CN103592465B - Triaxial micromachine accelerometer static correction method based on particle swarm optimization - Google Patents

Abstract

The invention discloses a triaxial micromachine accelerometer static correction method based on particle swarm optimization. The method includes the steps that a triaxial MEMS accelerometer and a relevant peripheral circuit are welded to a designated circuit board to form a test and using circuit of the accelerometer, and normal operation and data output of the accelerometer are guaranteed; the circuit board is fixed to a certain space point, and the circuit board can be freely rotated on three axial directions of space; the circuit board provided with the accelerometer is manually or automatically driven by an external device to rotate in the space, the rotating speed is smaller than 10 degrees per second, the rotation track of the accelerometer covers main distribution of a spherical surface, and actual output data of the accelerometer are recorded; after the collected data are preprocessed, an error estimation algorithm of the accelerometer is designed according to a standard PSO processing method; an error correction model of the accelerometer is set according to the estimated error model, and measurement accuracy of the accelerometer is improved. By the adoption of the method, random errors in the sampling process are effectively eliminated, and measurement accuracy of the MEMS accelerometer is further improved.

Description

Based on three axle micro-mechanical accelerometer static modification methods of particle group optimizing

Technical field

The present invention relates to a kind of static modification method of the three axle micro-mechanical accelerometers based on particle group optimizing, during the method exports according to the error model correction three axis accelerometer calculated zero partially, the dominant static error such as quantization error and quadrature error.

Background technology

Along with the raising day by day of MEMS (micro electro mechanical system) (Microelectromechanical Systems, is abbreviated as MEMS) processing and manufacturing technology, the use of MEMS obtains the extensive concern of people and promotes rapidly, and wherein the application of mems accelerometer is the most wide.Due to manufacture reason and the using method of MEMS sensor, there is certain measuring error in the Static output of mems accelerometer, limits its result of use, needs to revise to improve measuring accuracy to it.

At present, use comparatively general static modification method to be max min method in engineering, the method only can estimate scaling factor error and zero drift, and is easily subject to the interference of stochastic error.Another kind of comparatively conventional modification method tests with gyroscope composition Inertial Measurement Unit, concrete has combination calibration compensation method, the reference data that the method mainly relies on rotation platform to provide carries out error correction, can calibrate static and dynamic errors, the method model is perfect simultaneously, and correction effect precision is high, but in makeover process, need multi-pass operations rotation platform, makeover process complexity is loaded down with trivial details, and the cost of rotation platform is also beyond the tolerance range of domestic consumer.

Summary of the invention

The object of the invention is the measuring accuracy improving mems accelerometer further, effectively the static error parameter of estimated acceleration meter, set up a more perfect and comprehensive error model, thus improve the result of use of mems accelerometer.

Instant invention overcomes shortcoming of the prior art, a kind of static modification method of the three axle micro-mechanical accelerometers based on particle group optimizing is provided.By analyzing measuring principle and the using method thereof of mems accelerometer, determine its main source of error, on original error model basis, propose the accelerometer static modification method based on particle group optimizing (Particle Swarm Optimization, PSO).Because the existence of measuring error, actual measured value is distributed in center for (bx, by, bz) on ellipsoid, therefore, the parameter estimation be converted to the static modification of mems accelerometer under ellipsoid constraint, determine the fitness function of PSO algorithm for estimating, according to standard particle group algorithm determination error model.

In order to solve the technical matters of above-mentioned existence, realize goal of the invention, the present invention is achieved by the following technical solutions:

Based on a static modification method for three axle micro-mechanical accelerometers of particle group optimizing, its content comprises the steps:

3 axis MEMS accelerometer and relevant peripheral circuit are welded on the circuit board of specifying by the first step, the test of composition accelerometer and use circuit, ensure that the normal running of accelerometer and data export; Then circuit board is fixed on a certain spatial point, ensures that circuit board axially can rotate freely three of space;

Second step passes through external device (ED), manual or automated manner drive installation has the circuit board of accelerometer at Space Rotating, rotational speed is less than 10 °/s, ensures that the rotational trajectory of accelerometer covers the main distribution of sphere, records the actual output data of accelerometer simultaneously;

After the data of collection are carried out pre-service by the 3rd step, according to the disposal route of standard P SO, the error estimation algorithm of design acceleration meter;

4th step, according to the error model estimated, arranges the VEC of acceleration, improves the measuring accuracy of accelerometer.

