CN114666100B - Intelligent vehicle network attack security detection system and method - Google Patents
Intelligent vehicle network attack security detection system and method Download PDFInfo
- Publication number
- CN114666100B CN114666100B CN202210198115.0A CN202210198115A CN114666100B CN 114666100 B CN114666100 B CN 114666100B CN 202210198115 A CN202210198115 A CN 202210198115A CN 114666100 B CN114666100 B CN 114666100B
- Authority
- CN
- China
- Prior art keywords
- vehicle
- representing
- network attack
- moment
- state
- 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
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L63/00—Network architectures or network communication protocols for network security
- H04L63/14—Network architectures or network communication protocols for network security for detecting or protecting against malicious traffic
- H04L63/1408—Network architectures or network communication protocols for network security for detecting or protecting against malicious traffic by monitoring network traffic
- H04L63/1416—Event detection, e.g. attack signature detection
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L67/00—Network arrangements or protocols for supporting network services or applications
- H04L67/01—Protocols
- H04L67/12—Protocols specially adapted for proprietary or special-purpose networking environments, e.g. medical networks, sensor networks, networks in vehicles or remote metering networks
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L2209/00—Additional information or applications relating to cryptographic mechanisms or cryptographic arrangements for secret or secure communication H04L9/00
- H04L2209/84—Vehicles
Landscapes
- Engineering & Computer Science (AREA)
- Computer Security & Cryptography (AREA)
- Computing Systems (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Computer Hardware Design (AREA)
- General Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Navigation (AREA)
- Control Of Driving Devices And Active Controlling Of Vehicle (AREA)
Abstract
The invention discloses a system and a method for detecting network attack security of an intelligent vehicle, wherein the method comprises the following steps: establishing a four-degree-of-freedom vehicle model; acquiring position signals Xs and Ys of the vehicle through a combined navigation sensor; according to the established four-degree-of-freedom vehicle model, a state estimator is designed through an extended Kalman filtering algorithm, and position signals Xe and Ye of the vehicle are estimated; and according to the obtained position signal of the vehicle and the position signal of the vehicle estimated by the state estimator, carrying out dynamic threshold detection, and carrying out network attack safety detection and fault tolerance in real time. The invention can carry out warning and fault tolerance to the problems of sensor failure and network attack safety, thereby greatly improving the network safety of the intelligent vehicle; the high-freedom nonlinear vehicle model is adopted, and the nonlinear state estimator is designed, so that the detection accuracy is enhanced.
Description
Technical Field
The invention belongs to the technical field of intelligent traffic systems, and particularly relates to a system and a method for detecting network attack safety of an intelligent vehicle.
Background
With the progress of sensing and communication technologies, the number of vehicle-mounted sensors and communication networks is remarkably increased, so that the intelligence of the vehicle is enhanced; however, these additional sensors and communication networks make vehicle systems vulnerable to network attacks. Malicious software interfering with an intelligent vehicle may cause serious potential safety hazard and may cause traffic accidents, further causing physical danger to users or passengers, and potential network safety vulnerability of a vehicle system is a problem to be solved urgently.
At present, the methods for detecting the attack security of the automatic driving network are less researched, and a part of research is limited to a linear time-invariant system, and although a linear time-invariant system model provides a more accurate model for the intelligent vehicle under the conditions of constant speed and fixed turning radius, the described vehicle kinematic model is inaccurate under the conditions of lane change, turning radius and speed change due to the nonlinearity of vehicle dynamics. On the other hand, observer-based residual detectors detect possible sensor attacks and trigger an alarm if the residual is greater than a threshold. This method is detected by a fixed threshold and may trigger false alarms when transient noise of large amplitude is encountered. The improvement in robustness and accuracy of current model-based approaches remains a challenge for autonomous vehicle network attack detection.
Disclosure of Invention
In view of the above deficiencies of the prior art, the present invention provides a system and a method for detecting network attack of an intelligent vehicle, so as to overcome the problem of the prior art that the robustness and the accuracy of the network attack detection of the intelligent vehicle are not strong. The invention particularly considers the influence of sensor noise and observer error, and can quickly and accurately monitor the network attack safety of the intelligent vehicle through the dynamic threshold.
