CN112950812B - Vehicle state fault-tolerant estimation method based on long-time and short-time memory neural network - Google Patents
Vehicle state fault-tolerant estimation method based on long-time and short-time memory neural network Download PDFInfo
- Publication number
- CN112950812B CN112950812B CN202110156245.3A CN202110156245A CN112950812B CN 112950812 B CN112950812 B CN 112950812B CN 202110156245 A CN202110156245 A CN 202110156245A CN 112950812 B CN112950812 B CN 112950812B
- Authority
- CN
- China
- Prior art keywords
- vehicle
- speed
- time
- moment
- representing
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- G—PHYSICS
- G07—CHECKING-DEVICES
- G07C—TIME OR ATTENDANCE REGISTERS; REGISTERING OR INDICATING THE WORKING OF MACHINES; GENERATING RANDOM NUMBERS; VOTING OR LOTTERY APPARATUS; ARRANGEMENTS, SYSTEMS OR APPARATUS FOR CHECKING NOT PROVIDED FOR ELSEWHERE
- G07C5/00—Registering or indicating the working of vehicles
- G07C5/08—Registering or indicating performance data other than driving, working, idle, or waiting time, with or without registering driving, working, idle or waiting time
- G07C5/0808—Diagnosing performance data
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
- G01S19/00—Satellite radio beacon positioning systems; Determining position, velocity or attitude using signals transmitted by such systems
- G01S19/01—Satellite radio beacon positioning systems transmitting time-stamped messages, e.g. GPS [Global Positioning System], GLONASS [Global Orbiting Navigation Satellite System] or GALILEO
- G01S19/13—Receivers
- G01S19/14—Receivers specially adapted for specific applications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/044—Recurrent networks, e.g. Hopfield networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/047—Probabilistic or stochastic networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/048—Activation functions
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Biomedical Technology (AREA)
- Computing Systems (AREA)
- Life Sciences & Earth Sciences (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Artificial Intelligence (AREA)
- General Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Radar, Positioning & Navigation (AREA)
- Remote Sensing (AREA)
- Probability & Statistics with Applications (AREA)
- Computer Networks & Wireless Communication (AREA)
- Navigation (AREA)
- Control Of Driving Devices And Active Controlling Of Vehicle (AREA)
Abstract
The invention discloses a vehicle state fault-tolerant estimation method based on a long-time and short-time memory neural network, which comprises the following steps of: establishing a one-step prediction model at any moment based on a vehicle two-degree-of-freedom dynamic model and a vehicle longitudinal kinematics model; establishing an observation model based on the measurement of a left rear wheel speed sensor, the measurement of a right rear wheel speed sensor and the fault-tolerant overall vehicle absolute speed of the vehicle at any moment; and (4) combining the one-step prediction model and the observation model with a Kalman filtering theory to estimate the longitudinal speed, the transverse speed and the yaw angular speed of the vehicle at any moment. The method of the invention estimates the vehicle state by fusing the vehicle-mounted sensor and the GPS signal through a multi-sensor fusion technology, and solves the reconstructed vehicle absolute speed signal by using the LSTM-RNN when the GPS is absent, thereby solving the problem of estimating system failure caused by the GPS deficiency, ensuring the vehicle state estimation precision and having stronger fault tolerance.
Description
Technical Field
The invention belongs to the technical field of system state estimation, and particularly relates to a vehicle state fault-tolerant estimation method based on a long-time and short-time memory neural network.
Background
In a high-grade autonomous vehicle, real-time recognition of the vehicle's own state is important for vehicle energy conservation, safety, and the like. Key vehicle state variables such as vehicle longitudinal speed, vehicle lateral speed, vehicle yaw rate and the like are the basis of normal operation of vehicle safety systems such as ESP, ABS and lateral stability control of a vehicle, and are important branches of vehicle intelligent research.
Because the single sensor has factors such as measurement noise and inherent deviation, and the vehicle is difficult to accurately measure the self state through the single sensor, the multi-sensor fusion technology is developed to fully utilize the advantages of various sensors and improve the confidence coefficient of vehicle state estimation. With the improvement of the GPS positioning precision, the GPS measurement information is used for being fused with the measurement information of the vehicle-mounted sensor to estimate the vehicle state. However, the GPS module requires the vehicle-end receiver to receive signals of four satellites at the same time to measure effective positioning and navigation information, so that the GPS module often suffers from signal loss during use. When the GPS is absent, the vehicle state estimation method based on the GPS measurement information will not be able to accurately estimate the vehicle state. Therefore, a vehicle state fault-tolerant estimation method in the absence of GPS is needed to ensure the accuracy of vehicle state estimation.
