CN108827301A - A kind of improvement error quaternion Kalman filtering robot pose calculation method - Google Patents
A kind of improvement error quaternion Kalman filtering robot pose calculation method Download PDFInfo
- Publication number
- CN108827301A CN108827301A CN201810346303.7A CN201810346303A CN108827301A CN 108827301 A CN108827301 A CN 108827301A CN 201810346303 A CN201810346303 A CN 201810346303A CN 108827301 A CN108827301 A CN 108827301A
- Authority
- CN
- China
- Prior art keywords
- gravitational acceleration
- quaternion
- error
- error quaternion
- matrix
- 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
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01C—MEASURING DISTANCES, LEVELS OR BEARINGS; SURVEYING; NAVIGATION; GYROSCOPIC INSTRUMENTS; PHOTOGRAMMETRY OR VIDEOGRAMMETRY
- G01C21/00—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00
- G01C21/20—Instruments for performing navigational calculations
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01C—MEASURING DISTANCES, LEVELS OR BEARINGS; SURVEYING; NAVIGATION; GYROSCOPIC INSTRUMENTS; PHOTOGRAMMETRY OR VIDEOGRAMMETRY
- G01C21/00—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00
- G01C21/10—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00 by using measurements of speed or acceleration
Landscapes
- Engineering & Computer Science (AREA)
- Radar, Positioning & Navigation (AREA)
- Remote Sensing (AREA)
- Automation & Control Theory (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Manipulator (AREA)
- Gyroscopes (AREA)
Abstract
The present invention is a kind of improvement error quaternion Kalman filtering robot pose calculation method, is mainly acted in pose of mobile robot resolving system, is allowed to obtain accurate attitude angle and is convenient for subsequent gesture stability.This method obtains attitude data using gyroscope accelerometer, construct systematic error quaternary number Kalman filtering state equation and measurement equation, and determine that formula judges whether there is non-gravitational acceleration according to non-gravitational acceleration, and adaptive equalization non-gravitational acceleration influences attitude algorithm bring, then error quaternion Kalman filter is utilized, error quaternion updated value is obtained, true attitude quaternion is calculated, attitude algorithm is carried out according to Quaternion Method and obtains attitude angle.The present invention can be realized full attitude algorithm, and can obviously inhibit the influence of noise, drift error and non-gravitational acceleration to attitude algorithm.
Description
Technical field
The present invention is a kind of improvement error quaternion Kalman filtering robot pose calculation method, and the invention belongs to numbers
Filtering and multisensor Data Fusion technology field.The invention is suitable for the attitude algorithm of mobile robot, is also applied for simultaneously
Mobile electric vehicle.
Background technique
The accuracy and speed of attitude algorithm will directly affect the stability, reliability and the difficulty or ease of realization of gesture stability algorithm
Degree, so, the premise that attitude algorithm is mobile or flight control is realized.The inexpensive posture that mobile robot generally uses passes
Sensor, due to sensor self character, the influence of accelerometer noise is very big, and gyroscope signal has biased error, works as inertia
When sensor works long hours, error will cause system can not work normally with accumulated time.Pose of mobile robot at present
Calculation method mainly has Kalman filtering method, Extended Kalman filter method and particle filter method, Extended Kalman filter method appearance
State calculation method real-time is preferable, but linearisation has lost part quadratic value, and linearized stability is larger.Kalman filtering method
Attitude algorithm method modeling is simple, and real-time is preferable, but has ignored non-linear factor, and there are sample degeneracies to ask for particle filter method
Topic, real-time is poor, and not easy enough.The invention patent is a kind of effective pose of mobile robot solution regarding to the issue above
Calculation method.
Summary of the invention
The present invention solves the technical problem of the deficiencies for being directed to above-mentioned background technique, provide a kind of improvement error four
First number Kalman filtering robot pose calculation method.The present invention adopts the following technical scheme that for achieving the above object.