In order to achieve the purpose, the technical scheme adopted by the invention is as follows:
the invention discloses an intelligent vehicle network attack security detection system, which comprises: the system comprises a vehicle-mounted acceleration sensor, a front wheel turning angle sensor, a yaw rate sensor, a combined navigation sensor, a state estimation module, a network attack detection module and a network attack fault-tolerant module;
the vehicle-mounted acceleration sensor is used for acquiring the transverse acceleration a of the vehicle x Longitudinal acceleration a y ；
The front wheel steering angle sensor is used for acquiring the front wheel steering angle of the vehicle;
the yaw velocity sensor is used for acquiring the yaw velocity of the vehicle;
the combined navigation sensor is used for acquiring position signals Xs and Ys of the vehicle;
the state estimation module is used for receiving signals sent by a vehicle-mounted acceleration sensor, a front wheel steering angle sensor, a yaw rate sensor and a combined navigation sensor, designing a state estimator through an extended Kalman filtering algorithm, and estimating position signals Xe and Ye of a vehicle;
the network attack detection module is used for respectively receiving a vehicle position signal sent by the integrated navigation sensor and a vehicle position signal sent by the state estimation module, carrying out network attack detection according to a preset threshold value and sending a network attack flag bit Fault to the network attack Fault-tolerant module;
the network attack Fault-tolerant module carries out Fault-tolerant processing on the vehicle position information by judging the network attack zone bit Fault value sent by the network attack detection module, and sends a reconstructed signal after the Fault-tolerant processing to the vehicle as a feedback signal.
Further, the integrated navigation sensor can realize switching between vehicle-mounted inertia and GPS integrated navigation and vehicle-mounted inertia navigation.
The invention discloses a network attack safety detection method of an intelligent vehicle, which is based on the system and comprises the following steps:
(1) Establishing a four-degree-of-freedom vehicle model;
(2) Acquiring position signals Xs and Ys of the vehicle through a combined navigation sensor;
(3) According to the established four-degree-of-freedom vehicle model, a state estimator is designed through an extended Kalman filtering algorithm, and position signals Xe and Ye of the vehicle are estimated;
(4) And (3) carrying out dynamic threshold detection according to the position signal of the vehicle obtained in the step (2) and the position signal of the vehicle estimated by the state estimator in the step (3), and carrying out network attack safety detection and fault tolerance in real time.
Further, the step (1) is specifically as follows: considering the longitudinal dynamic state and the vertical load transfer of the vehicle, a four-degree-of-freedom vehicle model is established as follows:
wherein m represents the mass of the vehicle; x and y respectively represent the transverse displacement and the longitudinal displacement of the vehicle motion in a local coordinate system, and the first differential and the second differential of the two respectively represent the transverse displacement and the longitudinal displacement in the local coordinate systemVelocity and acceleration of direction;
represents a yaw angle, the first derivative of which represents a yaw angular velocity; m is s Representing the vehicle spring mass; f Xij ，F Yij Respectively, longitudinal and lateral tire forces for a wheel ij (ij =11 for a front left wheel; ij =12 for a front right wheel; ij =21 for a rear left wheel; ij =22 for a rear right wheel); i is z Representing the moment of inertia of the vehicle about the Z axis; i is x Representing the moment of inertia of the vehicle about the X-axis; h is a total of c Representing the height of the center of mass of the vehicle; e represents the distance from the roll center of the vehicle to the center of mass; phi represents a roll angle; c φ Represents a roll coefficient; k φ Represents roll stiffness; c d Representing a wind resistance coefficient; a. The f Representing the area of the front of the vehicle; ρ represents an air density; l represents the wheelbase; a and b respectively represent the distance from the front axis and the rear axis to the centroid; I.C. A xeq Representing the equivalent moment of inertia of the sprung mass about the X axis; d represents a wheel tread; mu represents the road surface friction coefficient, delta f Indicating a front wheel turning angle;
the conversion relationship between the local coordinate system and the global coordinate system is as follows:
in the formula, X and Y respectively represent the transverse displacement and the longitudinal displacement of the vehicle motion in a global coordinate system, and the first order differential and the second order differential of the X and the Y respectively represent the speed and the acceleration in the transverse and longitudinal directions in the global coordinate system;
the vehicle dynamic vertical load is modeled as:
in the formula, F Zij Represents the dynamic vertical load of the wheel ij; g is the gravity proportionality coefficient.