Disclosure of Invention
In view of the above disadvantages of the prior art, an object of the present invention is to provide a long-short time memory neural network-based vehicle state fault-tolerant estimation method, so as to solve the problem of low confidence in estimation of a vehicle state caused by GPS signal loss in the prior art.
In order to achieve the purpose, the technical scheme adopted by the invention is as follows:
the invention discloses a vehicle state fault-tolerant estimation method based on a long-time and short-time memory neural network, which comprises the following steps of:
step 1) establishing a one-step prediction model at any k moment based on a vehicle two-degree-of-freedom dynamic model and a vehicle longitudinal kinematics model;
step 2) establishing an observation model based on the measurement quantity of a left rear wheel speed sensor, the measurement quantity of a right rear wheel speed sensor and the fault-tolerant overall vehicle absolute speed of the vehicle at any k moment;
and 3) combining the one-step prediction model and the observation model with a Kalman filtering theory to estimate the longitudinal speed, the transverse speed and the yaw angular speed of the vehicle at any k moment.
Further, the further prediction model at any k time in step 1) is:
in the formula, systemThe state quantity is: x s (k)＝[v x (k) v y (k) ω z (k)] T The system input amount is
U m (k)＝[δ fm (k) a xm (k)] T K, k +1 denotes a time stamp, T denotes a sampling time, C f Denotes front axle cornering stiffness, C r Representing the cornering stiffness of the rear axle, a representing the distance of the vehicle's centre of mass to the front axle, b representing the distance of the vehicle's centre of mass to the rear axle, m representing the sprung mass of the vehicle, I z Representing the moment of inertia of the vehicle about a vertical axis, X being represented by f (-), and s (k+1),X s (k),U m (k) the mapping relationship between the two is shown in formula (1):
X s (k+1)＝f(X s (k),U m (k)) (2)
the one-step prediction model estimates the vehicle state at the k +1 moment by using the input quantity and the state quantity at the k moment.
Further, the establishing of the observation model in the step 2) is as follows:
in the formula, v xm (k) Represents an estimated measurement value v of the longitudinal speed of the vehicle, which is obtained by calculating the wheel speed signal of the left rear wheel and the wheel speed signal of the right rear wheel of the vehicle at any k moment ym (k) The method represents that the absolute vehicle speed v (k) and the longitudinal vehicle speed estimated and measured value v of the whole vehicle at any time k are estimated through fault tolerance xm (k) Calculating a vehicle transverse speed estimation measurement value; v. of 1 (k) For the measurement of the speed sensor of the left and rear wheels of the vehicle at any time k 2 (k) And v (k) is the measurement quantity of a wheel speed sensor of the right rear wheel of the vehicle at any time k, and v (k) is the fault-tolerant absolute vehicle speed of the whole vehicle.
Further, the solution method of the fault-tolerant whole vehicle absolute vehicle speed v (k) in the step 2) is as follows:
21) when the GPS signal is normally available:
v(k)＝v GPS (k)
in the formula, v GPS (k) The absolute speed of the whole vehicle measured by the GPS is represented;
22) v when the GPS receiver receives less than 4 satellite signals GPS (k) The measured value is not credible, and at the moment, the reconstructed absolute speed of the whole vehicle is solved by adopting a trained Long-Short-Term Memory Neural Network (LSTM-RNN)
For replacing v GPS (k) T is a timestamp, and t is k; the long-time and short-time memory neural network comprises: the input layer, LSTM layer, output layer, arbitrary moment t input layer model is:
in the formula (I), the compound is shown in the specification,
a measured value of the angle of rotation of the front wheels of the vehicle is indicated,
a measure of the longitudinal acceleration of the vehicle is indicated,
representing an estimated measurement of vehicle speed in the longitudinal direction of the vehicle,
the LSTM layer at any time t consists of 50 cell units, each modeled as:
i t ＝σ(W ih h t-1 +W ix x t +b i )
f t ＝σ(W fh h t-1 +W fx x t +b f )
o t ＝σ(W oh h t-1 +W ox x t +b o )
h t ＝o t ⊙tanh(C t )
in the formula, sigma (·) represents a sigmoid activation function; i all right angle t ,
Representing an input threshold layer; f. of t Representing a forgetting threshold layer; o. o t ,h t Representing an output threshold layer; c t Indicating the state of the cell;
