GNSS / INS (Global Navigation Satellite System / Inertial Navigation System) vehicle-mounted fault-tolerant integrated navigation method and system based on isolated forest fault detection, medium and product

Abstract

The invention discloses a GNSS / INS (Global Navigation Satellite System / Inertial Navigation System) vehicle-mounted fault-tolerant integrated navigation method and system based on isolated forest fault detection, a medium and a product, belonging to the field of vehicle-mounted integrated navigation and aiming at improving the navigation positioning precision and reliability of an automatic driving vehicle in an urban environment. The purpose of the invention is realized through the following technical method, and the method comprises the following steps: collecting navigation data and carrying out inertial navigation calculation; an isolated forest model is constructed based on historical navigation incremental data, and anomaly detection of GNSS measurement is realized; calculating an abnormal score of the isolated forest to recognize a GNSS measurement fault; a switchable factor of GNSS measurement is adjusted in a self-adaptive mode; and solving an optimal navigation state based on the optimized factor graph to realize fault-tolerant processing of GNSS abnormal measurement. The method effectively reduces the influence of signal interference such as multi-path effect and non-line-of-sight receiving on the navigation system, enhances the stability and robustness of the system, and is suitable for complex urban application scenes such as automatic driving and intelligent transportation.

Description

Technical Field

[0001] The present invention belongs to the field of vehicle integrated navigation, and particularly relates to a GNSS / INS vehicle fault-tolerant integrated navigation method, system, medium and product based on isolated forest fault detection. Background Art

[0002] In fields involving life safety such as intelligent transportation and vehicle driving, in order to ensure the safety and reliability of vehicles, vehicle integrated navigation systems must provide continuous and high-precision navigation solutions in complex dynamic environments. The integrated navigation technology combining the Global Navigation Satellite System (GNSS) and the Inertial Navigation System (INS) has been widely applied to vehicle navigation systems due to its complementarity. However, in urban environments, GNSS signals are vulnerable to the influence of complex scenarios such as high-rise buildings, glass curtain walls, tunnels and viaducts, resulting in multipath effects and non-line-of-sight reception problems, thereby reducing the GNSS positioning accuracy and affecting the overall performance of the integrated navigation system. In addition, GNSS receivers may also be affected by interference signals (such as electromagnetic interference, spoofing attacks, etc.), which may lead to positioning anomalies. Since GNSS faults usually have uncertainty and randomness, it is difficult to detect and identify them.

[0003] In order to improve the positioning accuracy of the integrated navigation system, researchers usually use the measurement innovation in Kalman filtering to construct a fault detection statistic to identify GNSS measurement anomalies and perform isolation processing. For example, in the patent document with the patent application number 202211388034.3 and the title "A Fault Detection and Identification Method for an Integrated Navigation System", the filtering innovation is obtained through the observable quantities of visible stars and INS observable quantities, and the satellite system fault is detected and judged by using the innovation residual. The standardized innovation proportional to the corresponding satellite fault probability is used to determine the faulty satellite and remove it from the system. However, in the actual navigation environment, the statistical characteristics of GNSS measurement noise deviate from the Gaussian white noise assumption induced by complex and dynamically changing external factors, resulting in a decrease in fault detection accuracy. Another example is in the patent document with the patent application number 202111031783.6 and the title "A Method for Detecting Slow-Varying Faults of Satellites in GNSS / INS Integrated Navigation", an equivalent weight matrix is constructed by using the observation data matrix, and robust processing is performed on the abnormal measurements based on the standard t-distribution. The joint autonomous integrity detection extrapolation method and the residual chi-square statistic are used to improve the detection sensitivity to slow-varying faults. However, this method is affected by the setting of the sliding window length in the extrapolation method. If the sliding window length is too large, the probability of missing GNSS faults will increase, and then the fault detection benchmark will be contaminated by GNSS outliers. Summary of the Invention

[0004] The object of the present invention is a GNSS / INS vehicle-mounted fault-tolerant integrated navigation method, system, medium and product based on isolated forest fault detection, which can improve the navigation accuracy of a carrier in a complex urban environment and provide accurate and continuous navigation information for a vehicle.