Further, the step (3) is specifically as follows:
through the steps of(1) Tire force F obtained by middle-four-degree-of-freedom vehicle model Xij ，F Yij And as the input of the state estimator, the longitudinal position and the transverse position of the vehicle are the output of the state estimator, an automatic driving system equation is established by adopting an extended Kalman filtering method, and the system is described as a discrete time-varying system:
in the formula, the state vector X = [ X ] e ,Y e ] T The system input is
Measurement output z = [ X ] s ,Y s ] T ；f cd Expressing a prediction equation; u. of k An input representing time k; w is a k Representing the predicted noise; g cd Representing an observation equation; v. of k Representing observation noise; x is the number of k+1 Represents the state quantity at the time of k + 1; x is the number of k Represents the state quantity at time k, y k+1 Represents an observed value at time k + 1;
the jacobian matrix defining the non-linear prediction equation and the observation equation is:
in the formula, F represents a Jacobian matrix of a prediction equation; g represents a Jacobian matrix of the observation equation;
an estimated value representing the state quantity at the time k-1; />
Is an estimated value of the state quantity at the k moment;
converting a nonlinear system into a linear system:
the update process is as follows:
in the formula, K k+1∣k A state gain matrix representing state estimation at time k to time k + 1; p k+1∣k A prediction error covariance matrix representing state estimation at the k moment to the k +1 moment; p k∣k A prediction error covariance matrix representing state estimation at time k versus time k; f k A Jacobian matrix representing the prediction equation at time k;
a transposed matrix representing a Jacobian matrix of the prediction equation at time k; q represents a covariance matrix of the predicted noise; g k A Jacobian matrix representing the observation equation at the k moment; />
A transpose of a Jacobian matrix representing the observation equation at time k; s k+1∣k An observation error covariance matrix representing state estimation at the k moment to the k +1 moment; r represents a covariance matrix of the observed noise; k k+1∣k A state gain matrix representing state estimation at time k to time k + 1; p k+1∣k+1 A prediction error covariance matrix representing state estimation at the k +1 moment and the k +1 moment; i represents an identity matrix;
in the formula (I), the compound is shown in the specification,
a predicted value representing state estimation at time k; x is the number of k+1∣k The predicted value of state estimation at the k moment to the k +1 moment is represented; y is k+1∣k Representing the observed value of the integrated navigation sensor at the k moment to the k +1 moment;
the estimated value is obtained by:
in the formula (I), the compound is shown in the specification,
the predicted value of state estimation at the moment of k +1 is represented; y is K+1 Representing the measured values of the sensor.
Further, the step (4) specifically includes:
estimating the position signal of the vehicle in real time through the state estimator established in the step (3), simultaneously carrying out network attack detection according to a dynamic threshold detection method, and judging whether the deviation between the position signals Xs and Ys of the vehicle collected by the combined navigation sensor and the vehicle position signals Xe and Ye estimated by the state estimator exceeds a dynamic threshold:
through continuous N times of sampling, at the moment of k + N +1, if the deviation between vehicle position signals Xe and Ye estimated by a state estimator and vehicle position signals Xs and Ys acquired by a combined navigation sensor exceeds a dynamic threshold sigma and a network attack flag bit Fault =0 is set, the combined navigation sensor is considered to be under network attack, the measurement data of the combined navigation sensor is isolated, the estimation of the vehicle position is triggered, and the vehicle position signals Xe and Ye are output and sent to a vehicle as feedback signals X and Y;
and if the deviation between the position signals Xe and Ye estimated by the state estimator and the vehicle position signals Xs and Ys acquired by the combined navigation sensor does not exceed the dynamic threshold sigma and the network attack flag bit Fault =1 is set, the network attack is considered not to occur, and the position signals Xs and Ys of the vehicle are output as feedback signals X and Y to the vehicle.
Further, the method for detecting the dynamic threshold specifically includes:
and simultaneously considering the preset threshold of the error between the estimated value of the state estimator to the vehicle position signal and the measured value of the combined navigation sensor to the vehicle position signal, and establishing the threshold error as follows:
in the formula, epsilon represents the measurement error of the combined navigation sensor at the moment of k + N + 1;
representing the measured value of the combined navigation sensor at the moment k; />
Representing the estimation value of the state estimation module at the k moment; k-k + N represents the detection step length when no network attack occurs; epsilon is adjusted with the state of the vehicle at different time periods;
considering the estimation error η of the state estimator, designing an off-line preset corresponding dynamic threshold as follows:
σ＝||ε±η||
in the formula, σ represents an attack detection threshold value at the time of k + N + 1; η represents the estimation error of the state estimator.
The invention has the beneficial effects that:
1. the invention can warn and fault-tolerant the problems of sensor failure and network attack safety, thus greatly improving the network safety of the intelligent vehicle; the high-freedom nonlinear vehicle model is adopted, and the nonlinear state estimator is designed, so that the detection accuracy is enhanced.
2. On the basis of estimating vehicle positioning, the invention considers sensor noise and measurement noise to prevent false alarm of network attack, and sets a corresponding off-line dynamic threshold value according to the change of vehicle state by adopting a dynamic threshold value method under a specific driving scene, thereby further enhancing the accuracy and stability of network security detection.
Drawings
FIG. 1 is a flow chart of the method of the present invention;
FIG. 2 is a diagram illustrating the network attack detection result under the condition of 36 km/h.
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.