A kind of improvement error quaternion Kalman filtering robot pose calculation method, includes the following steps：
Step 1：The corresponding reference frame of mobile robot is chosen, posture changing matrix are determined, with nonlinear exponent mould
Type establishes gyroscopic drift mathematical model in conjunction with Levenberg-Marquardt iterative least squares approximating method, utilizes gyro
Relationship between instrument output data and error quaternion establishes the error quaternion differential equation；
Step 2：Error quaternion Kalman is established according to the error quaternion differential equation and gyroscopic drift mathematical model
Filter state equation；
Step 3：Error quaternion Kalman is established according to the relationship between accelerometer output data and error quaternion
Filter observational equation；
Step 4：Judge that mobile robot whether there is non-gravitational acceleration according to accelerometer output data, and adaptive
Update the covariance matrix of mobile robot non-gravitational acceleration；
Step 5：It is compensated using the non-gravitational acceleration covariance matrix of adaptive updates to noise matrix is measured, and
Error quaternion updated value is obtained using error quaternion Kalman filter, true attitude quaternion is calculated, according to
Quaternion Method carries out attitude algorithm and obtains attitude angle.
Further, in step 1, northeast day coordinate system is chosen as navigational coordinate system, between navigation system and carrier system
Following posture changing matrix can be used in transformation relationIt indicates：
Wherein, θ is pitch angle, and γ is roll angle, and ψ is course angle.
Output from Gyroscope is unrelated with moving condition, but gyroscope signal has biased error.Therefore gyroscope is surveyed
Amount model is angular speed, biased error, the summation of white noise：
ωo=ω+εg+ng
Wherein, ωoIndicate the output of gyroscope, ω indicates actual angular speed, εgIndicate the drift of gyroscope, ngIndicate top
Spiral shell instrument zero-mean white noise.
With nonlinear exponent model, gyro is established in conjunction with Levenberg-Marquardt iterative least squares approximating method
Instrument drift mathematical model；
Wherein, C1、C2, T be model of fit three parameters, TsFor the sampling time.
Due to ωo≠ ω, so the quaternary number solved by gyroscope output valve is not true quaternary number, we claim
The estimated value for quaternary number, be denoted asIt is to have small error between the estimated value of quaternary number and true quaternary number, him
Between error be known as error quaternion, be denoted as
DefinitionFor following formula：
In above-mentioned formula, definition：
Wherein, qe1、qe2、qe3For three parameters of error quaternion.
Relationship between error quaternion and quaternary number estimated value and true quaternary number is as follows：
Wherein, Q is true quaternary number.
To formula derivation above and abbreviation can obtain：
Wherein, [ωo×] indicate its antisymmetric matrix.
Further, in step 2, selecting system quantity of state is x=[qe1 qe2 qe3 εgx εgy εgz]T。
Wherein, εgx、εgy、εgzIndicate three axis component of gyroscopic drift.
Following error quaternion Kalman can be established according to the error quaternion differential equation and gyroscopic drift mathematical model
Filter state equation：
xk=Axk-1+Buk-1+wk
Wherein, A indicates that systematic procedure matrix, B indicate that system controls matrix coefficient, and u-system controls matrix, indicates that w is indicated
System mode noise.
B=I
Wherein, I indicates unit matrix.
Further, in step 3, acceleration signal model is acceleration of gravity, non-gravitational acceleration, zero-mean white noise
Summation：
Wherein, faIndicate the output of accelerometer, abIndicate non-gravitational acceleration, naIndicate accelerometer zero-mean white noise
Sound,For state transition matrix.
Wherein, g indicates local gravitational acceleration, and specific value is subject to local gravitational acceleration.
State transition matrixAndBetween relationship it is as follows
Accelerometer output formula will be substituted by formula above, can obtained：
Wherein,For its antisymmetric matrix.Error quaternion Kalman filtering state side as available from the above equation
Journey：
zk=Hxk+vk
Wherein,
V=ab+na
Wherein, z indicates system quantities measurement, and H indicates that system measurements matrix, v indicate system measurements noise.