[0005] The object of the present invention is achieved by the following technical solutions:

[0006] A GNSS / INS vehicle-mounted fault-tolerant integrated navigation method based on isolated forest fault detection includes the following steps:

[0007] Step 1: Install the initial navigation parameters of GNSS / INS, initialize the factor graph optimization parameters, and then enter the navigation working mode;

[0008] Step 2: Collect the navigation data under normal working conditions, calculate the navigation increment data within the unit time interval of INS, including attitude increment speed increment and position increment Calculate the position increment data within the unit time interval of GNSS and store them in sequence;

[0009] Step 3: Compose all the navigation increment data samples [x1,…,x k T from the initial moment to the k-th moment into a training data set D, randomly select M samples from it to form a sample set χ as the input of the isolated forest model, and at the same time set the limited depth l of the isolated tree, randomly divide the data, and recursively generate isolated trees to form an isolated forest for fault detection;

[0010] Step 4: Collect the increment data x k+1 at the (k + 1)-th moment in real time, calculate the anomaly score S k+1 , S k+1 as the input for fault detection, and perform isolated forest fault detection;

[0011] Step 5: Judge whether the current GNSS measurement is faulty according to the anomaly score S k+1 . If the anomaly score of the GNSS measurement is lower than the threshold, the GNSS is fault-free and the switching factor is 1; if it exceeds the threshold, evaluate the measurement quality and calculate the switchable factor ρ k+1 ;

[0012] Step 6: Calculate the INS pre-integration factor and the GNSS factor and perform a fault-tolerant factor graph optimization algorithm to output the optimal estimated navigation information, realizing the fault tolerance processing of the GNSS abnormal measurement; at the same time, perform a closed-loop feedback correction on the navigation state of the integrated navigation system.

[0013] ​Further, in step 1, the heading information of GNSS is used as the initial heading ψ0 of the carrier, where 0 represents the initial moment; given that the vehicle is in a stationary state on the road surface, the initial pitch angle θ0 and roll angle γ0 are set to zero; the initial velocity information E, 0v N,0 v U,0 T is also set to zero, where ν E,0 , ν N,0 and ν U,0 respectively represent the initial eastward, northward, and upward velocities of the vehicle; the initial position information [L0λ0h0] T is provided by GNSS, where L0, λ0, and h0 respectively represent the initial latitude, longitude, and altitude of the vehicle.

[0014] Further, step 2 is specifically as follows:

[0015] Step 2.1: The normal operating condition navigation data collected includes: gyroscope information ω b , accelerometer information f b , and the latitude, longitude, and altitude information [L GNSS λ GNSS h GNSS T , where ω b is the angular velocity information of the gyroscope in the body coordinate system b, and f b is the acceleration information of the accelerometer in the body coordinate system b;

[0016] Step 2.2: Based on the gyroscope information ω b and the accelerometer information f b , inertial navigation solution is performed. The attitude , velocity , and position at the (k - 1)-th moment are integrated to obtain the attitude , velocity , and position of INS at the k-th moment. The navigation information of INS before and after the moment is subtracted to calculate the attitude increment , velocity increment , and position increment

[0017]

[0018] Step 2.3: The measured position information of the k-th moment collected by GNSS is differenced from the position information of GNSS at the (k - 1)-th moment to obtain the position increment ​​

[0019]

[0020] Step 2.4: Record the GNSS and INS incremental information samples collected by the system at the k-th moment and save them as the k-th navigation incremental data sample.

[0021] Furthermore, in step 3, a dimension μ is randomly specified as the starting node, and a value is randomly selected between the maximum and minimum values of this dimension. Samples smaller than this value are assigned to the left subtree, and samples greater than or equal to this value are assigned to the right subtree; the above steps are recursively executed to continuously construct new isolation trees until there is only one sample in the isolation tree or the sub-nodes have reached the specified depth l.

[0022] Furthermore, the fault detection of GNSS measurements in step 4 based on the isolation forest is specifically as follows: Real-time collect the navigation incremental data x at the (k + 1)-th moment k+1 , and calculate the anomaly score S k+1 in the isolation forest:

[0023]

[0024] where L(x k+1 ) is the path length from the root node to the external node of the sample x k+1 in the isolation tree, E(L(x k+1 )) is the path expectation value of the sample x k+1 in each isolation tree in the isolation forest, and C(M) represents the average path length of an isolation tree constructed with M samples;

[0025] where L(x k+1 ) is:

[0026] L(x k+1 ) = e + C(M)

[0027]

[0028] where e represents the path length of the sample from the root node to the leaf node of the isolation tree, C(M) represents the average path length of an isolation tree constructed with M samples, and B(M - 1) is the harmonic number, and the calculation formula is B(M - 1) = ln(M - 1) + 0.5772156649.