The invention discloses an intelligent vehicle network attack safety detection system, which comprises: the system comprises a vehicle-mounted acceleration sensor, a front wheel turning angle sensor, a yaw rate sensor, a combined navigation sensor, a state estimation module, a network attack detection module and a network attack fault-tolerant module;
the vehicle-mounted acceleration sensor is used for acquiring the transverse acceleration a of the vehicle x Longitudinal acceleration a y ；
The front wheel steering angle sensor is used for acquiring the front wheel steering angle of the vehicle;
the yaw velocity sensor is used for acquiring the yaw velocity of the vehicle;
the integrated navigation sensor is used for acquiring position signals Xs and Ys (the Xs and the Ys respectively represent a transverse position and a longitudinal position) of the vehicle;
the integrated navigation sensor can realize the switching between the vehicle-mounted inertia and GPS integrated navigation and the vehicle-mounted inertia navigation;
the state estimation module is used for receiving signals sent by a vehicle-mounted acceleration sensor, a front wheel steering angle sensor, a yaw rate sensor and a combined navigation sensor, designing a state estimator through an extended Kalman filtering algorithm, and estimating position signals Xe and Ye of a vehicle;
the network attack detection module is used for respectively receiving a vehicle position signal sent by the integrated navigation sensor and a vehicle position signal sent by the state estimation module, carrying out network attack detection according to a preset threshold value and sending a network attack flag bit Fault to the network attack Fault-tolerant module;
the network attack Fault-tolerant module carries out Fault-tolerant processing on the vehicle position information by judging the Fault value of the network attack flag bit sent by the network attack detection module, and sends a reconstructed signal after the Fault-tolerant processing to the vehicle as a feedback signal.
Referring to fig. 1, the intelligent vehicle network attack security detection method of the present invention, based on the above system, includes the following steps:
(1) Establishing a four-degree-of-freedom vehicle model;
wherein, the step (1) is as follows: considering the longitudinal dynamic state and the vertical load transfer of the vehicle, a four-degree-of-freedom vehicle model is established as follows:
wherein m represents the mass of the vehicle; x and y respectively represent the transverse displacement and the longitudinal displacement of the vehicle motion under a local coordinate system, and the first differential and the second differential of the transverse displacement and the longitudinal displacement respectively represent the speed and the acceleration in the transverse and longitudinal directions under the local coordinate system;
represents a yaw angle, the first derivative of which represents a yaw angular velocity; m is s Representing the vehicle spring mass; f Xij ，F Yij Respectively, longitudinal and lateral tire forces for a wheel ij (ij =11 for a front left wheel; ij =12 for a front right wheel; ij =21 for a rear left wheel; ij =22 for a rear right wheel); i is z Representing the moment of inertia of the vehicle about the Z axis; i is x Representing the moment of inertia of the vehicle about the X-axis; h is c Representing the height of the center of mass of the vehicle; e represents the distance from the roll center of the vehicle to the center of mass; phi represents a roll angle; c φ Represents a roll coefficient; k φ Represents roll stiffness; c d Representing a wind resistance coefficient; a. The f Representing the area of the front of the vehicle; ρ represents an air density; l represents the wheelbase; a and b respectively represent the distance from the front and rear axes to the centroid; i is xeq Representing the equivalent moment of inertia of the sprung mass about the X axis; d represents a wheel tread; mu represents the road surface friction coefficient, delta f Indicating a front wheel turning angle;
the conversion relationship between the local coordinate system and the global coordinate system is as follows:
in the formula, X and Y respectively represent the transverse displacement and the longitudinal displacement of the vehicle motion in a global coordinate system, and the first order differential and the second order differential of the X and the Y respectively represent the speed and the acceleration in the transverse and longitudinal directions in the global coordinate system;
the vehicle dynamic vertical load is modeled as:
in the formula, F Zij Represents the dynamic vertical load of the wheel ij; g is the gravity proportionality coefficient.