b i ,b f ,b o and
W ih ,W ix ,W fh ,W fx ,W oh ,W ox are parameters to be trained for each cell; t and t-1 both represent timestamps; each cell unit has an output h t (ii) a The output layer outputs h of 50 cell units t Added and solved by an activation function softmax (·)
The expression is as follows:
at this moment, the fault-tolerant whole vehicle absolute speed is as follows:
further, the kalman filtering algorithm in step 3) comprises the following steps:
31) and (3) one-step prediction:
one-step prediction of state quantity:
X s (k+1,k)＝f(X s (k),U m (k)) (3)
where (k +1, k) is a time stamp indicating that the system state at the time k +1 is predicted from the state quantity at the time k and the input quantity, and X on the left side of equation (2) s (k +1) is replaced by X s (k +1, k), to obtain formula (3); the initial state of the system is:
X s (0)＝[v x (0) v y (0) ω z (0)] T
one-step prediction of error covariance P (k +1, k):
P(k+1,k)＝A(k+1,k)P(k)A T (k+1,k)+Q(k)
where p (k) represents the k time error covariance matrix, and the initial time error covariance matrix is:
in the formula, p 1 ,p 2 ,p 3 For a constant value, the superscript T represents the transpose of the matrix, a (k +1, k) represents the jacobian matrix, and its expression is:
q (k) is the process error covariance matrix, and:
32) solving for kalman gain K (K + 1):
K(k+1)＝P(k+1,k)H T (k+1)[H(k+1)P(k+1,k)H T (k+1)+R(k+1)] -1
in the formula (I), the compound is shown in the specification,
r 1 is a constant value representing the estimated measured value v of the vehicle longitudinal speed xm (k) Error variance of r 2 Is a constant value and represents the estimated and measured value v of the vehicle transverse speed ym (k) Error variance of (2), superscript -1 Is an indicator of an inverse matrix;
33) and (3) updating the state:
updating the state quantity:
X s (k+1)＝X s (k+1,k)+K(k+1)[Z(k+1)-H(k+1)X s (k+1,k)]
in the formula, Z (k +1) represents a system observation model at the time of k + 1;
updating an error covariance matrix:
P(k+1)＝[I-K(k+1)H(k+1)]P(k+1,k)
in the formula, I represents an identity matrix;
by iterative solution of the steps 31), 32) and 33), the method can be obtained
X s (k+1)＝[v x (k+1) v y (k+1) ω z (k+1)] T This is the estimation of the longitudinal speed, lateral speed and yaw rate of the vehicle at any time k + 1.
The invention has the beneficial effects that:
the method of the invention estimates the vehicle state by fusing the vehicle-mounted sensor and the GPS signal through a multi-sensor fusion technology, and solves the reconstructed vehicle absolute speed signal by using the LSTM-RNN when the GPS is absent, thereby solving the problem of estimating system failure caused by the GPS deficiency, ensuring the vehicle state estimation precision and having stronger fault tolerance.
Drawings
FIG. 1 is a flow chart of the steps of the method of the present invention;
FIG. 2 is a diagram showing the network structure and cell unit structure of LSTN-RNN.
Detailed Description
In order to facilitate understanding of those skilled in the art, the present invention will be further described with reference to the following examples and drawings, which are not intended to limit the present invention.
Referring to fig. 1, the vehicle state fault-tolerant estimation method based on the long-time and short-time memory neural network includes the following steps:
step 1) establishing a one-step prediction model at any k moment based on a vehicle two-degree-of-freedom dynamic model and a vehicle longitudinal kinematics model;
the one-step prediction model at any k time is as follows:
in the formula, the system state quantity is: x s (k)＝[v x (k) v y (k) ω z (k)] T The system input quantity is
U m (k)＝[δ fm (k) a xm (k)] T K, k +1 denotes a time stamp, T denotes a sampling time, C f Denotes front axle cornering stiffness, C r Representing the cornering stiffness of the rear axle, a representing the distance of the vehicle's centre of mass to the front axle, b representing the distance of the vehicle's centre of mass to the rear axle, m representing the sprung mass of the vehicle, I z Representing the moment of inertia of the vehicle about a vertical axis, X being represented by f (-) s (k+1),X s (k),U m (k) The mapping relationship between the two is shown in formula (1):
X s (k+1)＝f(X s (k),U m (k)) (2)
the one-step prediction model estimates the vehicle state at the k +1 moment by using the input quantity and the state quantity at the k moment.