Owing to adopting technique scheme, the static modification method of a kind of three axle micro-mechanical accelerometers based on particle group optimizing provided by the invention, compared with prior art has such beneficial effect:

The inventive method eliminates the stochastic error in sampling process effectively, further increases the measuring accuracy of mems accelerometer.The method does not need three axle rotation platforms, reduces the use cost of user, calculates simply, is easy to realize, make mems accelerometer better meet engineering requirements, for the environment such as outdoor test provide new selection.

Accompanying drawing explanation

Fig. 1 is three axis accelerometer alignment error figure;

Fig. 2 is sphere and the ellipsoid distribution plan of three axis accelerometer vector;

Fig. 3 is PSO calculation flow chart;

Fig. 4 is the protocol procedures figure of three axis accelerometer static modification;

Embodiment

Below in conjunction with accompanying drawing and embodiment, a more detailed description of the present invention is to do:

In experiment, use the three axis accelerometer ADXL345 that ADI company produces.Arranging it at this is 13 A/D conversions, and measuring accuracy is 3.9mg/LSB, and range is-2g ~ 2g, therefore 1g=255Counts.

Based on a static modification method for three axle micro-mechanical accelerometers of particle group optimizing, as shown in Figure 4, its content comprises the steps:

First step degree of will speed up meter control circuit and carrier are fixed, and keep its Z axis parallel with bearer plane vertical line as far as possible, XY plane is parallel with bearer plane, to reduce alignment error.Ensure that circuit board rotates freely at spatial triaxial.

Second step is by external device (ED), and manual actuation is provided with the circuit board of accelerometer and carrier at Space Rotating, and rotational speed is less than 10 °/s; By I2C bus, carry out the sampling of gravity acceleration value raw data according to 40Hz, sample 1936 times altogether, sampled point substantially covers the institute of sphere a little, to ensure precision and the accuracy of parameter estimation.

3rd step, by analyzing measuring principle and the using method thereof of mems accelerometer, can determine that its main error comprises zero partially, quantization error, quadrature error, alignment error etc., as shown in Figure 1.Consider the measuring accuracy of accelerometer and the practicality of correction result, ignore second order error item at this.According to its error features, the error model of MEMS triaxial accelerometer in certain MEMS gyro strapdown inertial navitation system (SINS) can be obtained, be shown below.

$\left[\begin{array}{c}{h}_x\\ h_y\\ h_z\end{array}\right]=\left[\begin{array}{c}{b}_x+n_x\\ b_y+n_y\\ b_z+n_z\end{array}\right]+\left[\begin{array}{ccc}{S}_x& -a_4.S_y& -a_5.S_z\\ a_1.S_x& S_y& -a_6.S_z\\ -a_2.S_x& a_3.S_y& S_z\end{array}\right]\left[\begin{array}{c}{g}_x\\ g_y\\ g_z\end{array}\right]=\left[\begin{array}{c}{b}_x+n_x\\ b_y+n_y\\ b_z+n_z\end{array}\right]+\left[\begin{array}{ccc}{S}_x& K_{12}& K_{13}\\ K_{21}& S_y& K_{23}\\ K_{32}& K_{32}& S_z\end{array}\right]\left[\begin{array}{c}{g}_x\\ g_y\\ g_z\end{array}\right]$

In formula, h _x, h _y, h _zfor the gravity field vector of the actual measurement of mems accelerometer; b _x, b _y, b _zfor mems accelerometer is each axial zero inclined; a ₁, a ₂, a ₃, a ₄, a ₅, a ₆for the quadrature error in mems accelerometer installation process; S _x, S _y, S _zfor the quantization error of mems accelerometer, n _x, n _y, n _zfor mems accelerometer random noise.

When accelerometer is in static, accelerometer exports as 1g, gets || g||=9.8m/s ², therefore acceleration true value should be distributed in the center of circle for (0,0,0), radius and is || g _ref||=9.8m/s ²sphere on.Because the existence of measuring error, actual measured value be distributed on an ellipsoid, the sphere of accelerometer vector and ellipsoid distribution more as shown in Figure 2.Therefore the static modification of accelerometer be converted to ellipsoid constraint under parameter estimation.In order to contrast, now provide the accelerometer error model that the least square method based on Ellipsoidal Restrictions draws, parameter to be estimated is 9, is shown below:

$\left[\begin{array}{c}{g}_x\\ g_y\\ g_z\end{array}\right]=\left[\begin{array}{ccc}{K}_{11}& K_{12}& K_{13}\\ 0& K_{22}& K_{23}\\ 0& 0& K_{33}\end{array}\right]\left[\begin{array}{c}{h}_x-b_x\\ h_y-b_y\\ h_z-b_z\end{array}\right]$

In order to effectively simplify the static modification method of accelerometer, the inventive method, on original error model basis, proposes the accelerometer PSO static modification method based on Ellipsoidal Restrictions.