Due to abAnd naIt is irrelevant, so measurement noise matrix R is：
R=E (vvT)=E (abab T+nana T)=Ra+Racc
Wherein, RaFor accelerometer measures noise covariance matrix, RaccIt is the association side of mobile robot non-gravitational acceleration
Poor matrix.
Further, in step 4, consider that mobile robot whether there is non-gravitational acceleration using following formula：
Wherein, fx、fyAnd fzIndicate three axis output quantity of accelerometer, δgFor constant, it is related that noise is measured with accelerometer.
RaccAdaptive updates：
If above-mentioned formula meets, show mobile robot under static or stable state：
Racc=0
If formula is unsatisfactory for, show that there are non-gravitational accelerations for movement：
Racc≈max(α||ra| |, 0)
Wherein, α is constant, raFor residual information, | | ra| | indicate raMould.
ra=z-Hx
Further, in step 5, using the non-gravitational acceleration covariance matrix of adaptive updates to measurement noise matrix
It compensates, and obtains error quaternion updated value using error quaternion Kalman filter, true posture is calculated
Quaternary number carries out attitude algorithm according to Quaternion Method and obtains attitude angle.
Step1：Initialization, initialization error quaternary numerical value, the biased error of gyroscope, gyroscope zero-mean white noise,
Gyroscopic drift zero-mean white noise, accelerometer zero-mean white noise；
Step2：According to Kalman filtering algorithm process, state update is carried out；
Wherein, xk|k-1For estimated using the maximum likelihood estimation state of etching system when k-1 k when etching system state
Amount,For the optimal estimation value of k-1 moment quantity of state, Pk-1ForCovariance, Pk|k-1For xk|k-1Covariance, QdTo be
System process noise matrix；
Step3：Judge that mobile robot whether there is non-gravitational acceleration according to accelerometer data, and to RaccIt carries out
Adaptive updates are compensated using the non-gravitational acceleration covariance matrix of adaptive updates to noise matrix is measured；
Step4：According to Kalman filtering algorithm process, measurement update is carried out；
Wherein, K is Kalman filtering matrix；
Step5：By above step, error quaternion updated value is obtained, true attitude quaternion is calculated, according to four
First number method carries out attitude algorithm and obtains attitude angle；
ByObtain updated true quaternary numerical value.
Wherein, q0、q1、q2And q3Indicate four parameters of quaternary number.
It can be obtained by posture changing matrix in the step 2 of formula combination above：
It in order to more clearly explain the embodiment of the invention or the technical proposal in the existing technology, below will be to institute in embodiment
Attached drawing to be used is needed to be briefly described.
Detailed description of the invention
The improved error quaternion Kalman filtering structures block diagram of Fig. 1
Fig. 2 mobile robot platform non-gravitational acceleration
Fig. 3 mobile robot platform pitch angle estimated value
Specific embodiment
The embodiment of the present invention is described below in detail, the present embodiment is exemplary, for explaining only the invention, and cannot
It is interpreted as limitation of the present invention.Referring to Figure of description to a modification of the present invention error quaternion Kalman filtering machine
People's attitude algorithm method carries out explained in detail below：
As shown in the improved error quaternion Kalman filtering structures block diagram of Fig. 1.
Nonlinear exponent model is used first, is established in conjunction with Levenberg-Marquardt iterative least squares approximating method
Gyro error mathematical model, and the error quaternion differential equation is established using gyroscope output data.Then according to error four
First fractional differentiation equation and gyro error mathematical model establish error quaternion Kalman filtering state equation.According to accelerometer
Relationship between output data and error quaternion establishes error quaternion Kalman filter observational equation.According to non-gravity plus
Speed determines that formula judges whether there is non-gravitational acceleration, and adaptive equalization non-gravitational acceleration is to attitude algorithm bring
It influences.Data fusion is carried out to sensing data using error quaternion Kalman filter, and then obtains error quaternion more
New value, is calculated true attitude quaternion, then carries out attitude algorithm and obtains attitude angle.
1, the corresponding reference frame of mobile robot is chosen, determines posture changing matrix, with nonlinear exponent model, knot
It closes Levenberg-Marquardt iterative least squares approximating method and establishes gyroscopic drift mathematical model, it is defeated using gyroscope
The relationship between data and error quaternion establishes the error quaternion differential equation out；
Northeast day coordinate system is chosen as navigational coordinate system, following appearance can be used in the transformation relation between navigation system and carrier system
State transformation matrixIt indicates：
Wherein, θ is pitch angle, and γ is roll angle, and ψ is course angle.
Output from Gyroscope is unrelated with moving condition, but gyroscope signal has biased error.Therefore gyroscope is surveyed
Amount model is angular speed, biased error, the summation of white noise：
ωo=ω+εg+ng
Wherein, ωoIndicate the output of gyroscope, ω indicates actual angular speed, εgIndicate the drift of gyroscope, ngIndicate top
Spiral shell instrument zero-mean white noise.
With nonlinear exponent model, gyro is established in conjunction with Levenberg-Marquardt iterative least squares approximating method
Instrument drift mathematical model；
Wherein, C1、C2, T be model of fit three parameters, TsFor the sampling time.
Pass through experiment, gyroscope just bias error model fit parameter values such as Table I.
Table.I gyro sensor error model parameters
Due to ωo≠ ω, so the quaternary number solved by gyroscope output valve is not true quaternary number, we claim
The estimated value for quaternary number, be denoted asIt is to have small error between the estimated value of quaternary number and true quaternary number, him
Between error be known as error quaternion, be denoted as
Quaternary number estimated value can be obtained by following formula：
DefinitionFor following formula：
In above-mentioned formula, definition：
Wherein, qe1、qe2、qe3For three parameters of error quaternion.
Relationship between error quaternion and quaternary number estimated value and true quaternary number is as follows：
Wherein, Q is true quaternary number.
To formula derivation above and abbreviation can obtain：
Wherein, [ωo×] indicate its antisymmetric matrix.
2, error quaternion Kalman filtering is established according to the error quaternion differential equation and gyroscopic drift mathematical model
State equation；
Selecting system quantity of state is x=[qe1 qe2 qe3 εgx εgy εgz]T。
Wherein, εgx、εgy、εgzIndicate three axis component of gyroscopic drift.
Following error quaternion Kalman can be established according to the error quaternion differential equation and gyroscopic drift mathematical model
Filter state equation：
xk=Axk-1+Buk-1+wk
Wherein, A indicates that systematic procedure matrix, B indicate that system controls matrix coefficient, and u indicates that system controls matrix, and w is indicated
System mode noise, subscript indicate current time value.
B=I
Wherein, I indicates unit matrix.
3, error quaternion Kalman filtering is established according to the relationship between accelerometer output data and error quaternion
Observational equation；
Acceleration signal model is the summation of acceleration of gravity, non-gravitational acceleration, zero-mean white noise：
Wherein, faIndicate the output of accelerometer, abIndicate non-gravitational acceleration, naIndicate accelerometer zero-mean white noise
Sound,For state transition matrix.
Wherein, g indicates local gravitational acceleration, and specific value is subject to local gravitational acceleration.
State transition matrixAndBetween relationship it is as follows
Accelerometer output formula will be substituted by formula above, can obtained：
Wherein,For its antisymmetric matrix.
Error quaternion Kalman filtering state equation as available from the above equation：
zk=Hxk+vk
Wherein
V=ab+na
Wherein, z indicates system quantities measurement, and H indicates that system measurements matrix, v indicate system measurements noise.
Due to abAnd naIt is irrelevant, so measurement noise matrix R is：
R=E (vvT)=E (abab T+nana T)=Ra+Racc
Wherein, RaFor accelerometer measures noise covariance matrix, RaccIt is the association side of mobile robot non-gravitational acceleration
Poor matrix.
4, judge that mobile robot whether there is non-gravitational acceleration, and adaptive updates according to accelerometer output data
The covariance matrix of mobile robot non-gravitational acceleration；
Consider that mobile robot whether there is non-gravitational acceleration using following formula：
Wherein, fx、fyAnd fzIndicate three axis output quantity of accelerometer, δgIt is related with accelerometer measurement noise for constant,
δ in this examplegIt is 0.015.If formula is unsatisfactory for, show that there are non-gravitational accelerations for movement.
RaccAdaptive updates：
If above-mentioned formula meets, show mobile robot under static or stable state：
Racc=0
If formula is unsatisfactory for, show that there are non-gravitational accelerations for movement：
Racc≈max(α||ra| |, 0)
Wherein, α is constant, and α is 0.2, r in this exampleaFor residual information, | | ra| | indicate raMould.
ra=z-Hx
5, it is compensated, and utilized to noise matrix is measured using the non-gravitational acceleration covariance matrix of adaptive updates
Error quaternion Kalman filter obtains error quaternion updated value, true attitude quaternion is calculated, according to quaternary
Number method carries out attitude algorithm and obtains attitude angle.
Step1：Initialization, initialization error quaternary numerical value, the biased error of gyroscope, gyroscope zero-mean white noise,
Accelerometer zero-mean white noise；
x0=[0 0000 0]T
ng=[0.01 0.01 0.01]T
na=[0.0001 0.0001 0.0001]T
Step2：According to Kalman filtering algorithm process, state update is carried out；
Wherein, xk|k-1For estimated using the maximum likelihood estimation state of etching system when k-1 k when etching system state
Amount,For the optimal estimation value of k-1 moment quantity of state, Pk-1ForCovariance, Pk|k-1For xk|k-1Covariance, QdTo be
System process noise matrix；
Step3：Judge that mobile robot whether there is non-gravitational acceleration according to accelerometer data, and to RaccIt carries out
Adaptive updates are compensated using the non-gravitational acceleration covariance matrix of adaptive updates to noise matrix is measured；
Step4：According to Kalman filtering algorithm process, measurement update is carried out；
Wherein, K is Kalman filtering matrix.
Step5：By above step, error quaternion updated value is obtained, true attitude quaternion is calculated, according to four
First number method carries out attitude algorithm and obtains true attitude angle；
ByObtain updated true quaternary numerical value.
Wherein, q0、q1、q2And q3Indicate four parameters of quaternary number.
It can be obtained by posture changing matrix in the step 2 of formula combination above：
In conclusion in Fig. 2, blue line indicates non-gravitational acceleration, red line δ to this method progress effect analysisg, from
Fig. 2 can be seen that 0s to 30s mobile robot platform with non-gravitational acceleration, and there is no non-heavy in experiment from 30s to 60s
Power acceleration.In Fig. 3, red line indicates to improve error quaternion Kalman filtered results, and blue line expression does not improve error quaternion
Kalman filtering, from figure 3, it can be seen that two kinds of algorithms show different estimated results in pitching angular estimation from 0s to 30s, from
Two kinds of algorithms of 30s to 60s show identical estimated result in pitching angular estimation.The present invention can be realized full attitude algorithm,
And it can obviously inhibit the influence of noise, drift error and non-gravitational acceleration to attitude algorithm.
Claims (3)
1. a kind of improvement error quaternion Kalman filtering robot pose calculation method, feature include the following steps：
Step 1：The corresponding reference frame of mobile robot is chosen, determines posture changing matrix, with nonlinear exponent model, knot
It closes Levenberg-Marquardt iterative least squares approximating method and establishes gyroscopic drift mathematical model, it is defeated using gyroscope
The relationship between data and error quaternion establishes the error quaternion differential equation out；
Step 2：Error quaternion Kalman filtering is established according to the error quaternion differential equation and gyroscopic drift mathematical model
State equation；
Step 3：Error quaternion Kalman filtering is established according to the relationship between accelerometer output data and error quaternion
Observational equation；
Step 4：Judge that mobile robot whether there is non-gravitational acceleration, and adaptive updates according to accelerometer output data
The covariance matrix of mobile robot non-gravitational acceleration；
Step 5：It is compensated using the non-gravitational acceleration covariance matrix of adaptive update to noise matrix is measured, and utilizes mistake
Poor quaternary number Kalman filter obtains error quaternion updated value, true attitude quaternion is calculated, according to quaternary number
Method carries out attitude algorithm and obtains attitude angle.
2. a kind of improvement error quaternion Kalman filtering robot pose calculation method according to claim 1, special
Sign is described in step 4, judges that mobile robot whether there is non-gravitational acceleration according to accelerometer output data, and adaptive
The covariance matrix of mobile robot non-gravitational acceleration should be updated, its step are as follows：
Consider that mobile robot whether there is non-gravitational acceleration using following formula：
Wherein, fx、fyAnd fzIndicate three axis output quantity of accelerometer, δgFor constant, it is related that noise is measured with accelerometer；
The covariance matrix of mobile robot non-gravitational acceleration is Racc, RaccAdaptive updates are as follows：
If above-mentioned formula meets, show mobile robot under static or stable state：
Racc=0
If formula is unsatisfactory for, show that there are non-gravitational accelerations for movement：
Racc≈max(α||ra| |, 0)
Wherein, α is constant, raFor residual information, | | ra| | indicate raMould；
ra=z-Hx
Wherein, z indicates system quantities measurement, and H indicates that system measurements matrix, x indicate system state amount.
3. a kind of improved error quaternion Kalman filtering robot pose calculation method according to claim 1,
It is characterized in that described in step 5, is mended using the non-gravitational acceleration covariance matrix of adaptive updates to noise matrix is measured
It repays, its step are as follows：
Covariance matrix using updated mobile robot non-gravitational acceleration is Racc, mended to noise matrix R is measured
It repays；
R=E (vvT)=E (abab T+nana T)=Ra+Racc
Wherein, v indicates system measurements noise, abIndicate non-gravitational acceleration, naIndicate accelerometer zero-mean white noise, RaFor
Accelerometer measures noise covariance matrix；Utilize the compensated error quaternion Kalman for measuring noise matrix R and foundation
Algorithm filter obtains error quaternion updated value, and true attitude quaternion is calculated, and carries out posture according to Quaternion Method
Resolving obtains attitude angle, reduces influence of the non-gravitational acceleration to attitude algorithm.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810346303.7A CN108827301A (en) | 2018-04-16 | 2018-04-16 | A kind of improvement error quaternion Kalman filtering robot pose calculation method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810346303.7A CN108827301A (en) | 2018-04-16 | 2018-04-16 | A kind of improvement error quaternion Kalman filtering robot pose calculation method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN108827301A true CN108827301A (en) | 2018-11-16 |
Family
ID=64154410
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810346303.7A Pending CN108827301A (en) | 2018-04-16 | 2018-04-16 | A kind of improvement error quaternion Kalman filtering robot pose calculation method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108827301A (en) |
Cited By (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110108279A (en) * | 2019-06-05 | 2019-08-09 | 河南理工大学 | A kind of shaft tower inclinometric system and inclination calculation method |
| CN110440797A (en) * | 2019-08-28 | 2019-11-12 | 广州小鹏汽车科技有限公司 | Vehicle attitude estimation method and system |
| CN110561424A (en) * | 2019-07-28 | 2019-12-13 | 华南理工大学 | online robot kinematic calibration method based on multi-sensor hybrid filter |
| WO2020124678A1 (en) * | 2018-12-19 | 2020-06-25 | 上海交通大学 | Method and system employing functional iterative integration to solve inertial navigation |
| CN112084458A (en) * | 2020-08-07 | 2020-12-15 | 深圳市瑞立视多媒体科技有限公司 | Rigid body posture tracking method and device, equipment and storage medium thereof |
| CN112114339A (en) * | 2020-11-20 | 2020-12-22 | 四川中科川信科技有限公司 | GNSS data differential iterative filtering resolving method |
| CN113405553A (en) * | 2020-11-30 | 2021-09-17 | 辽宁工业大学 | Unmanned special vehicle attitude measurement method |
| CN113674412A (en) * | 2021-08-12 | 2021-11-19 | 浙江工商大学 | Pose fusion optimization-based indoor map construction method and system and storage medium |
| CN114176576A (en) * | 2021-12-11 | 2022-03-15 | 江苏智恒文化科技有限公司 | Method for identifying human motion state based on acceleration |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101726295A (en) * | 2008-10-24 | 2010-06-09 | 中国科学院自动化研究所 | Unscented Kalman filter-based method for tracking inertial pose according to acceleration compensation |
| KR101456576B1 (en) * | 2013-07-10 | 2014-10-31 | 한국원자력연구원 | Attitude estimation apparatus and method in in-pipe robot |
| CN104316055A (en) * | 2014-09-19 | 2015-01-28 | 南京航空航天大学 | Two-wheel self-balancing robot attitude calculation method based on improved Extended Kalman Filter algorithm |
| JP5747752B2 (en) * | 2011-09-06 | 2015-07-15 | トヨタ自動車株式会社 | Posture estimation device, posture estimation method, posture estimation program |
| CN106052683A (en) * | 2016-05-25 | 2016-10-26 | 速感科技（北京）有限公司 | Robot motion attitude estimating method |
| CN106767797A (en) * | 2017-03-23 | 2017-05-31 | 南京航空航天大学 | A kind of inertia based on dual quaterion/GPS Combinated navigation methods |
| CN107036598A (en) * | 2017-03-30 | 2017-08-11 | 南京航空航天大学 | Dual quaterion inertia/celestial combined navigation method based on gyro error amendment |
| CN107478223A (en) * | 2016-06-08 | 2017-12-15 | 南京理工大学 | A kind of human body attitude calculation method based on quaternary number and Kalman filtering |
-
2018
- 2018-04-16 CN CN201810346303.7A patent/CN108827301A/en active Pending
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101726295A (en) * | 2008-10-24 | 2010-06-09 | 中国科学院自动化研究所 | Unscented Kalman filter-based method for tracking inertial pose according to acceleration compensation |
| JP5747752B2 (en) * | 2011-09-06 | 2015-07-15 | トヨタ自動車株式会社 | Posture estimation device, posture estimation method, posture estimation program |
| KR101456576B1 (en) * | 2013-07-10 | 2014-10-31 | 한국원자력연구원 | Attitude estimation apparatus and method in in-pipe robot |
| CN104316055A (en) * | 2014-09-19 | 2015-01-28 | 南京航空航天大学 | Two-wheel self-balancing robot attitude calculation method based on improved Extended Kalman Filter algorithm |
| CN106052683A (en) * | 2016-05-25 | 2016-10-26 | 速感科技（北京）有限公司 | Robot motion attitude estimating method |
| CN107478223A (en) * | 2016-06-08 | 2017-12-15 | 南京理工大学 | A kind of human body attitude calculation method based on quaternary number and Kalman filtering |
| CN106767797A (en) * | 2017-03-23 | 2017-05-31 | 南京航空航天大学 | A kind of inertia based on dual quaterion/GPS Combinated navigation methods |
| CN107036598A (en) * | 2017-03-30 | 2017-08-11 | 南京航空航天大学 | Dual quaterion inertia/celestial combined navigation method based on gyro error amendment |
Non-Patent Citations (3)
| Title |
|---|
| 叶锃锋等: "基于四元数和卡尔曼滤波的两轮车姿态稳定方法", 《传感技术学报》 * |
| 胡佳佳等: "一种自适应残差补偿算法在移动机器人姿态估计中的应用研究", 《传感技术学报》 * |
| 贾瑞才: "基于四元数EKF的低成本MEMS姿态估计算法", 《传感技术学报》 * |
Cited By (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2020124678A1 (en) * | 2018-12-19 | 2020-06-25 | 上海交通大学 | Method and system employing functional iterative integration to solve inertial navigation |
| CN110108279A (en) * | 2019-06-05 | 2019-08-09 | 河南理工大学 | A kind of shaft tower inclinometric system and inclination calculation method |
| CN110561424A (en) * | 2019-07-28 | 2019-12-13 | 华南理工大学 | online robot kinematic calibration method based on multi-sensor hybrid filter |
| CN110440797A (en) * | 2019-08-28 | 2019-11-12 | 广州小鹏汽车科技有限公司 | Vehicle attitude estimation method and system |
| CN112084458A (en) * | 2020-08-07 | 2020-12-15 | 深圳市瑞立视多媒体科技有限公司 | Rigid body posture tracking method and device, equipment and storage medium thereof |
| CN112114339A (en) * | 2020-11-20 | 2020-12-22 | 四川中科川信科技有限公司 | GNSS data differential iterative filtering resolving method |
| CN113405553A (en) * | 2020-11-30 | 2021-09-17 | 辽宁工业大学 | Unmanned special vehicle attitude measurement method |
| CN113405553B (en) * | 2020-11-30 | 2023-05-26 | 辽宁工业大学 | Unmanned special vehicle attitude measurement method |
| CN113674412A (en) * | 2021-08-12 | 2021-11-19 | 浙江工商大学 | Pose fusion optimization-based indoor map construction method and system and storage medium |
| CN113674412B (en) * | 2021-08-12 | 2023-08-29 | 浙江工商大学 | Pose fusion optimization-based indoor map construction method, system and storage medium |
| CN114176576A (en) * | 2021-12-11 | 2022-03-15 | 江苏智恒文化科技有限公司 | Method for identifying human motion state based on acceleration |
| CN114176576B (en) * | 2021-12-11 | 2024-05-24 | 江苏智恒文化科技有限公司 | Method for identifying human motion state based on acceleration |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108827301A (en) | A kind of improvement error quaternion Kalman filtering robot pose calculation method | |
| CN104736963B (en) | mapping system and method | |
| US7463953B1 (en) | Method for determining a tilt angle of a vehicle | |
| CN106052685B (en) | A kind of posture and course estimation method of two-stage separation fusion | |
| WO2017063388A1 (en) | A method for initial alignment of an inertial navigation apparatus | |
| CN104316055B (en) | A kind of double-wheel self-balancing robot attitude algorithm method based on improved expanded Kalman filtration algorithm | |
| CN106500693B (en) | A kind of AHRS algorithm based on adaptive extended kalman filtering | |
| CN103712622B (en) | The gyroscopic drift estimation compensation rotated based on Inertial Measurement Unit and device | |
| CN108318038A (en) | A kind of quaternary number Gaussian particle filtering pose of mobile robot calculation method | |
| CN108458714B (en) | Euler angle solving method without gravity acceleration in attitude detection system | |
| CN106482746B (en) | Lever arm calibration and compensation method in a kind of accelerometer for hybrid inertial navigation system | |
| CN101290229A (en) | Silicon micro-navigation attitude system inertia/geomagnetism assembled method | |
| CN104459828B (en) | Based on the non-aligned bearing calibration of earth magnetism vector system around method of principal axes | |
| CN109870173A (en) | A kind of track correct method of the submarine pipeline inertial navigation system based on checkpoint | |
| CN110954102B (en) | Magnetometer-assisted inertial navigation system and method for robot positioning | |
| CN105371846B (en) | Attitude of carrier detection method and its system | |
| CN106403952A (en) | Method for measuring combined attitudes of Satcom on the move with low cost | |
| CN106370178B (en) | Attitude measurement method and device of mobile terminal equipment | |
| US8782913B2 (en) | Method of determining heading by turning an inertial device | |
| CN111272158B (en) | Dynamic azimuth angle resolving method of MEMS electronic compass in complex magnetic disturbance scene | |
| US10488432B2 (en) | Systems and methods for compensating for the absence of a sensor measurement in a heading reference system | |
| CN107402022A (en) | A kind of accelerometer calibration method and device of stable head | |
| CN109916398B (en) | Adaptive quaternion particle filtering attitude data fusion method | |
| CN107607112A (en) | Aircraft inexpensive pose measuring apparatus and measuring method | |
| CN110779553A (en) | Calibration method for magnetometer data |
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 | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20181116 |
|
| WD01 | Invention patent application deemed withdrawn after publication |