Step 2) establishing an observation model based on the measurement quantity of a left rear wheel speed sensor, the measurement quantity of a right rear wheel speed sensor and the fault-tolerant overall vehicle absolute speed of the vehicle at any k moment;
the observation model established in the step 2) is as follows:
in the formula, v xm (k) When represents an arbitrary kThe estimated measured value of the longitudinal speed of the vehicle is obtained by calculating the wheel speed signal of the left rear wheel of the vehicle and the wheel speed signal of the right rear wheel of the vehicle ym (k) Representing the absolute vehicle speed v (k) and the estimated and measured value v of the longitudinal vehicle speed of the vehicle at any moment k through fault tolerance xm (k) Calculating a vehicle transverse speed estimation measurement value; v. of 1 (k) For the measurement of the speed sensor of the left and rear wheels of the vehicle at any time k 2 (k) And v (k) is the measurement quantity of a wheel speed sensor of the right rear wheel of the vehicle at any time k, and v (k) is the fault-tolerant absolute vehicle speed of the whole vehicle.
The solving method of the fault-tolerant whole vehicle absolute vehicle speed v (k) in the step 2) is as follows:
21) when the GPS signal is normally available:
v(k)＝v GPS (k)
in the formula, v GPS (k) The absolute speed of the whole vehicle measured by the GPS is represented;
22) v when the GPS receiver receives less than 4 satellite signals GPS (k) The measured value is not credible, and at the moment, the reconstructed absolute speed of the whole vehicle is solved by adopting a trained Long-Short-Term Memory Neural Network (LSTM-RNN)
For substituting v GPS (k) T is a timestamp, and t is k; the long-time and short-time memory neural network comprises: input layer, LSTM layer, output layer, as shown in fig. 2; the input layer model at any time t is as follows:
in the formula (I), the compound is shown in the specification,
a measured value of the angle of rotation of the front wheels of the vehicle is indicated,
a measure of the longitudinal acceleration of the vehicle is indicated,
representing a measure of vehicle longitudinal vehicle speed estimate,
the LSTM layer at any time t consists of 50 cell units, each modeled as:
i t ＝σ(W ih h t-1 +W ix x t +b i )
f t ＝σ(W fh h t-1 +W fx x t +b f )
o t ＝σ(W oh h t-1 +W ox x t +b o )
h t ＝o t ⊙tanh(C t )
in the formula, σ (·) represents a sigmoid activation function; i.e. i t ,
Representing an input threshold layer; f. of t Representing a forgetting threshold layer; o. o t ,h t Representing an output threshold layer; c t Indicating the state of the cell;
b i ,b f ,b o and
W ih ,W ix ,W fh ,W fx ,W oh ,W ox are parameters to be trained for each cell; t and t-1 both represent timestamps; each cell unit has oneAn output h t (ii) a The output layer outputs h of 50 cell units t Added and solved by an activation function softmax (·)
The expression is as follows:
at this moment, the fault-tolerant whole vehicle absolute speed is as follows:
step 3) combining the one-step prediction model and the observation model with a Kalman filtering theory to estimate the longitudinal speed, the transverse speed and the yaw angular speed of the vehicle at any k moment;
the Kalman filtering algorithm comprises the following steps:
31) and (3) one-step prediction:
one-step prediction of state quantity:
X s (k+1,k)＝f(X s (k),U m (k)) (3)
where (k +1, k) is a time stamp indicating that the system state at time k +1 is predicted from the state quantity at time k and the input quantity, and X on the left side of equation (2) s (k +1) by X s (k +1, k), to obtain formula (3); the initial state of the system is:
X s (0)＝[v x (0) v y (0) ω z (0)] T
one-step prediction of error covariance P (k +1, k):
P(k+1,k)＝A(k+1,k)P(k)A T (k+1,k)+Q(k)
where p (k) represents the k time error covariance matrix, and the initial time error covariance matrix is:
in the formula, p 1 ,p 2 ,p 3 For a constant value, the superscript T represents the transpose of the matrix, a (k +1, k) represents the jacobian matrix, and its expression is:
q (k) is a process error covariance matrix, and:
32) solving for kalman gain K (K + 1):
K(k+1)＝P(k+1,k)H T (k+1)[H(k+1)P(k+1,k)H T (k+1)+R(k+1)] -1
in the formula (I), the compound is shown in the specification,
r 1 is a constant value and represents an estimated measured value v of the longitudinal speed of the vehicle xm (k) Error variance of r 2 Is a constant value and represents the estimated and measured value v of the vehicle transverse speed ym (k) Error variance of (2), superscript -1 Is an indicator of an inverse matrix;
33) and (3) updating the state:
updating the state quantity:
X s (k+1)＝X s (k+1,k)+K(k+1)[Z(k+1)-H(k+1)X s (k+1,k)]
in the formula, Z (k +1) represents a system observation model at the time of k + 1;
updating an error covariance matrix:
P(k+1)＝[I-K(k+1)H(k+1)]P(k+1,k)
in the formula, I represents an identity matrix.