Can determine that the fitness function of PSO algorithm for estimating is

$O\left(p\right)=\frac{1}{N}\underset{n=1}{\overset{n=N}{\mathrm{\Σ}}}{\left(\right|\left|g_{\mathrm{ref}}\right||-|\left|g_n\right|\left|\right)}^2$ Wherein $\left[\begin{array}{c}{g}_{\mathrm{nx}}\\ g_{\mathrm{ny}}\\ g_{\mathrm{nz}}\end{array}\right]=\left[\begin{array}{ccc}{K}_{11}& K_{12}& K_{13}\\ K_{21}& K_{22}& K_{23}\\ K_{31}& K_{32}& K_{33}\end{array}\right]\·\left[\begin{array}{c}{h}_x-b_x\\ h_y-b_y\\ h_z-b_z\end{array}\right]$

Wherein N is sampling number, g _reffor local gravitational acceleration value, h is actual samples data, g _nfor revised acceleration of gravity.Parameter to be estimated is 12.According to standard particle colony optimization algorithm, first corresponding parameter is set, in possible region of search, random initializtion is carried out to population, comprise position and speed.Then the fitness of each particle is calculated; According in the ring topology of call number, choose the radius of neighbourhood, calculate the personal best particle of each particle and the fitness of local optimal location; The following speed to particle and position upgrade; The personal best particle of final updating particle and fitness thereof, and and the fitness of local optimum position in contiguous range relatively after, upgrade its local optimum position and its fitness.If iteration terminates, then export optimal location and fitness thereof, otherwise recalculate the fitness of each particle, its detailed process as shown in Figure 3.

4th step according to the method described above, establishes the accelerometer static modification error model based on PSO.The validity of the method is verified by experiment.

According to PSO method, static modification process is carried out to sampling the data obtained.According to said method obtain VEC result as follows:

$\left[\begin{array}{c}{g}_x\\ g_y\\ g_z\end{array}\right]=\left[\begin{array}{ccc}0.9871& 0.0029& -0.0144\\ 0.0751& 0.9813& 0.0548\\ 0.0243& -0.0268& 0.9803\end{array}\right]\·\left[\begin{array}{c}{h}_x+4.7816\\ h_y-1.0140\\ h_z-33.5546\end{array}\right]$

Experimental result is as shown in the table, least square method based on Ellipsoidal Restrictions and the PSO algorithm based on Ellipsoidal Restrictions are compared, analyze from mean value and standard deviation, two kinds of static modification methods all eliminate main error, demonstrate the validity of Ellipsoidal Restrictions, and PSO revised law substantially reduces the measuring error of acceleration of gravity, cause the main cause of this difference to be that least square method has carried out simplify processes to mathematical model, can only estimate fractional error; And the error parameter that PSO static modification method is estimated, include the main error parameter of acceleration, give the optimal estimation in whole space, provide error model more comprehensively, therefore closer to desirable error model, so its parameter estimation effect is even more ideal.

Amplitude com parison before and after the correction of table 1 acceleration of gravity

Claims (1)

1., based on a static modification method for three axle micro-mechanical accelerometers of particle group optimizing, it is characterized in that: the method content comprises the steps:

Three axle micro-mechanical accelerometers and relevant peripheral circuit are welded on the circuit board of specifying by the first step, the test of composition accelerometer and use circuit, ensure that the normal running of accelerometer and data export; Then circuit board is fixed on a certain spatial point, ensures that circuit board axially can rotate freely three of space;

Second step passes through external device (ED), manual or automated manner drive installation has the circuit board of accelerometer at Space Rotating, rotational speed is less than 10 °/s, ensures that the rotational trajectory of accelerometer covers the main distribution of sphere, records the actual output data of accelerometer simultaneously;

After the data of collection are carried out pre-service by the 3rd step, the disposal route optimized according to standard particle group, the error estimation algorithm of design acceleration meter;

4th step, according to the error model estimated, arranges the VEC of acceleration, improves the measuring accuracy of accelerometer.