based on MEMS accelerometer | |
| CN115688288B (en) | Aircraft pneumatic parameter identification method and device, computer equipment and storage medium | |
| CN112946318A (en) | Calibration algorithm of acceleration sensor | |
| CN112906185A (en) | MEMS inertial sensor heterogeneous array based on artificial intelligence and design method thereof | |
| CN116702005A (en) | Neural network-based data anomaly diagnosis method and electronic equipment | |
| CN109405884B (en) | System and method for realizing humidity calibration function based on temperature and humidity sensor | |
| CN111598215A (en) | Temperature compensation method and system based on neural network | |
| CN114084764B (en) | Elevator transportation quality detection method and detection system | |
| CN116362140A (en) | Sensor error compensation method, system, device and medium based on fusion of transducer and GRU | |
| CN112747773B (en) | Method for improving precision of gyroscope based on Allan variance and random polynomial | |
| CN113820062B (en) | Temperature compensation method of six-dimensional force sensor | |
| Hernandez | Improving the response of a rollover sensor placed in a car under performance tests by using a RLS lattice algorithm | |
| CN113780520A (en) | Landslide MEMS acceleration sensor error compensation method | |
| CN114705184A (en) | Nine-axis attitude sensor integrated intelligent error compensation method based on neural network |
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 | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20210611 |
|
| RJ01 | Rejection of invention patent application after publication |