By iterative solution of the steps 31), 32) and 33), the method can be obtained
X s (k+1)＝[v x (k+1) v y (k+1) ω z (k+1)] T This is kEstimation of the longitudinal speed, lateral speed, and yaw rate of the vehicle at time + 1.
While the invention has been described in terms of its preferred embodiments, it will be understood by those skilled in the art that various changes in form and details may be made therein without departing from the spirit and scope of the invention.
Claims (1)
1. A vehicle state fault-tolerant estimation method based on a long-time and short-time memory neural network is characterized by comprising the following steps:
step 1) establishing a one-step prediction model at any k moment based on a vehicle two-degree-of-freedom dynamic model and a vehicle longitudinal kinematics model;
step 2) establishing an observation model based on the measurement of a left rear wheel speed sensor, the measurement of a right rear wheel speed sensor and the fault-tolerant whole vehicle absolute speed of the vehicle at any k moment;
step 3) combining the one-step prediction model and the observation model with a Kalman filtering theory to estimate the longitudinal speed, the transverse speed and the yaw angular speed of the vehicle at any k moment;
the one-step prediction model at any k time in the step 1) is as follows:
in the formula, the system state quantity is: x s (k)＝[v x (k) v y (k) ω z (k)] T The system input is U m (k)＝[δ fm (k) a xm (k)] T K, k +1 denotes a time stamp, T denotes a sampling time, C f Representing front axle yaw stiffness, C r Representing the lateral deflection stiffness of the rear axle, a representing the distance from the vehicle's center of mass to the front axle, b representing the distance from the vehicle's center of mass to the rear axle, m representing the sprung mass of the vehicle, I z Representing the moment of inertia of the vehicle about a vertical axis, X being represented by f (-), and s (k+1),X s (k),U m (k) mapping relationship betweenThen equation (1) is expressed as:
X s (k+1)＝f(X s (k),U m (k)) (2)
the one-step prediction model estimates the vehicle state at the k +1 moment by using the input quantity and the state quantity at the k moment;
the observation model established in the step 2) is as follows:
in the formula, v xm (k) Represents an estimated measurement value v of the longitudinal speed of the vehicle, which is obtained by calculating the wheel speed signal of the left rear wheel and the wheel speed signal of the right rear wheel of the vehicle at any k moment ym (k) The method represents that the absolute vehicle speed v (k) and the longitudinal vehicle speed estimated and measured value v of the whole vehicle at any time k are estimated through fault tolerance xm (k) Calculating a vehicle transverse speed estimation measurement value; v. of 1 (k) For any k-time measurement of the left and rear wheel speed sensor of the vehicle, v 2 (k) The measured quantity of a wheel speed sensor of the right rear wheel of the vehicle at any time k, and v (k) is the fault-tolerant absolute vehicle speed of the whole vehicle;
the solving method of the fault-tolerant whole vehicle absolute vehicle speed v (k) in the step 2) is as follows:
21) when the GPS signal is normally available:
v(k)＝v GPS (k)
in the formula, v GPS (k) Representing the absolute speed of the whole vehicle measured by a GPS;
22) v when the GPS receiver receives less than 4 satellite signals GPS (k) The measured value is not credible, and the reconstructed absolute speed of the whole vehicle is solved by adopting the trained long-time and short-time memory neural network
For replacing v GPS (k) T is a timestamp, and t is k; the long-time and short-time memory neural network comprises: the input layer, LSTM layer, output layer, arbitrary moment t input layer model is:
in the formula (I), the compound is shown in the specification,
a measured value of the turning angle of the front wheels of the vehicle is indicated,
a measure of the longitudinal acceleration of the vehicle is indicated,
representing a measure of vehicle longitudinal vehicle speed estimate,
the LSTM layer at any time t consists of 50 cell units, each modeled as:
i t ＝σ(W ih h t-1 +W ix x t +b i )