[0029] Furthermore, the quality assessment of GNSS measurements in step 5 based on the isolation forest is specifically as follows:

[0030]

[0031] where ρ k+1is the measurement switchable factor at the (k + 1)-th moment of GNSS, and T is the detection threshold of the Isolation Forest anomaly score.

[0032] Furthermore, the pre-integration factor in step 6

[0033]

[0034] where is the measurement residual of INS at the (k + 1)-th moment, is the process noise covariance matrix of INS at the (k + 1)-th moment; is the Mahalanobis distance of the residual, and the calculation formula is

[0035] GNSS factor

[0036]

[0037] where is the position difference between GNSS and INS at the (k + 1)-th moment, is the GNSS measurement matrix, and X k+1 is the state estimation vector at the (k + 1)-th moment; is the measurement noise covariance matrix of GNSS at the (k + 1)-th moment;

[0038] Fault-tolerant factor graph optimization algorithm:

[0039]

[0040] where f Prior is the prior factor of factor graph optimization, which is provided by the initial state information of integrated navigation; is the INS pre-integration factor at the j-th moment when traversing the navigation system; ρ j is the switchable factor at the j-th moment when traversing the navigation system, is the GNSS factor at the j-th moment when traversing the navigation system.

[0041] A computer device / system, including a memory, a processor, and a computer program stored on the memory, and the processor executes the computer program to implement the steps of the GNSS / INS vehicle-borne fault-tolerant integrated navigation method based on Isolation Forest fault detection.

[0042] A computer-readable storage medium, on which a computer program / instruction is stored, and when the computer program / instruction is executed by a processor, the steps of the GNSS / INS vehicle-borne fault-tolerant integrated navigation method based on Isolation Forest fault detection are implemented.

[0043] A computer program product includes a computer program / instructions, and when the computer program / instructions are executed by a processor, the steps of a GNSS / INS vehicle fault-tolerant integrated navigation method based on isolated forest fault detection are implemented.

[0044] The beneficial effects of the present invention are as follows:

[0045] 1. The present invention proposes an unsupervised isolated forest anomaly detection method based on machine learning, which is particularly suitable for the frequent occurrence of GNSS measurement anomalies caused by signal interferences such as multipath effects and non-line-of-sight reception in urban environments. This method constructs an isolated forest model through a data-driven approach and utilizes the random characteristics of isolated trees to improve the accuracy and sensitivity of fault detection, thereby effectively avoiding the contamination of the GNSS / INS integrated navigation system by GNSS faults.

[0046] 2. The present invention innovatively combines the anomaly scores of the isolated forest model to qualitatively and quantitatively evaluate the GNSS measurement quality, and combines the switchable factors in factor graph optimization to further explore the potential utilization value of abnormal measurements, thereby improving the positioning accuracy and reliability of the GNSS / INS integrated navigation system, and having certain engineering application value. Description of the Drawings

[0047] Figure 1 is the flow chart of the implementation algorithm of the present invention;

[0048] Figure 2 is the schematic diagram of isolated tree fault detection and path length of the present invention;

[0049] Figure 3 is the horizontal positioning error calculated by the algorithm proposed by the present invention. Detailed Embodiment

[0050] The following further describes the present invention with reference to the drawings.

[0051] A GNSS / INS vehicle fault-tolerant integrated navigation method, system, medium and product based on isolated forest fault detection of the present invention fully preheats the inertial measurement elements of the integrated navigation system and binds the initial navigation parameters of GNSS / INS; collects navigation increment data in urban environments; randomly divides the data, recursively generates isolated trees, and forms an isolated forest for fault detection; calculates the path length and anomaly scores of real-time data to determine whether GNSS measurements are faulty; qualitatively and quantitatively evaluates GNSS observations; adjusts the GNSS switchable factors and optimizes the factor graph weights to achieve optimal fusion. The present invention can achieve accurate fault detection and optimal fusion of GNSS in a complex urban environment for the integrated navigation system.

[0052] According to Figure 1 and 2As shown below, the specific steps of the method are as follows:

[0053] Step 1: Static start the GNSS / INS vehicle integrated navigation system, and use the heading information of GNSS as the initial heading ψ0 of the carrier, where 0 represents the initial moment; Given that the vehicle is in a stationary state on the road surface, the initial pitch angle θ0 and roll angle γ0 are set to zero; The initial velocity information [v E,0 v N,0 v U,0 T is also set to zero, where v E,0 , v N,0 and v U,0 respectively represent the initial eastward, northward, and upward velocities of the vehicle; The initial position information [L0λ0h0] T is provided by GNSS, where L0, λ0, and h0 respectively represent the initial latitude, longitude, and altitude of the vehicle. Subsequently, the above initial navigation parameters are loaded into the integrated navigation system.

[0054] Step 2: Collect navigation data under normal working conditions in an urban environment, including gyroscope information ω b , accelerometer information f b and the latitude, longitude, and altitude information [L GNSS λ GNSS h GNSS T , where ω b is the angular velocity information of the gyroscope in the body coordinate system b, f b is the acceleration information of the accelerometer in the body coordinate system b. The data collection should fully cover the dynamic characteristics of the urban environment.

[0055] Step 3: Perform inertial navigation solution based on the gyroscope information ω b and the accelerometer information f b . Integrate the attitude velocity and position at the (k - 1)-th moment to obtain the attitude velocity and position of INS at the k-th moment. Take the difference between the navigation information of INS at the previous and current moments, and calculate the attitude increment velocity increment and position increment

[0056]

[0057] Step 4: Collect the measured position information at the k-th moment by GNSS and the position information of GNSS at the (k - 1)-th moment​​ Differencing is performed to obtain the position increment of GNSS within the interval [k - 1, k].

[0058]

[0059] Step 5: Record the GNSS and INS increment information collected by the system at the k-th moment and save it as the navigation increment data sample at the k-th moment. Subsequently, all navigation increment data samples [x1, …, x k T from the initial moment to the current moment are composed to form the training dataset D, and M samples are randomly selected from it to form the sample set χ, which serves as the input to the isolation forest model. At the same time, the limited depth l of the isolation tree is set.

[0060] Step 6: Randomly specify a dimension μ as the starting node, and randomly select a value between the maximum and minimum values of this dimension. Samples less than this value are assigned to the left subtree, and samples greater than or equal to this value are assigned to the right subtree.

[0061] Step 7: Recursively execute Step 6 to continuously construct new isolation trees until there is only one sample in the isolation tree or the child node has reached the limited depth l. A forest of isolation trees jointly constructed by multiple basic isolation trees can detect faults in the integrated navigation system.

[0062] Step 8: Continue to collect the increment data x k+1 at the (k + 1)-th moment in real time, and calculate its path length G(x k+1 ) from the root node to the external node in each isolation tree:

[0063] G(x k+1 ) = e + C(M) (5)

[0064]

[0065] where e represents the path length of the sample from the root node to the leaf node of the isolation tree, C(M) represents the average path length of an isolation tree constructed with M samples, and B(M - 1) is the harmonic number, and its calculation formula is B(M - 1) = ln(M - 1) + 0.5772156649.

[0066] Step 9: Calculate the anomaly score S k+1 of the real-time detection data x k+1 in the isolation forest:

[0067]

[0068] where E(G(x k+1 )) is the sample x k+1 ​The path expected value of each isolated tree in the isolated forest.

[0069] Step 10: Use the real-time anomaly score S k+1 as the input for fault detection:

[0070]

[0071] where T is the detection threshold of the isolated forest anomaly score (usually T = 0.6 in the urban canyon environment and T = 0.55 in the urban open environment).

[0072] Step 11: Judge whether the current GNSS measurement is faulty according to the anomaly score. If the GNSS measurement anomaly score is lower than the threshold, keep the GNSS switchable factor as 1; if it exceeds the threshold, evaluate the measurement quality and calculate the optimized GNSS switchable factor ρ k+1 :

[0073]

[0074] Step 12: Calculate the INS pre-integration factor of the fault tolerance factor graph optimization method according to the gyroscope information ω b and the accelerometer information f b ,

[0075]

[0076] where is the measurement residual at the (k + 1)-th moment of the INS, is the process noise covariance matrix at the (k + 1)-th moment of the INS. is the Mahalanobis distance of the residual, and the calculation formula is

[0077] Step 13: Calculate the GNSS factor of the fault tolerance factor graph optimization method according to the measurement information of GNSS at the (k + 1)-th moment

[0078]

[0079] where is the position difference between GNSS and INS at the (k + 1)-th moment, is the GNSS measurement matrix, and X k+1 is the state estimation vector at the (k + 1)-th moment. is the measurement noise covariance matrix of GNSS at the (k + 1)-th moment.

[0080] Step 14: Based on the GNSS switchable factor ρ obtained in Step 11 k+1, adaptively scale the covariance matrix of the GNSS factor. The calculation expression of the improved fault tolerance factor graph optimization method is as follows:

[0081]

[0082] Among them, f Prior is the prior factor for factor graph optimization, provided by the initial state information of integrated navigation; is the INS pre-integration factor at the j-th moment when traversing the navigation system; ρ j is the GNSS switchable factor at the j-th moment when traversing the navigation system, is the GNSS factor at the j-th moment when traversing the navigation system.

[0083] Step 15: Obtain the optimal navigation state estimate through Step 14, and use the calculated navigation variables as the output of the integrated navigation system to achieve fault tolerance processing for GNSS abnormal measurements. At the same time, perform closed-loop feedback correction on the navigation state of the integrated navigation system to further reduce the nonlinear characteristics of the system and improve the information fusion accuracy.

[0084] Thus, the GNSS / INS vehicle-mounted fault tolerance integrated navigation method based on isolated forest fault detection is completed.

[0085] Example 1:

[0086] To illustrate the effectiveness of the algorithm, an urban vehicle-mounted test is conducted on the algorithm. The test conditions are as follows: The experimental platform includes a test data acquisition device and a high-precision reference device. The test device uses a self-developed low-cost inertial measurement unit. The gyroscope zero bias is 10° / h, the accelerometer zero bias is 5mg, and the sampling rate is 100Hz. The GNSS data is provided by a single-antenna receiver module u-blox M8T. The ground reference device is a high-precision integrated navigation system PHINS. The measurement data is collected through a dual-antenna receiver module Huace CHC-CGI-610, and the positioning enhancement service is enabled. The installation error angles of the two devices have been calibrated, and the coordinate system alignment has been achieved through a reference board. The test results are as shown in the appendix Figure 3 shown, the horizontal positioning error is small and the accuracy is high.

[0087] The functions of the GNSS / INS vehicle fault-tolerant integrated navigation system based on isolated forest fault detection according to the present invention can be described by the aforementioned GNSS / INS vehicle fault-tolerant integrated navigation method based on isolated forest fault detection. The system includes a Central Processing Unit (CPU) 301, which can perform various appropriate actions and processes according to the program stored in the Read-Only Memory (ROM) 302 or the program loaded into the Random Access Memory (RAM) 303 from the storage section 308, such as executing the method described in the above embodiments. In the RAM 303, various programs and data required for the operation of the rescue response system are also stored. The CPU 301, ROM 302, and RAM 303 are connected to each other via a bus 304. An Input / Output (I / O) interface 305 is also connected to the bus 304.

[0088] The following components are connected to the I / O interface 305: an input section 306 including an audio input device, button switches, etc.; an output section 307 including a Liquid Crystal Display (LCD), an audio output device, an indicator light, etc.; a storage section 308 including a hard disk, etc.; and a communication section 309 including a network interface card such as a LAN (Local Area Network) card, a modem, etc. The communication section 309 performs communication processing via a network such as the Internet. A drive 310 is also connected to the I / O interface 305 as needed. A removable medium 311, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc., is installed on the drive 310 as needed so that a computer program read from it can be installed into the storage section 308 as needed.

[0089] In particular, according to an embodiment of the present invention, the process described above with reference to the flowchart can be implemented as a computer software program. For example, an embodiment of the present invention includes a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program includes a computer program for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from the network via the communication section 309, and / or installed from the removable medium 311. When the computer program is executed by the Central Processing Unit (CPU) 301, various functions defined in the present invention are executed.

[0090] It should be noted that specific examples of computer-readable storage media may include, but are not limited to: electrical connections with one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM), flash memory, optical fibers, portable compact disc read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the above. In the present invention, a computer-readable storage medium can be any tangible medium that contains or stores a program, and this program can be used by or in combination with an instruction execution system, device, or apparatus.

[0091] Specifically, the GNSS / INS vehicle-mounted fault-tolerant integrated navigation system based on isolated forest fault detection in this embodiment includes a processor and a memory. A computer program is stored on the memory. When the computer program is executed by the processor, it implements the GNSS / INS vehicle-mounted fault-tolerant integrated navigation method based on isolated forest fault detection provided in the above embodiment.

[0092] As another aspect, the present invention also provides a computer-readable storage medium. This storage medium can be included in the GNSS / INS vehicle-mounted fault-tolerant integrated navigation system based on isolated forest fault detection described in the above embodiment; or it can exist alone without being assembled into the GNSS / INS vehicle-mounted fault-tolerant integrated navigation system based on isolated forest fault detection. The above storage medium carries one or more computer programs. When the one or more computer programs are executed by a processor of the GNSS / INS vehicle-mounted fault-tolerant integrated navigation system based on isolated forest fault detection, the GNSS / INS vehicle-mounted fault-tolerant integrated navigation system based on isolated forest fault detection is enabled to implement the GNSS / INS vehicle-mounted fault-tolerant integrated navigation method provided in the above embodiment.

[0093] The above are only the preferred embodiments of the present invention and are not used to limit the present invention. For those skilled in the art, the present invention can have various changes and modifications. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention should be included within the protection scope of the present invention.

Claims

1. A GNSS / INS vehicle fault-tolerant integrated navigation method based on isolated forest fault detection, characterized in that Including the following steps: Step 1: Bind the initial navigation parameters of GNSS / INS, initialize the factor graph optimization parameters, and then enter the navigation working mode; Step 2: Collect navigation data under normal working conditions, and calculate the navigation increment data within the INS unit time interval, including attitude increment velocity increment and position increment Calculate the position increment data within the GNSS unit time interval and store them sequentially; Step 3: All navigation increment data samples from the initial moment to the k-th moment [x1, …, x k T are used to form the training dataset D, and M samples are randomly selected from it to form the sample set χ, which serves as the input to the isolation forest model. At the same time, the limited depth l of the isolation tree is set, and the data is randomly divided and recursively used to generate isolation trees, forming an isolation forest for fault detection;​ Step 4: Collect the incremental data x at the (k + 1)-th moment in real time k+1 , calculate the anomaly score S k+1 , S k+1 is used as the input for fault detection, and isolation forest fault detection is performed; Step 5: Based on the anomaly score S k+1 Determine whether the current GNSS measurement is faulty. If the anomaly score of the GNSS measurement is lower than the threshold, the GNSS has no fault and the switching factor is 1; If the threshold is exceeded, the measurement quality is evaluated and the switchable factor ρ is calculated k+1 ; Step 6: Calculate the INS pre-integration factor and the GNSS factor And perform the fault-tolerant factor graph optimization algorithm to output the optimal estimated navigation information, realizing the fault tolerance processing of GNSS abnormal measurements; at the same time, perform closed-loop feedback correction on the navigation state of the integrated navigation system.

2. The GNSS / INS vehicle-mounted fault-tolerant integrated navigation method based on isolated forest fault detection according to claim 1, characterized in that: The heading information of GNSS in step 1 is used as the initial heading ψ0 of the carrier, where 0 represents the initial moment; considering that the vehicle is in a stationary state on the road surface, the initial pitch angle θ0 and roll angle γ0 are set to zero; the initial velocity information [v E, 0v N,0 v U,0 T is also set to zero, where ν E,0 , ν N,0 and ν U,0 respectively represent the initial eastward, northward and upward velocities of the vehicle; the initial position information [L0λ0h0] T is provided by GNSS, where L0, λ0 and h0 respectively represent the initial latitude, longitude and altitude of the vehicle.

3. A GNSS / INS vehicle fault-tolerant integrated navigation method based on isolated forest fault detection according to claim 1, characterized in that: The specific content of Step 2 is as follows: Step 2.1: The collected navigation data under normal working conditions include: gyroscope information ω b , accelerometer information f b , and the latitude, longitude, and altitude information [L GNSS λ GNSS h GNSS provided by GNSS T , where ω b is the angular velocity information of the gyroscope in the body coordinate system b, and f b is the acceleration information of the accelerometer in the body coordinate system b; Step 2.2: Based on the gyroscope information ω b and the accelerometer information f b perform inertial navigation solution, and integrate from the attitude velocity and position at the (k - 1)-th moment to obtain the attitude velocity and position of the INS at the k-th moment. Take the difference between the navigation information of the INS at the previous and current moments, and calculate the attitude increment velocity increment and position increment Step 2.3: Collect the measured position information at the k-th moment by GNSS and the position information of GNSS at the (k-1)-th moment to perform differencing to obtain the position increment of GNSS in the interval [k-1, k] Step 2.4: Record the GNSS and INS incremental information samples collected by the system at the k-th moment and save them as the k-th navigation incremental data sample.

4. A GNSS / INS vehicle fault-tolerant integrated navigation method based on isolated forest fault detection according to claim 1, characterized in that: In Step 3, a dimension μ is randomly specified as the starting node, and a value is randomly selected between the maximum and minimum values of this dimension. Samples smaller than this value are assigned to the left subtree, and samples greater than or equal to this value are assigned to the right subtree; the above steps are recursively executed to continuously construct a new isolation tree until there is only one sample in the isolation tree or the sub-node has reached the limited depth l.

5. A GNSS / INS vehicle fault-tolerant integrated navigation method based on isolated forest fault detection according to claim 1, characterized in that: The fault detection of GNSS measurements based on the isolation forest described in step 4 is specifically as follows: real-time collect the navigation increment data x at the (k + 1)-th moment k+1 , and calculate the anomaly score S in the isolation forest k+1 : Among them, L(x k+1 ) is the path length from the root node to the external node of the sample x k+1 in the isolation tree, and E(L(x k+1 )) is the path expectation value of each isolation tree in the isolation forest for the sample x k+1 ; C(M) represents the average path length of an isolation tree constructed with M samples. where L(x k+1 ) is: L(x k+1 ) = e + C(M) Among them, e represents the path length of the sample from the root node to the leaf node of the isolation tree, C(M) represents the average path length of an isolation tree constructed with M samples, and B(M - 1) is the harmonic number, and the calculation formula is B(M - 1) = ln(M - 1) + 0.5772156649.

6. A GNSS / INS vehicle fault-tolerant integrated navigation method based on isolated forest fault detection according to claim 1, characterized in that: The specific content of Step 5 for quality assessment of GNSS measurements based on the isolation forest is as follows: Among them, ρ k+1 is the measurement switchable factor at the (k + 1)-th moment of GNSS, and T is the detection threshold of the anomaly score of the Isolation Forest.

7. A GNSS / INS vehicle fault-tolerant integrated navigation method based on isolated forest fault detection according to claim 1, characterized in that: The pre-integration factor in step 6 wherein, is the measurement residual at the (k + 1)-th moment of the INS, is the process noise covariance matrix at the (k + 1)-th moment of the INS; is the Mahalanobis distance of the residual, and the calculation formula is GNSS factor Among them, is the position difference between GNSS and INS at the (k + 1)-th moment, is the GNSS measurement matrix, and X k+1 is the state estimation vector at the (k + 1)-th moment; is the measurement noise covariance matrix of GNSS at the (k + 1)-th moment; Fault-tolerant factor graph optimization algorithm: Among them, f Prior is the prior factor for factor graph optimization, provided by the initial state information of integrated navigation; is the INS pre-integration factor at the j-th moment when traversing the navigation system; ρ j is the switchable factor at the j-th moment when traversing the navigation system, is the GNSS factor at the j-th moment when traversing the navigation system.

8. A computer device / equipment / system, comprising a memory, a processor, and a computer program stored on the memory, characterized in that: The processor executes the computer program to implement the steps of the method described in any one of claims 1 to 7.

9. A computer-readable storage medium having computer programs / instructions stored thereon, characterized in that: When the computer program / instructions are executed by the processor, the steps of the method described in any one of claims 1 to 7 are implemented.

10. A computer program product comprising a computer program / instructions, characterized in that: When the computer program / instructions are executed by the processor, the steps of the method described in any one of claims 1 to 7 are implemented.

Images