(2) Acquiring position signals Xs and Ys of the vehicle through a combined navigation sensor;
(3) According to the established four-degree-of-freedom vehicle model, a state estimator is designed through an extended Kalman filtering algorithm, and position signals Xe and Ye of the vehicle are estimated;
wherein the step (3) is specifically as follows:
tire force F obtained through the four-degree-of-freedom vehicle model in step (1) Xij ，F Yij And as the input of the state estimator, the longitudinal position and the transverse position of the vehicle are the output of the state estimator, an automatic driving system equation is established by adopting an extended Kalman filtering method, and the system is described as a discrete time-varying system:
wherein, the state vector X = [ X ] e ,Y e ] T The system input is
Measurement output z = [ X ] s ,Y s ] T ；f cd Expressing a prediction equation; u. of k An input representing time k; w is a k Representing the prediction noise; g cd Representing an observation equation; v. of k Representing observation noise; x is the number of k+1 Represents the state quantity at the time of k +1；x k Represents the state quantity at time k, y k+1 Represents an observed value at time k + 1;
the jacobian matrix defining the non-linear prediction equation and the observation equation is:
in the formula, F represents a Jacobian matrix of a prediction equation; g represents a Jacobian matrix of the observation equation;
an estimated value representing the state quantity at the time k-1; />
Is an estimated value of the state quantity at the k moment;
converting a nonlinear system into a linear system:
the update process is as follows:
in the formula, K k+1∣k A state gain matrix representing state estimation at time k to time k + 1; p k+1∣k A prediction error covariance matrix representing state estimation at the k moment to the k +1 moment; p k∣k A prediction error covariance matrix representing state estimation at time k versus time k; f k A Jacobian matrix representing the prediction equation at time k;
a transposed matrix representing a Jacobian matrix of the prediction equation at time k; q represents a covariance matrix of the predicted noise; g k A Jacobian matrix representing the observation equation at the k moment; />
A transpose of a Jacobian matrix representing the observation equation at time k; s k+1∣k An observation error covariance matrix representing state estimation at the k moment to the k +1 moment; r represents a covariance matrix of the observed noise; k k+1∣k A state gain matrix representing state estimation at time k to time k + 1; p k+1∣k+1 A prediction error covariance matrix representing state estimation at the k +1 moment and at the k +1 moment; i represents an identity matrix;
in the formula (I), the compound is shown in the specification,
a predicted value representing state estimation at time k; x is the number of k+1∣k The predicted value of state estimation at the k moment to the k +1 moment is shown; y is k+1∣k Representing the observation value of the integrated navigation sensor at the k moment to the k +1 moment;
the estimated value is obtained by:
in the formula (I), the compound is shown in the specification,
the predicted value of state estimation at the moment of k +1 is represented; y is K+1 Representing the measured values of the sensor.
(4) Performing dynamic threshold detection according to the position signal of the vehicle obtained in the step (2) and the position signal of the vehicle estimated by the state estimator, and performing network attack safety detection and fault tolerance in real time;
wherein the step (4) specifically comprises:
estimating the position signal of the vehicle in real time through the state estimator established in the step (3), simultaneously carrying out network attack detection according to a dynamic threshold detection method, and judging whether the deviation between the position signals Xs and Ys of the vehicle collected by the combined navigation sensor and the vehicle position signals Xe and Ye estimated by the state estimator exceeds a dynamic threshold:
through continuous N times of sampling, at the moment of k + N +1, if the deviation between vehicle position signals Xe and Ye estimated by a state estimator and vehicle position signals Xs and Ys acquired by a combined navigation sensor exceeds a dynamic threshold sigma and a network attack flag bit Fault =0 is set, the combined navigation sensor is considered to be under network attack, the measurement data of the combined navigation sensor is isolated, the estimation of the vehicle position is triggered, and the vehicle position signals Xe and Ye are output and sent to a vehicle as feedback signals X and Y;
and if the deviation between the position signals Xe and Ye estimated by the state estimator and the vehicle position signals Xs and Ys acquired by the combined navigation sensor does not exceed the dynamic threshold sigma and the network attack flag bit Fault =1 is set, the network attack is considered not to occur, and the position signals Xs and Ys of the vehicle are output as feedback signals X and Y to the vehicle.
Further, the method for detecting the dynamic threshold specifically includes:
and simultaneously considering the preset threshold of the error between the estimated value of the state estimator to the vehicle position signal and the measured value of the combined navigation sensor to the vehicle position signal, and establishing the threshold error as follows:
in the formula, epsilon represents the measurement error of the combined navigation sensor at the moment of k + N + 1;
representing the measured value of the combined navigation sensor at the moment k; />
Representing the estimation value of the state estimation module at the k moment; k-k + N represents the detection step length when no network attack occurs; e follows the state of the vehicle at different time periodsAdjusting the rows;
considering the estimation error eta of the state estimator, designing an offline preset corresponding dynamic threshold as follows:
σ＝||ε±η||
in the formula, σ represents an attack detection threshold value at the time of k + N + 1; η represents the estimation error of the state estimator.
In the example, a high-simulation class A vehicle is selected from Carsim, the vehicle speed is set to be 36km/h, the vehicle runs at a constant speed under the condition of double-lane change, and the combined navigation sensor is subjected to network attack when the transverse distance is 100 m. According to the test, the preset threshold is designed off-line at 0.5 meters in this driving situation.
Referring to fig. 2, the total calculation time of the attack detection records based on the extended kalman filter algorithm from the network attack occurrence time to the detection identification time is 3.52 seconds, so that the network attack can be rapidly and accurately detected, and the network attack detection robustness and accuracy are better. After the network attack is identified, the vehicle state estimation can continue to work, and the proposed method can provide high-precision position information for vehicle control despite the network attack on the vehicle.
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 (5)
1. A security detection method for network attacks of intelligent vehicles is based on a security detection system for network attacks of intelligent vehicles, and the system comprises: the system comprises a vehicle-mounted acceleration sensor, a front wheel turning angle sensor, a yaw rate sensor, a combined navigation sensor, a state estimation module, a network attack detection module and a network attack fault-tolerant module;
the vehicle-mounted acceleration sensor is used for acquiring the transverse acceleration a of the vehicle x Longitudinal acceleration a y ；
The front wheel steering angle sensor is used for acquiring the front wheel steering angle of the vehicle;
the yaw velocity sensor is used for acquiring the yaw velocity of the vehicle;
the combined navigation sensor is used for acquiring position signals Xs and Ys of the vehicle;
the state estimation module is used for receiving signals sent by a vehicle-mounted acceleration sensor, a front wheel steering angle sensor, a yaw rate sensor and a combined navigation sensor, designing a state estimator through an extended Kalman filtering algorithm, and estimating position signals Xe and Ye of a vehicle;
the network attack detection module is used for respectively receiving a vehicle position signal sent by the integrated navigation sensor and a vehicle position signal sent by the state estimation module, carrying out network attack detection according to a preset threshold value and sending a network attack flag bit Fault to the network attack Fault-tolerant module;
the network attack Fault-tolerant module carries out Fault-tolerant processing on the vehicle position information by judging the Fault value of the network attack flag bit sent by the network attack detection module, and sends a reconstructed signal after the Fault-tolerant processing to the vehicle as a feedback signal;
the method is characterized by comprising the following steps:
(1) Establishing a four-degree-of-freedom vehicle model;
(2) Acquiring position signals Xs and Ys of the vehicle through a combined navigation sensor;
(3) According to the established four-degree-of-freedom vehicle model, a state estimator is designed through an extended Kalman filtering algorithm, and position signals Xe and Ye of the vehicle are estimated;
(4) And (3) carrying out dynamic threshold detection according to the position signal of the vehicle obtained in the step (2) and the position signal of the vehicle estimated by the state estimator in the step (3), and carrying out network attack safety detection and fault tolerance in real time.
2. The method for detecting the network attack security of the intelligent vehicle according to claim 1, wherein the step (1) is as follows: considering the longitudinal dynamic state and the vertical load transfer of the vehicle, a four-degree-of-freedom vehicle model is established as follows:
wherein m represents the mass of the vehicle; x and y respectively represent the transverse displacement and the longitudinal displacement of the vehicle motion under a local coordinate system, and the first differential and the second differential of the transverse displacement and the longitudinal displacement respectively represent the speed and the acceleration in the transverse and longitudinal directions under the local coordinate system;
representing the yaw angle, the first derivative of which represents the yaw rate; m is a unit of s Representing the vehicle spring mass; f Xij ，F Yij Represents the longitudinal tire force and the lateral tire force of the wheel ij, respectively; i is z Representing the moment of inertia of the vehicle about the Z axis; i is x Representing the moment of inertia of the vehicle about the X-axis; h is c Representing the height of the center of mass of the vehicle; e represents the distance from the roll center of the vehicle to the center of mass; phi represents a roll angle; c φ Represents a roll coefficient; k is φ Represents roll stiffness; c d Representing a wind resistance coefficient; a. The f Representing the area of the front of the vehicle; ρ represents an air density; l represents the wheelbase; a and b respectively represent the distance from the front axis and the rear axis to the centroid; i is xeq Representing the equivalent moment of inertia of the sprung mass about the X-axis; d represents a wheel tread; mu represents the road surface friction coefficient, delta f Indicating a front wheel turning angle;
the conversion relationship between the local coordinate system and the global coordinate system is as follows:
in the formula, X and Y respectively represent the transverse displacement and the longitudinal displacement of the vehicle motion in a global coordinate system, and the first order differential and the second order differential of the X and the Y respectively represent the speed and the acceleration in the transverse and longitudinal directions in the global coordinate system;
the vehicle dynamic vertical load is modeled as:
in the formula, F Zij Represents the dynamic vertical load of the wheel ij; g is the gravity proportionality coefficient.
3. The intelligent vehicle network attack security detection method according to claim 2, wherein the step (3) is specifically as follows:
tire force F obtained through the four-degree-of-freedom vehicle model in step (1) Xij ，F Yij And as the input of the state estimator, the longitudinal position and the transverse position of the vehicle are the output of the state estimator, an automatic driving system equation is established by adopting an extended Kalman filtering method, and the system is described as a discrete time-varying system:
wherein, the state vector X = [ X ] e ,Y e ] T The system input is
Measurement output z = [ X ] s ,Y s ] T ；f cd Expressing a prediction equation; u. of k An input representing time k; w is a k Representing the prediction noise; g cd Representing an observation equation; v. of k Representing observation noise; x is a radical of a fluorine atom k+1 Represents the state quantity at the time of k + 1; x is a radical of a fluorine atom k Represents the state quantity at time k, y k+1 Represents an observed value at time k + 1;
the jacobian matrix defining the non-linear prediction equation and the observation equation is:
in the formula, F represents a Jacobian matrix of a prediction equation; g represents a Jacobian matrix of the observation equation;
an estimated value representing the state quantity at the time k-1;
is an estimated value of the state quantity at the k moment;
converting a nonlinear system into a linear system:
the update process is as follows:
in the formula, K k+1∣k A state gain matrix representing state estimation at time k to time k + 1; p is k+1∣k A prediction error covariance matrix representing state estimation at the k moment to the k +1 moment; p is k∣k A prediction error covariance matrix representing state estimation at time k versus time k; f k A Jacobian matrix representing the prediction equation at the k moment;
a transposed matrix representing a Jacobian matrix of the prediction equation at time k; q represents a covariance matrix of the predicted noise; g k A Jacobian matrix representing the observation equation at the time k;
a transpose of a Jacobian matrix representing the observation equation at time k; s k+1∣k An observation error covariance matrix representing state estimation at the k moment to the k +1 moment; r represents a covariance matrix of the observed noise; k k+1∣k A state gain matrix representing state estimation at time k to time k + 1; p k+1∣k+1 A prediction error covariance matrix representing state estimation at the k +1 moment and the k +1 moment; i represents an identity matrix;
in the formula (I), the compound is shown in the specification,
a predicted value representing state estimation at time k; x is the number of k+1∣k The predicted value of state estimation at the k moment to the k +1 moment is shown; y is k+1∣k Representing the observed value of the integrated navigation sensor at the k moment to the k +1 moment;
the estimated value is obtained by:
in the formula (I), the compound is shown in the specification,
the predicted value of state estimation at the moment of k +1 is represented; y is K+1 Representing the measured values of the sensor.
4. The intelligent vehicle network attack security detection method according to claim 3, wherein the step (4) specifically comprises:
estimating the position signal of the vehicle in real time through the state estimator established in the step (3), simultaneously carrying out network attack detection according to a dynamic threshold detection method, and judging whether the deviation between the position signals Xs and Ys of the vehicle collected by the combined navigation sensor and the vehicle position signals Xe and Ye estimated by the state estimator exceeds a dynamic threshold:
through continuous N times of sampling, at the moment of k + N +1, if the deviation between vehicle position signals Xe and Ye estimated by a state estimator and vehicle position signals Xs and Ys acquired by a combined navigation sensor exceeds a dynamic threshold sigma and a network attack flag bit Fault =0 is set, the combined navigation sensor is considered to be under network attack, the measurement data of the combined navigation sensor is isolated, the estimation of the vehicle position is triggered, and the vehicle position signals Xe and Ye are output and sent to a vehicle as feedback signals X and Y;
and if the deviation between the position signals Xe and Ye estimated by the state estimator and the vehicle position signals Xs and Ys acquired by the combined navigation sensor does not exceed the dynamic threshold sigma and the network attack flag bit Fault =1 is set, the network attack is considered not to occur, and the position signals Xs and Ys of the vehicle are output as feedback signals X and Y to the vehicle.
5. The intelligent vehicle network attack security detection method according to claim 4, wherein the dynamic threshold detection method specifically comprises the following steps:
and simultaneously considering the preset threshold of the error between the estimated value of the state estimator to the vehicle position signal and the measured value of the combined navigation sensor to the vehicle position signal, and establishing the threshold error as follows:
in the formula, epsilon represents the measurement error of the combined navigation sensor at the moment of k + N + 1;
representing the measured value of the combined navigation sensor at the moment k;
representing the estimation value of the state estimation module at the k moment; k-k + N represents the detection step length when no network attack occurs; epsilon is adjusted with the state of the vehicle at different time periods;
considering the estimation error η of the state estimator, designing an off-line preset corresponding dynamic threshold as follows:
σ＝||ε±η||
in the formula, σ represents an attack detection threshold value at the time of k + N + 1; η represents the estimation error of the state estimator.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210198115.0A CN114666100B (en) | 2022-03-02 | 2022-03-02 | Intelligent vehicle network attack security detection system and method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210198115.0A CN114666100B (en) | 2022-03-02 | 2022-03-02 | Intelligent vehicle network attack security detection system and method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN114666100A CN114666100A (en) | 2022-06-24 |
| CN114666100B true CN114666100B (en) | 2023-03-24 |
Family
ID=82027434
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210198115.0A Active CN114666100B (en) | 2022-03-02 | 2022-03-02 | Intelligent vehicle network attack security detection system and method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114666100B (en) |
Family Cites Families (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| DE102014211180A1 (en) * | 2014-06-11 | 2015-12-17 | Continental Teves Ag & Co. Ohg | Method and system for improved detection and / or compensation of error values |
| CN109361678B (en) * | 2018-11-05 | 2021-10-12 | 浙江工业大学 | False data injection attack detection method for intelligent networked automobile automatic cruise system |
| CN109358344A (en) * | 2018-11-07 | 2019-02-19 | 西安电子科技大学 | A kind of anti-GPS fraud system of unmanned plane based on Multi-source Information Fusion and method |
| JP7272910B2 (en) * | 2019-09-03 | 2023-05-12 | 株式会社ゼンリンデータコム | In-vehicle device, state estimation method and program |
| CN110889862B (en) * | 2019-11-08 | 2023-04-18 | 北京航空航天大学 | Combined measurement method for multi-target tracking in network transmission attack environment |
| CN111796522B (en) * | 2020-07-16 | 2022-06-03 | 上海智驾汽车科技有限公司 | Vehicle state estimation method |
| CN112622862B (en) * | 2020-12-24 | 2021-11-30 | 北京理工大学前沿技术研究院 | Automatic driving automobile brake abnormity/attack on-line monitoring method and system |
| CN113933858A (en) * | 2021-09-28 | 2022-01-14 | 中国科学院深圳先进技术研究院 | Abnormal detection method and device of positioning sensor and terminal equipment |
-
2022
- 2022-03-02 CN CN202210198115.0A patent/CN114666100B/en active Active
Also Published As
| Publication number | Publication date |
|---|---|
| CN114666100A (en) | 2022-06-24 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Glaser et al. | Integrated driver–vehicle–infrastructure road departure warning unit | |
| CN108765942B (en) | Intelligent networking automobile curve danger early warning system and method | |
| Han et al. | Monitoring system design for lateral vehicle motion | |
| Li et al. | Vehicle sideslip angle estimation based on hybrid Kalman filter | |
| Zeeman et al. | Combining speed and acceleration to detect reckless driving in the informal public transport industry | |
| CN103616013B (en) | A kind of rescue obstacles removing car pose estimation method | |
| CN114629698A (en) | Automatic driving network attack detection system and method based on vehicle state estimation | |
| WO2023050586A1 (en) | Abnormality detection method and apparatus for positioning sensor, and terminal device | |
| Wang et al. | An integrated scheme for coefficient estimation of tire–road friction with mass parameter mismatch under complex driving scenarios | |
| Bouton et al. | A rollover indicator based on the prediction of the load transfer in presence of sliding: application to an all terrain vehicle | |
| CN103279675A (en) | Method for estimating tire-road adhesion coefficients and tire slip angles | |
| Lampiri | Sensor anomaly detection and recovery in a nonlinear autonomous ground vehicle model | |
| Doumiati et al. | Unscented Kalman filter for real-time vehicle lateral tire forces and sideslip angle estimation | |
| CN114666100B (en) | Intelligent vehicle network attack security detection system and method | |
| Cao et al. | Adaptive complementary filter-based post-impact control for independently-actuated and differentially-steered autonomous vehicles | |
| Hac et al. | Estimation of vehicle roll angle and side slip for crash sensing | |
| Yoon et al. | Rollover index-based rollover mitigation control system | |
| CN112278072A (en) | Intelligent vehicle steering safety control system and control method | |
| Han et al. | Development and experimental evaluation of an online estimation system for vehicle mass | |
| JP7028223B2 (en) | Self-position estimator | |
| CN113002552A (en) | Rollover discrimination method based on vehicle roll angle | |
| Iijima et al. | Vehicle side slip angle estimation methodology using a drive recorder | |
| KR100844758B1 (en) | Method and apparatus for detecting malfunction of yaw rate sensor using image sensor | |
| Wenzel et al. | Simplified extended Kalman filter for automotive state estimation | |
| KR101733880B1 (en) | Apparatus and method for estimating position of vehicle using nonlinear tire model |
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 |