f t ＝σ(W fh h t-1 +W fx x t +b f )
o t ＝σ(W oh h t-1 +W ox x t +b o )
h t ＝o t ⊙tanh(C t )
in the formula, sigma (·) represents a sigmoid activation function;
representing an input threshold layer; f. of t Representing a forgetting threshold layer; o. o t ,h t Representing an output threshold layer; c t Indicating the state of the cell;
and
are parameters to be trained for each cell; t and t-1 both represent timestamps; each cell unit has an output h t (ii) a The output layer outputs h of 50 cell units t Added and solved by an activation function softmax (·)
The expression is as follows:
at this moment, the fault-tolerant whole vehicle absolute speed is as follows:
the Kalman filtering algorithm in the step 3) comprises the following steps:
31) and (3) one-step prediction:
one-step prediction of state quantity:
X s (k+1,k)＝f(X s (k),U m (k)) (3)
wherein (k +1, k) is a time stamp indicating that k +1 is predicted from the state quantity and the input quantity at the time kX on the left side of equation (2) for system status s (k +1) by X s (k +1, k), to obtain formula (3); the initial state of the system is:
X s (0)＝[v x (0) v y (0) ω z (0)] T
one-step prediction of error covariance P (k +1, k):
P(k+1,k)＝A(k+1,k)P(k)A T (k+1,k)+Q(k)
where p (k) represents the k time error covariance matrix, and the initial time error covariance matrix is:
in the formula, p 1 ,p 2 ,p 3 For a constant value, the superscript T represents the transpose of the matrix, a (k +1, k) represents the jacobian matrix, and its expression is:
q (k) is a process error covariance matrix, and:
32) solving for kalman gain K (K + 1):
K(k+1)＝P(k+1,k)H T (k+1)[H(k+1)P(k+1,k)H T (k+1)+R(k+1)] -1
in the formula (I), the compound is shown in the specification,
r 1 is a constant value and represents an estimated measured value v of the longitudinal speed of the vehicle xm (k) Error variance of r 2 Is a constant value and represents the estimated and measured value v of the vehicle transverse speed ym (k) Error variance, superscript of -1 Is an indicator of an inverse matrix;
33) and (3) state updating:
updating the state quantity:
X s (k+1)＝X s (k+1,k)+K(k+1)[Z(k+1)-H(k+1)X s (k+1,k)]
in the formula, Z (k +1) represents a system observation model at the time of k + 1;
updating an error covariance matrix:
P(k+1)＝[I-K(k+1)H(k+1)]P(k+1,k)
in the formula, I represents an identity matrix;
by iterative solution of steps 31), 32) and 33), X can be obtained s (k+1)＝[v x (k+1) v y (k+1) ω z (k+1)] T This is the estimation of the longitudinal speed, lateral speed and yaw rate of the vehicle at any time k + 1.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110156245.3A CN112950812B (en) | 2021-02-04 | 2021-02-04 | Vehicle state fault-tolerant estimation method based on long-time and short-time memory neural network |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110156245.3A CN112950812B (en) | 2021-02-04 | 2021-02-04 | Vehicle state fault-tolerant estimation method based on long-time and short-time memory neural network |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112950812A CN112950812A (en) | 2021-06-11 |
| CN112950812B true CN112950812B (en) | 2022-07-26 |
Family
ID=76244044
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110156245.3A Active CN112950812B (en) | 2021-02-04 | 2021-02-04 | Vehicle state fault-tolerant estimation method based on long-time and short-time memory neural network |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112950812B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113311714A (en) * | 2021-07-28 | 2021-08-27 | 中国科学院自动化研究所 | Fault diagnosis and fault-tolerant control method and system for multi-joint bionic robot fish sensor |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108715166A (en) * | 2018-04-28 | 2018-10-30 | 南京航空航天大学 | Intact stability index method of estimation based on deep learning |
| CN110516311A (en) * | 2019-07-31 | 2019-11-29 | 江苏大学 | A kind of comprehensive compensation construction of strategy method for automobile-used acceleration transducer constant error |
| CN110562263A (en) * | 2019-09-19 | 2019-12-13 | 中国人民解放军陆军装甲兵学院 | Wheel hub motor driven vehicle speed estimation method based on multi-model fusion |
| CN111497866A (en) * | 2020-04-01 | 2020-08-07 | 南京航空航天大学 | Steering wheel corner sensor fault diagnosis method based on improved Mahalanobis distance |
| CN111693044A (en) * | 2020-06-19 | 2020-09-22 | 南京晓庄学院 | Fusion positioning method |
-
2021
- 2021-02-04 CN CN202110156245.3A patent/CN112950812B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108715166A (en) * | 2018-04-28 | 2018-10-30 | 南京航空航天大学 | Intact stability index method of estimation based on deep learning |
| CN110516311A (en) * | 2019-07-31 | 2019-11-29 | 江苏大学 | A kind of comprehensive compensation construction of strategy method for automobile-used acceleration transducer constant error |
| CN110562263A (en) * | 2019-09-19 | 2019-12-13 | 中国人民解放军陆军装甲兵学院 | Wheel hub motor driven vehicle speed estimation method based on multi-model fusion |
| CN111497866A (en) * | 2020-04-01 | 2020-08-07 | 南京航空航天大学 | Steering wheel corner sensor fault diagnosis method based on improved Mahalanobis distance |
| CN111693044A (en) * | 2020-06-19 | 2020-09-22 | 南京晓庄学院 | Fusion positioning method |
Non-Patent Citations (1)
| Title |
|---|
| A Reliable Vehicle Lateral Velocity Estimation Methodology Based on SBI-LSTM During GPS-Outage;bo zhang，wanzhong zhao，songchun zou，han zhang，zhongkai luan;《IEEE sensors hournal》;20200907;全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN112950812A (en) | 2021-06-11 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109466558B (en) | Road adhesion coefficient estimation method based on EKF (extended Kalman Filter) and BP (Back propagation) neural network | |
| Zhang et al. | Fault-tolerant control for intelligent electrified vehicles against front wheel steering angle sensor faults during trajectory tracking | |
| Marco et al. | Multi-modal sensor fusion for highly accurate vehicle motion state estimation | |
| CN110562263A (en) | Wheel hub motor driven vehicle speed estimation method based on multi-model fusion | |
| US20070027586A1 (en) | Online estimation of vehicle side-slip under linear operating region | |
| CN111238471B (en) | Sideslip angle estimation method and estimator suitable for agricultural machine linear navigation | |
| CN113970330B (en) | Vehicle-mounted multi-sensor fusion positioning method, computer equipment and storage medium | |
| CN112950812B (en) | Vehicle state fault-tolerant estimation method based on long-time and short-time memory neural network | |
| CN112099378B (en) | Front vehicle lateral motion state real-time estimation method considering random measurement time lag | |
| CN103279675A (en) | Method for estimating tire-road adhesion coefficients and tire slip angles | |
| CN113830088A (en) | Intelligent semi-trailer tractor trajectory tracking prediction control method and vehicle | |
| Hsiao et al. | Robust estimation of the friction forces generated by each tire of a vehicle | |
| CN111189454A (en) | Unmanned vehicle SLAM navigation method based on rank Kalman filtering | |
| CN113682282A (en) | Vehicle stability control method and system, vehicle and storage medium | |
| CN111942399A (en) | Vehicle speed estimation method and system based on unscented Kalman filtering | |
| CN112270039A (en) | Distributed asynchronous fusion-based nonlinear state estimation method for drive-by-wire chassis vehicle | |
| Hrgetic et al. | Adaptive EKF-based estimator of sideslip angle using fusion of inertial sensors and GPS | |
| CN113008229B (en) | Distributed autonomous integrated navigation method based on low-cost vehicle-mounted sensor | |
| CN113771857A (en) | Longitudinal speed estimation method and system for vehicle control | |
| CN111231976B (en) | Vehicle state estimation method based on variable step length | |
| Hac et al. | Estimation of vehicle roll angle and side slip for crash sensing | |
| CN113978476B (en) | Wire-controlled automobile tire lateral force estimation method considering sensor data loss | |
| CN116749982A (en) | Engineering vehicle road surface adhesion coefficient state estimation method based on improved double-layer Kalman filtering | |
| CN113060143B (en) | System and method for determining road adhesion coefficient | |
| CN113341997B (en) | Transverse control method and system based on multi-state parameter collaborative estimation |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |