Sensor error correction platform and method for inertial navigation system

Abstract

The invention discloses a sensor error correction platform and method for an inertial navigation system, and belongs to the technical field of error correction, and the method comprises the steps: carrying out the pre-estimation of the inclination of a vehicle through an auto-encoder according to an acceleration value, a vibration quantity and a rotation angle which are sampled in real time; obtaining a vehicle driving inclination estimation queue; estimating the corrected gradient of each time point in the queue according to the vehicle driving gradient in the last time period, calculating the matching type criticality of each time point, and determining a vehicle safety determination factor; according to the vehicle safety affirmation factor, affirmation of the vehicle safety estimation factor, construction of a vehicle safety hidden danger level and execution of estimation of vehicle driving safety, the accuracy of measurement of the inclination degree of the electrical vehicle of the building is improved, and thus estimation of the vehicle safety is more accurate.

Description

Technical Field

[0001] The present invention belongs to the technical field of error correction, and in particular relates to a sensor error correction platform and method for an inertial navigation system. Background Art

[0002] An inertial navigation system is an autonomous navigation system that uses inertial sensitive devices, reference directions and initial position information to determine the position, direction and speed of a vehicle in inertial space, and is sometimes referred to as an inertial navigation system.

[0003] In practical applications, current inertial navigation system technology often uses the prior art solution mentioned in patent publication number "CN111735450B", which includes a connected vehicle control system and a vehicle inertial sensor. The vehicle inertial sensor is used to sample the vehicle's inertial navigation data and transmit it to the vehicle control system. The vehicle control system is used to perform navigation planning for the vehicle based on the collected vehicle's inertial navigation data.

[0004] The vehicle's inertial navigation data includes acceleration values, inclination, vibration and rotation angle. When the vehicle's inclination is a predetermined inclination, the vehicle will tip over. To reduce the risk of vehicle tipping over, the vehicle's inclination must be detected to estimate the safety of the vehicle's driving.

[0005] Currently, the safety of vehicle driving is estimated by averaging the vehicle's inclination during a period of time. Because the vehicle's inclination fluctuates and is not constant, and the vehicle's inclination is sampled by the vehicle's inertial sensor, which can often only sample the inclination of one point of the vehicle. Therefore, current calculations often use the vehicle's vibration and acceleration values ​​to perform a prior estimate of the vehicle's inclination. Because when the vehicle makes a sharp turn, the vehicle's inclination often presents a problem of causing the vehicle to tip over in a short period of time. This type of problem is defined as a latent problem. Using a constant criticality to estimate the vehicle's inclination in the short term will carry the constant criticality's sensitivity to short-term inclination fluctuations, resulting in a low accuracy in the estimation of potential problems with the vehicle's inclination, thereby weakening the accuracy of the estimation of the vehicle's safety. Summary of the invention

[0006] In order to solve the defects in the prior art, the present invention proposes a sensor error correction platform and method for an inertial navigation system, which uses an autoencoder to perform pre-estimation of the vehicle's inclination through the real-time sampled acceleration value, vibration amount, and rotation angle; through the vehicle inclination at each time point, the potential distortion factor of the vehicle inclination is determined, and a potential distortion inclination queue is obtained; based on the inclination with potential distortion, it is corrected, and the corrected inclination at each time point in the vehicle safety inclination queue is determined to obtain a vehicle driving inclination estimation queue; based on the corrected inclination at each time point in the vehicle driving inclination estimation queue of the last period, the matching criticality at each time point is calculated to determine the vehicle safety determination factor; based on the vehicle safety determination factor, the vehicle safety estimation factor is determined, and a vehicle safety hidden danger level is constructed, and the vehicle driving safety is estimated, thereby improving the accuracy of the building electrical vehicle inclination measurement, so as to make the vehicle safety estimation more accurate.

[0007] The present invention utilizes the following technical solutions.

[0008] A sensor error correction method for an inertial navigation system, comprising:

[0009] The vehicle inertial sensor samples the vehicle's inertial navigation data and transmits it to the vehicle control system. The vehicle control system performs navigation planning for the vehicle based on the received vehicle's inertial navigation data. The vehicle's inertial navigation data includes acceleration value, inclination, vibration amount and rotation angle;

[0010] The vehicle control system also performs the following method on the inertial navigation data of the vehicle it collects:

[0011] Step 1: Collect the inertial navigation data sampled by the vehicle inertial sensor and transmitted by the vehicle, obtain the vehicle's pre-estimated inclination, and obtain the vehicle's safety inclination queue;

[0012] Step 2: According to the arrangement and classification of the inclination at each time in the vehicle safety inclination queue, the potential distortion factor of the vehicle inclination is determined to obtain the potential distortion inclination queue;

[0013] Step 3: According to the relative arrangement of the inclination at each time point in each potential distortion inclination queue and the potential distortion factor of the vehicle inclination, the corrected inclination at each time point in the vehicle safety inclination queue is determined to obtain the vehicle driving inclination estimation queue;

[0014] Step 4: Estimate the corrected inclination at each time point in the queue based on the vehicle driving inclination, determine the vehicle safety determination factor, and perform an estimate on the vehicle driving safety.

[0015] Furthermore, in Step 1, the real-time sampled acceleration value, vibration amount, and rotation angle of the vehicle are sent to the autoencoder to perform learning using the autoencoder. The vehicle inertial sensor samples the acceleration value, vibration amount, and rotation angle of the vehicle in real time and transmits them to the autoencoder on the vehicle control system, and the vehicle's pre-estimated inclination is obtained through the autoencoder.

[0016] Furthermore, in Step 1, the real-time sampling speed is 2 times / second, and the total sampling time is ten minutes. All the pre-estimated inclinations are arranged in the order of the sampling time points and defined as a vehicle safety inclination queue.

[0017] Furthermore, in Step 2, the vehicle safety tilt queue is equally divided into several sub-queues, which are defined as each distortion tilt queue;

[0018] Get the full range of each distortion tilt queue respectively, and form a tilt full range cluster with all the full ranges;

[0019] By the difference between the full range values, the state expansion and contraction factor is calculated.

[0020] Furthermore, in Step 2, the method of calculating the expansion and contraction factor of the identified state includes:

[0021] The amount obtained by multiplying the base position decrement of the tilt range cluster by 2 and dividing it by the total range of the tilt range cluster is defined as a ratio of one;

[0022] If the ratio one is not lower than the pre-defined recognition threshold, the state expansion factor is determined to be ratio one; if the ratio one is lower than the pre-defined recognition threshold or the full range of the inclination full range cluster is zero, the value of the state expansion factor is determined to be one.

[0023] Furthermore, in Step 2, the calculation method of reducing the reference position of the tilt full-range cluster includes:

[0024] After arranging the full ranges in the tilt full range cluster in order from low to high to form a queue with p elements, obtain the R1th element and the R2th element in the queue, and the value obtained by subtracting the value of the R1th element from the value of the R2th element and dividing it by two is the base position decrement of the tilt full range cluster;

[0025] Where:

[0026] Furthermore, the value of the predefined recognition threshold is one.

[0027] Furthermore, in Step 2, the potential distortion factor of the vehicle inclination is calculated, which includes:

[0028] Calculate the mean of the slope range cluster, which is defined as mean 1;

[0029] The amount obtained by adding the mean value 1 + the mean value 1 of the determined state expansion and contraction factor is determined as the potential distortion factor of the vehicle inclination;

[0030] The distortion tilt queue of all distortion tilt queues with a full range height exceeding the tilt potential distortion identification factor is defined as the potential distortion tilt queue.

[0031] Further, in Step 3, according to the difference between the inclination at each time point in each potential distortion inclination queue and the minimum amount of its potential distortion inclination queue, the corresponding difference amount at each time point in each potential distortion inclination queue is determined;

[0032] The difference between the corresponding difference amount and the potential distortion factor of the vehicle inclination is regarded as the vehicle driving inclination distortion parameter at each time point in each potential distortion inclination queue;

[0033] Then, according to the vehicle driving inclination distortion parameter, the inclination at each time point in the potential distortion inclination queue is corrected;

[0034] Furthermore, in Step 3, the method for correcting the tilt at each time point in the potential distortion tilt queue includes:

[0035] The quantity obtained by applying the modified linear unit to the slope at each time point in each potential distortion slope queue is defined as unit quantity 1;

[0036] Aggregate the minimum amount of each potential distortion inclination queue, its vehicle inclination potential distortion factor and the unit amount to determine the correction factor at each time point in each potential distortion inclination queue;

[0037] According to the corrected inclination at each time point, the inclination at the corresponding time point in the vehicle safety inclination queue is replaced to obtain the corrected inclination at each time point in the vehicle safety inclination queue;

[0038] The vehicle safety inclination queue is formed by selecting the inclination corrected at several time points from the tail of the queue to the head of the queue to form a vehicle driving inclination estimation queue.

[0039] Furthermore, in Step 3, the corresponding difference amount at each time point in each potential distortion slope queue is the amount obtained by subtracting the minimum amount of the potential distortion slope queue from the slope at each time point in each potential distortion slope queue.

[0040] Furthermore, the vehicle driving inclination distortion parameter at each time point in each potential distortion inclination queue is a quantity obtained by subtracting the vehicle inclination potential distortion factor from the corresponding difference quantity.

[0041] Furthermore, in Step 3, the unit quantity of each time point in each potential distortion inclination queue is multiplied by the potential distortion factor of the vehicle inclination, and the product value of each time point in each potential distortion inclination queue is defined as one, and the product value of one and the minimum quantity of each potential distortion inclination queue are used as the correction factor of each time point in each potential distortion inclination queue;

[0042] If the vehicle's driving inclination distortion parameter is not less than zero, then the inclination corrected at the corresponding time point is the correction factor; if the vehicle's driving inclination distortion parameter is less than zero, then the inclination at the corresponding time point maintains the original value.

[0043] Furthermore, the corrected inclinations at the last 120 time points in the vehicle safety inclination queue are used to form a vehicle driving inclination estimation queue.

[0044] Furthermore, in Step 4, the vehicle driving criticality determination factor is calculated, which includes:

[0045] According to the difference between the corrected inclination at each time point in the vehicle driving inclination estimation queue and the average of the vehicle driving inclination estimation queue, the actual inclination difference at each time point in the vehicle driving inclination estimation queue is determined;

[0046] The amount obtained by dividing the corrected inclination at each time point in the vehicle driving inclination estimation queue by the actual inclination difference is used as the vehicle driving criticality measurement factor at each time point in the vehicle driving inclination estimation queue.

[0047] Furthermore, the modulus of the quantity obtained by subtracting the average of the vehicle driving inclination estimation queue from the corrected inclination at each time point in the vehicle driving inclination estimation queue is taken as the actual inclination difference at each time point in the vehicle driving inclination estimation queue.

[0048] Furthermore, in Step 4, the matching criticality at each time point is calculated by arranging the vehicle driving criticality determination factors, which includes:

[0049] The mean of the vehicle driving criticality determination factors at all time points in the vehicle driving inclination estimation queue is defined as mean 2;

[0050] The amount obtained by dividing the vehicle driving criticality determination factor at each time point in the vehicle driving inclination estimation queue by the mean of two is used as the matching criticality at each time point in the vehicle driving inclination estimation queue.

[0051] Furthermore, in Step 4, the product value obtained by multiplying the corrected inclination at each time point in the vehicle driving inclination estimation queue by the matching formula key degree at the corresponding time point is obtained, and the amount obtained by accumulating all the product values ​​is divided by the number of elements in the vehicle driving inclination estimation queue, which is used as the vehicle safety determination factor;

[0052] Then, based on the vehicle safety identification factor, the vehicle safety of the building is estimated:

[0053] The vehicle safety identification factors are standardized using the Z-score method. If the standardized value is higher than the pre-defined critical value of the vehicle rollover hazard, then the vehicle driving has a dangerous hazard. The vehicle control system controls the speaker connected to it to issue a voice prompt of the vehicle driving danger to prevent the electric vehicle from rolling over. If the standardized value is not higher than the pre-defined critical value of the vehicle rollover hazard, then the vehicle driving has no dangerous hazard.

[0054] A sensor error correction platform for an inertial navigation system, comprising:

[0055] A connected vehicle control system and a vehicle inertial sensor, the vehicle inertial sensor is used to sample the vehicle's inertial navigation data and transmit it to the vehicle control system, the vehicle control system is used to perform navigation planning for the vehicle based on the collected vehicle's inertial navigation data, the vehicle's inertial navigation data includes acceleration value, inclination, vibration amount and rotation angle;

[0056] The modules running on the vehicle control system include:

[0057] An acquisition module is used to collect the inertial navigation data of the vehicle sampled by the vehicle inertial sensor and transmitted to obtain the vehicle's pre-estimated inclination and obtain the vehicle's safety inclination queue;

[0058] A classification module, which is used to classify the status of the vehicle safety inclination according to the arrangement of the inclination at each time in the vehicle safety inclination queue, identify the potential distortion factor of the vehicle inclination, and obtain the potential distortion inclination queue;

[0059] A recognition module, which is used to recognize the corrected inclination at each time point in the vehicle safety inclination queue according to the associated arrangement status of the inclination at each time point in each potential distortion inclination queue and the potential distortion factor of the vehicle inclination, and obtain the vehicle driving inclination estimation queue;

[0060] The estimation module is used to estimate the corrected inclination at each time point in the queue based on the vehicle driving inclination, determine the vehicle safety determination factor, and perform estimation on the vehicle driving safety.

[0061] The beneficial effects of the present invention are that, compared with the prior art, the technical effects of the present invention include:

[0062] The autoencoder is used to pre-estimate the vehicle's inclination through the real-time sampled acceleration value, vibration amount, and rotation angle, thereby avoiding the error caused by obtaining the vehicle inclination at only one place through the inclination sensor, which is conducive to improving the accuracy of subsequent vehicle safety estimation; through the vehicle inclination at each time point, the potential distortion factor of the vehicle inclination is determined, and a potential distortion inclination queue is obtained, thereby identifying and distinguishing the inclination with potential distortion; based on the inclination with potential distortion, it is corrected, and the corrected inclination at each time point in the vehicle safety inclination queue is determined to obtain the vehicle driving inclination estimation queue, thereby The slope is corrected to remove the disturbance of potential distortion and improve the accuracy of vehicle safety estimation. The corrected slope at each time point in the queue is estimated based on the vehicle driving inclination in the last period of time, and the matching criticality at each time point is calculated to determine the vehicle safety identification factor. Based on the different criticalities corresponding to the corrected slopes at different time points, the overall vehicle driving safety can be estimated more accurately. Based on the vehicle safety identification factor, the vehicle safety estimation factor is determined, the vehicle safety hazard level is constructed, and the vehicle driving safety is estimated, thereby improving the accuracy of the building electrical vehicle inclination measurement, so as to make the vehicle safety estimation more accurate. BRIEF DESCRIPTION OF THE DRAWINGS

[0063] Figure 1 is a flow chart of a sensor error correction method for an inertial navigation system according to the present invention;

[0064] Figure 2 It is a partial structural diagram of the sensor error correction platform for the inertial navigation system described in the present invention. DETAILED DESCRIPTION

[0065] In order to make the purpose, technical solution and advantages of the present invention clearer, the following will be combined with the drawings in the embodiments of the present invention to clearly and completely express the technical solution of the present invention. The embodiments expressed in this application are only partial embodiments of the present invention, not all embodiments. According to the spirit of the present invention, other embodiments obtained by technicians in this field without creative work are all within the protection scope of the present invention.

[0066] like Figure 1 As shown, a sensor error correction method for an inertial navigation system according to the present invention comprises:

[0067] The vehicle inertial sensor samples the vehicle's inertial navigation data and transmits it to the vehicle control system. The vehicle control system performs navigation planning for the vehicle based on the received vehicle's inertial navigation data. The vehicle's inertial navigation data includes acceleration value, inclination, vibration amount and rotation angle;

[0068] The vehicle control system also performs the following method on the inertial navigation data of the vehicle it collects:

[0069] Step 1: Collect the inertial navigation data sampled by the vehicle inertial sensor and transmitted by the vehicle, obtain the vehicle's pre-estimated inclination, and obtain the vehicle's safety inclination queue;

[0070] The vehicle's acceleration, vibration, inclination, and rotation angle are sampled by the vehicle's inertial sensor and transmitted to the vehicle control system. The acceleration, vibration, rotation angle, inclination, and other values ​​in the vehicle control system are used for learning using an autoencoder, so that the real-time sampled acceleration, vibration, and rotation angle sent to the autoencoder can be used to obtain the vehicle's estimated inclination in advance.

[0071] In a preferred but non-restrictive embodiment of the present invention, in Step 1, the real-time sampled acceleration value, vibration amount, and rotation angle of the vehicle are sent to the autoencoder to perform learning using the autoencoder. The vehicle inertial sensor samples the acceleration value, vibration amount, and rotation angle of the vehicle in real time and transmits them to the autoencoder on the vehicle control system, and the vehicle's pre-estimated inclination is obtained through the autoencoder.

[0072] In a preferred but non-limiting embodiment of the present invention, in Step 1, the real-time sampling speed is 2 times / second, the total sampling time is ten minutes, and all the pre-estimated inclinations are arranged in chronological order according to the sampling time points, and defined as a vehicle safety inclination queue.

[0073] So you get the vehicle safety gradient queue.

[0074] Step 2: According to the arrangement and classification of the inclination at each time in the vehicle safety inclination queue, the potential distortion factor of the vehicle inclination is determined to obtain the potential distortion inclination queue;

[0075] During the driving of the vehicle, it is very important to ensure the safety of the vehicle. When the vehicle suddenly turns, the vibration of the vehicle will suddenly increase. The vibration and inclination are positively correlated. Since the inclination of the vehicle is related to the vibration and acceleration value, the calculated vehicle inclination will increase. After the vehicle turns, the vibration will decrease to a reasonable level, so the vehicle inclination will temporarily appear in the pre-defined range, but it will be in a reasonable range.

[0076] Because the distortion of the vehicle's tilt is instantaneous, the distortion of the tilt value in a short period of time (that is, the sudden abnormal change value) is defined as potential distortion. Such potential distortion cannot bring hidden dangers to the safety of the vehicle. Therefore, when evaluating the safety of the vehicle, the interference of such potential distortion must be removed.

[0077] In a preferred but non-limiting embodiment of the present invention, in Step 2, the vehicle safety inclination queue is equally divided into a number of sub-queues, each of which is defined as a distortion inclination queue;

[0078] In a preferred but non-limiting embodiment of the present invention, in Step 2, the vehicle safety inclination queue is equally divided into 60 sub-queues.

[0079] Get the full range of each distortion tilt queue respectively, and form a tilt full range cluster with all the full ranges;

[0080] When there is potential distortion, the full range of the distortion inclination array in the potential distortion interval will be large, and there will be a large difference between the full range values ​​of the vehicle inclination in other valid intervals; therefore, the state expansion and contraction factor is calculated based on the difference between the full range values;

[0081] In a preferred but non-limiting embodiment of the present invention, in Step 2, the method for calculating the expansion and contraction factor of the identified state includes:

[0082] The amount obtained by multiplying the base position decrement of the tilt range cluster by 2 and dividing it by the total range of the tilt range cluster is defined as a ratio of one;

[0083] In a preferred but non-limiting embodiment of the present invention, in Step 2, the calculation method of the base position decrement of the tilt full-range cluster includes:

[0084] After arranging the full ranges in the tilt full range cluster in order from low to high to form a queue with p elements, obtain the R1th element and the R2th element in the queue, and the value obtained by subtracting the value of the R1th element from the value of the R2th element and dividing it by two is the base position decrement of the tilt full range cluster;

[0085] Where:

[0086] If the ratio one is not lower than the pre-defined recognition threshold, the state expansion factor is determined to be ratio one; if the ratio one is lower than the pre-defined recognition threshold or the full range of the inclination full range cluster is zero, the value of the state expansion factor is determined to be one.

[0087] In a preferred but non-limiting embodiment of the present invention, the value of the predefined identification threshold is one.

[0088] If the ratio is not less than 1, it means that the full ranges in the tilt range cluster are very close and the density of their arrangement is high, which means that the fluctuation of the full range value is not large and the probability of potential distortion is not high. Otherwise, when the ratio is lower than 1, it means that the tilt range cluster contains obvious singular values. These singular values ​​often represent abnormal fluctuations or potential distortion, thereby increasing the probability of potential distortion.

[0089] The inclination of the attribute with potential distortion is an attribute with short term and large value in the vehicle inclination, so the inclination of potential distortion is an attribute with singular points. To identify whether the distorted inclination queue is a potential distorted inclination, in a preferred but non-limiting embodiment of the present invention, in Step 2, the potential distortion factor of the vehicle inclination is calculated, which includes:

[0090] Calculate the mean of the slope range cluster, which is defined as mean 1;

[0091] The amount obtained by adding the mean value 1 + the mean value 1 of the determined state expansion and contraction factor is determined as the potential distortion factor of the vehicle inclination;

[0092] The distortion tilt queue of all distortion tilt queues with a full range height exceeding the tilt potential distortion identification factor is defined as the potential distortion tilt queue.

[0093] The mean one can reflect the midpoint value attribute of the vehicle inclination. Then, when there is no potential distortion, the mean one cannot make all values ​​reasonable. Therefore, the recognition state expansion and contraction factor of the distortion inclination array is used to correct the mean, so that the reasonable value can be cut into the reasonable value through the vehicle inclination potential distortion factor. When the full range of the distortion inclination array is higher than the vehicle inclination potential distortion factor, it means that the values ​​in the distortion inclination array are often potential distortion values, and the higher the difference between the full range of the distortion inclination array and the vehicle inclination potential distortion factor, the greater the probability that the distortion inclination array has potential distortion.

[0094] Therefore, the potential distortion factor of the vehicle inclination is obtained and the potential distortion inclination queue is identified.

[0095] Step 3: According to the relative arrangement of the inclination at each time point in each potential distortion inclination queue and the potential distortion factor of the vehicle inclination, the corrected inclination at each time point in the vehicle safety inclination queue is determined to obtain the vehicle driving inclination estimation queue;

[0096] In a preferred but non-limiting embodiment of the present invention, in Step 3, facing the distorted inclination queue with potential distortion, the safety of the vehicle in the firefighting is essentially within a reasonable range, and the potential danger is not high. Therefore, the values ​​in the distorted inclination queue with potential distortion can be corrected, which includes:

[0097] According to the difference between the inclination at each time point in each potential distortion inclination cohort and the minimum amount of its potential distortion inclination cohort, the corresponding difference amount at each time point in each potential distortion inclination cohort is determined;

[0098] In a preferred but non-limiting embodiment of the present invention, in Step 3, the corresponding difference amount at each time point in each potential distortion slope queue is the amount obtained by subtracting the minimum amount of its potential distortion slope queue from the slope of each time point in each potential distortion slope queue.

[0099] The difference between the corresponding difference amount and the potential distortion factor of the vehicle inclination is regarded as the vehicle driving inclination distortion parameter at each time point in each potential distortion inclination queue;

[0100] In a preferred but non-limiting embodiment of the present invention, the vehicle driving inclination distortion parameter at each time point in each potential distortion inclination queue is a quantity obtained by subtracting the vehicle inclination potential distortion factor from the corresponding difference quantity.

[0101] The lower the vehicle driving inclination distortion parameter is, the lower the difference between the inclination at each time point in the potential distortion inclination queue and the potential distortion identification factor of the inclination is, thereby increasing the probability of it belonging to a reasonable inclination fluctuation, and thus the inclination at that time point will be a more reasonable value.

[0102] Then, according to the vehicle driving inclination distortion parameter, the inclination at each time point in the potential distortion inclination queue is corrected;

[0103] In a preferred but non-limiting embodiment of the present invention, in Step 3, the method for correcting the tilt at each time point in the potential distortion tilt queue includes:

[0104] The quantity obtained by applying the modified linear unit to the slope at each time point in each potential distortion slope queue is defined as unit quantity 1;

[0105] Aggregate the minimum amount of each potential distortion inclination queue, its vehicle inclination potential distortion factor and the unit amount to determine the correction factor at each time point in each potential distortion inclination queue;

[0106] In a preferred but non-limiting embodiment of the present invention, in Step 3, the amount obtained by multiplying the unit amount of each time point in each potential distortion inclination queue by the potential distortion factor of the vehicle inclination is defined as the product value of each time point in each potential distortion inclination queue, and the product value of one and the minimum amount of each potential distortion inclination queue are used as the correction factor of each time point in each potential distortion inclination queue;

[0107] If the vehicle's driving inclination distortion parameter is not less than zero, then the inclination corrected at the corresponding time point is the correction factor; if the vehicle's driving inclination distortion parameter is less than zero, then the inclination at the corresponding time point maintains the original value.

[0108] The higher the corrected inclination at each time point in the potential distortion inclination queue, the greater the abnormality of the vehicle's inclination, thereby increasing the risk of vehicle rollover. A large inclination often indicates a low level of vehicle safety. Therefore, the vehicle safety of the building can be estimated through the corrected vehicle inclination.

[0109] According to the corrected inclination at each time point, the inclination at the corresponding time point in the vehicle safety inclination queue is replaced to obtain the corrected inclination at each time point in the vehicle safety inclination queue;

[0110] The vehicle safety inclination queue is formed by selecting the inclination corrected at several time points from the tail of the queue to the head of the queue to form a vehicle driving inclination estimation queue.

[0111] In a preferred but non-limiting embodiment of the present invention, the corrected inclinations at the last one hundred and twenty time points in the vehicle safety inclination queue are used to form a vehicle driving inclination estimation queue.

[0112] Therefore, the corrected inclination of the vehicle safety inclination queue at each time point is determined to obtain the vehicle driving inclination estimation queue.

[0113] Step 4: Estimate the corrected inclination at each time point in the queue based on the vehicle driving inclination, determine the vehicle safety determination factor, and perform an estimate on the vehicle driving safety.

[0114] In a preferred but non-limiting embodiment of the present invention, in Step 4, when the inclination of the vehicle increases, the vibration of the vehicle increases, thereby weakening the safety of the vehicle. With the continuous increase of the inclination, the probability of the vehicle rolling over increases accordingly. Therefore, the vehicle driving criticality determination factor is calculated based on the corrected inclination of the vehicle at each time point, which includes:

[0115] According to the difference between the corrected inclination at each time point in the vehicle driving inclination estimation queue and the average of the vehicle driving inclination estimation queue, the actual inclination difference at each time point in the vehicle driving inclination estimation queue is determined;

[0116] In a preferred but non-limiting embodiment of the present invention, the modulus of the quantity obtained by subtracting the average of the vehicle driving inclination estimation queue from the corrected inclination at each time point in the vehicle driving inclination estimation queue is taken as the actual inclination difference at each time point in the vehicle driving inclination estimation queue.

[0117] The vehicle driving inclination estimation factor at each time point in the vehicle driving column is positively correlated with the corrected inclination at the corresponding time point and negatively correlated with the actual inclination difference.

[0118] The amount obtained by dividing the corrected inclination at each time point in the vehicle driving inclination estimation queue by the actual inclination difference is used as the vehicle driving criticality measurement factor at each time point in the vehicle driving inclination estimation queue.

[0119] The higher the corrected inclination at each time point in the vehicle driving inclination estimation queue is, the greater the risk of vehicle rollover at that time point is. If the inclination at that time point is located in a densely packed area in the vehicle driving inclination estimation queue, it means that the risk of vehicle rollover at that time point is greater and the safety of vehicle driving is not high. Therefore, the higher the vehicle driving criticality determination factor is.

[0120] In a preferred but non-limiting embodiment of the present invention, in Step 4, the higher the vehicle driving criticality determination factor is, the greater the risk of vehicle rollover at that time point is. Therefore, the matching criticality at each time point is calculated by arranging the vehicle driving criticality determination factors, which includes:

[0121] The mean of the vehicle driving criticality determination factors at all time points in the vehicle driving inclination estimation queue is defined as mean 2;

[0122] The amount obtained by dividing the vehicle driving criticality determination factor at each time point in the vehicle driving inclination estimation queue by the mean of two is used as the matching criticality at each time point in the vehicle driving inclination estimation queue.

[0123] The higher the matching criticality, the greater the inclination value at that time point, which means that the risk of vehicle rollover at that time point is greater than the risk of vehicle rollover at other time points in the vehicle driving inclination estimation queue, and the vehicle safety is lower.

[0124] In a preferred but non-limiting embodiment of the present invention, in Step 4, the product value obtained by multiplying the corrected inclination at each time point in the vehicle driving inclination estimation queue by the matching formula key degree at the corresponding time point is obtained, and the amount obtained by accumulating all the product values ​​is divided by the number of elements in the vehicle driving inclination estimation queue, and is used as the vehicle safety determination factor;

[0125] Then, based on the vehicle safety identification factors, the vehicle safety of the building is estimated.

[0126] The vehicle safety identification factors are standardized using the Z-score method. If the standardized value is higher than the pre-defined critical value of the vehicle rollover hazard, then the vehicle driving has a dangerous hazard. The vehicle control system controls the speaker connected to it to issue a voice prompt of the vehicle driving danger to prevent the electric vehicle from rolling over. If the standardized value is not higher than the pre-defined critical value of the vehicle rollover hazard, then the vehicle driving has no dangerous hazard.

[0127] The value of the predefined critical amount of vehicle rollover hazard may be two fifths.

[0128] Based on this, the vehicle safety determination factor is calculated once for each distance and the vehicle safety is estimated.

[0129] Just as every sixty seconds an estimate is made of the vehicle's safety.

[0130] like Figure 2 As shown, a sensor error correction platform for an inertial navigation system according to the present invention comprises:

[0131] A connected vehicle control system and a vehicle inertial sensor, the vehicle inertial sensor is used to sample the vehicle's inertial navigation data and transmit it to the vehicle control system, the vehicle control system is used to perform navigation planning for the vehicle based on the collected vehicle's inertial navigation data, the vehicle's inertial navigation data includes acceleration value, inclination, vibration amount and rotation angle;

[0132] The modules running on the vehicle control system include:

[0133] An acquisition module is used to collect the inertial navigation data of the vehicle sampled by the vehicle inertial sensor and transmitted to obtain the vehicle's pre-estimated inclination and obtain the vehicle's safety inclination queue;

[0134] A classification module, which is used to classify the status of the vehicle safety inclination according to the arrangement of the inclination at each time in the vehicle safety inclination queue, identify the potential distortion factor of the vehicle inclination, and obtain the potential distortion inclination queue;

[0135] A recognition module, which is used to recognize the corrected inclination at each time point in the vehicle safety inclination queue according to the associated arrangement status of the inclination at each time point in each potential distortion inclination queue and the potential distortion factor of the vehicle inclination, and obtain the vehicle driving inclination estimation queue;

[0136] The estimation module is used to estimate the corrected inclination at each time point in the queue based on the vehicle driving inclination, determine the vehicle safety determination factor, and perform estimation on the vehicle driving safety.

[0137] The beneficial effects of the present invention are that, compared with the prior art, the technical effects of the present invention include:

[0138] The autoencoder is used to pre-estimate the vehicle's inclination through the real-time sampled acceleration value, vibration amount, and rotation angle, thereby avoiding the error caused by obtaining the vehicle inclination at only one place through the inclination sensor, which is conducive to improving the accuracy of subsequent vehicle safety estimation; through the vehicle inclination at each time point, the potential distortion factor of the vehicle inclination is determined, and a potential distortion inclination queue is obtained, thereby identifying and distinguishing the inclination with potential distortion; based on the inclination with potential distortion, it is corrected, and the corrected inclination at each time point in the vehicle safety inclination queue is determined to obtain the vehicle driving inclination estimation queue, thereby The slope is corrected to remove the disturbance of potential distortion and improve the accuracy of vehicle safety estimation. The corrected slope at each time point in the queue is estimated based on the vehicle driving inclination in the last period of time, and the matching criticality at each time point is calculated to determine the vehicle safety identification factor. Based on the different criticalities corresponding to the corrected slopes at different time points, the overall vehicle driving safety can be estimated more accurately. Based on the vehicle safety identification factor, the vehicle safety estimation factor is determined, the vehicle safety hazard level is constructed, and the vehicle driving safety is estimated, thereby improving the accuracy of the building electrical vehicle inclination measurement, so as to make the vehicle safety estimation more accurate.

[0139] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that the specific implementation methods of the present invention can still be modified or replaced by equivalents, and any modifications or equivalent replacements that do not deviate from the spirit and scope of the present invention should be covered within the protection space of the claims of the present invention.

Claims

1. A sensor error correction method for an inertial navigation system, characterized in that: include: The vehicle inertial sensor samples the vehicle's inertial navigation data and transmits it to the vehicle control system. The vehicle control system performs navigation planning for the vehicle based on the received vehicle's inertial navigation data. The vehicle's inertial navigation data includes acceleration value, inclination, vibration amount and rotation angle; The vehicle control system also performs the following method on the inertial navigation data of the vehicle it collects: Step 1: Collect the inertial navigation data sampled by the vehicle inertial sensor and transmitted by the vehicle, obtain the vehicle's pre-estimated inclination, and obtain the vehicle's safety inclination queue; Step 2: According to the arrangement and classification of the inclination at each time in the vehicle safety inclination queue, the potential distortion factor of the vehicle inclination is determined to obtain the potential distortion inclination queue; Step 3: According to the relative arrangement of the inclination at each time point in each potential distortion inclination queue and the potential distortion factor of the vehicle inclination, the corrected inclination at each time point in the vehicle safety inclination queue is determined to obtain the vehicle driving inclination estimation queue; Step 4: Estimate the corrected inclination at each time point in the queue based on the vehicle driving inclination, determine the vehicle safety determination factor, and perform an estimate on the vehicle driving safety.

2. The sensor error correction method for an inertial navigation system according to claim 1, characterized in that: In Step 1, the acceleration value, vibration amount, and rotation angle of the vehicle sampled in real time are sent to the autoencoder to perform learning using the autoencoder. The vehicle inertial sensor samples the acceleration value, vibration amount, and rotation angle of the vehicle in real time and transmits them to the autoencoder on the vehicle control system, and the vehicle's pre-estimated inclination is obtained through the autoencoder; In Step 1, the real-time sampling speed is 2 times / second, and the total sampling time is ten minutes. All the pre-estimated inclinations are arranged in the order of the sampling time points and defined as the vehicle safety inclination queue.

3. The sensor error correction method for an inertial navigation system according to claim 2, characterized in that: In Step 2, the vehicle safety tilt queue is divided into several sub-queues, which are defined as each distortion tilt queue; Get the full range of each distortion tilt queue respectively, and form a tilt full range cluster with all the full ranges; By the difference between the full range values, the state expansion and contraction factor is calculated and determined; In Step 2, the method for calculating the expansion and contraction factor of the identified state includes: The amount obtained by multiplying the base position decrement of the tilt range cluster by 2 and dividing it by the total range of the tilt range cluster is defined as a ratio of one; If the ratio one is not lower than the pre-defined recognition threshold, the state expansion factor is determined to be ratio one; if the ratio one is lower than the pre-defined recognition threshold or the full range of the inclination full range cluster is zero, the value of the state expansion factor is determined to be one.

4. The sensor error correction method for an inertial navigation system according to claim 3, characterized in that: In Step 2, the calculation method of the base position reduction of the tilt full-range cluster includes: After arranging the full ranges in the tilt full range cluster in order from low to high to form a queue with p elements, obtain the R1th element and the R2th element in the queue, and the value obtained by subtracting the value of the R1th element from the value of the R2th element and dividing it by two is the base position decrement of the tilt full range cluster; Where: The predefined recognition threshold value is one.

5. The sensor error correction method for an inertial navigation system according to claim 4, characterized in that: In Step 2, the potential distortion factor of vehicle inclination is calculated, which includes: Calculate the mean of the slope range cluster, which is defined as mean 1; The amount obtained by adding the mean value 1 + the mean value 1 of the determined state expansion and contraction factor is determined as the potential distortion factor of the vehicle inclination; The distortion tilt queue of all distortion tilt queues with a full range height exceeding the tilt potential distortion identification factor is defined as the potential distortion tilt queue.

6. The sensor error correction method for an inertial navigation system according to claim 5, characterized in that: In Step 3, according to the difference between the inclination at each time point in each potential distortion inclination queue and the minimum amount of its potential distortion inclination queue, the corresponding difference amount at each time point in each potential distortion inclination queue is determined; The difference between the corresponding difference amount and the potential distortion factor of the vehicle inclination is regarded as the vehicle driving inclination distortion parameter at each time point in each potential distortion inclination queue; Then, according to the vehicle driving inclination distortion parameter, the inclination at each time point in the potential distortion inclination queue is corrected; In Step 3, the method for correcting the tilt at each time point in the potential distortion tilt queue includes: The quantity obtained by applying the modified linear unit to the slope at each time point in each potential distortion slope queue is defined as unit quantity 1; Aggregate the minimum amount of each potential distortion inclination queue, its vehicle inclination potential distortion factor and the unit amount to determine the correction factor at each time point in each potential distortion inclination queue; According to the corrected inclination at each time point, the inclination at the corresponding time point in the vehicle safety inclination queue is replaced to obtain the corrected inclination at each time point in the vehicle safety inclination queue; The vehicle safety inclination queue is formed by selecting the inclination corrected at several time points from the tail of the queue to the head of the queue to form a vehicle driving inclination estimation queue.

7. The sensor error correction method for an inertial navigation system according to claim 6, characterized in that: In Step 3, the corresponding difference amount at each time point in each potential distortion tilt queue is the amount obtained by subtracting the minimum amount of its potential distortion tilt queue from the tilt at each time point in each potential distortion tilt queue; The vehicle driving inclination distortion parameter at each time point in each potential distortion inclination queue is the quantity obtained by subtracting the vehicle inclination potential distortion factor from the corresponding difference quantity.

8. The sensor error correction method for an inertial navigation system according to claim 7, characterized in that: In Step 3, the unit quantity at each time point in each potential distortion inclination queue is multiplied by the potential distortion factor of the vehicle inclination, and the product value of each time point in each potential distortion inclination queue is defined as one, and the product value of one and the minimum quantity of each potential distortion inclination queue are used as the correction factor at each time point in each potential distortion inclination queue; If the vehicle's running inclination distortion parameter is not less than zero, then the inclination corrected at the corresponding time point is the correction factor; if the vehicle's running inclination distortion parameter is less than zero, then the inclination at the corresponding time point maintains the original value; The corrected inclinations at the last 120 time points in the vehicle safety inclination queue are used to form a vehicle driving inclination estimation queue.

9. The sensor error correction method for an inertial navigation system according to claim 8, characterized in that: In Step 4, the vehicle driving criticality determination factors are calculated, which include: According to the difference between the corrected inclination at each time point in the vehicle driving inclination estimation queue and the average of the vehicle driving inclination estimation queue, the actual inclination difference at each time point in the vehicle driving inclination estimation queue is determined; The amount obtained by dividing the corrected inclination at each time point in the vehicle driving inclination estimation queue by the actual inclination difference is used as the vehicle driving criticality determination factor at each time point in the vehicle driving inclination estimation queue; The modulus of the amount obtained by subtracting the mean of the vehicle driving inclination estimation queue from the corrected inclination at each time point in the vehicle driving inclination estimation queue is used as the actual inclination difference at each time point in the vehicle driving inclination estimation queue; In Step 4, the matching criticality at each time point is calculated by arranging the vehicle driving criticality determination factors, which includes: The mean of the vehicle driving criticality determination factors at all time points in the vehicle driving inclination estimation queue is defined as mean 2; The amount obtained by dividing the vehicle driving criticality determination factor at each time point in the vehicle driving inclination estimation queue by the mean of two is used as the matching criticality at each time point in the vehicle driving inclination estimation queue. Furthermore, in Step 4, the product value obtained by multiplying the corrected inclination at each time point in the vehicle driving inclination estimation queue by the matching formula key degree at the corresponding time point is obtained, and the amount obtained by accumulating all the product values ​​is divided by the number of elements in the vehicle driving inclination estimation queue, which is used as the vehicle safety determination factor; Then, based on the vehicle safety identification factor, the vehicle safety of the building is estimated: The vehicle safety identification factors are standardized using the Z-score method. If the standardized value is higher than the pre-defined critical value of the vehicle rollover hazard, then the vehicle driving has a dangerous hazard. The vehicle control system controls the speaker connected to it to issue a voice prompt of the vehicle driving danger to prevent the electric vehicle from rolling over. If the standardized value is not higher than the pre-defined critical value of the vehicle rollover hazard, then the vehicle driving has no dangerous hazard.

10. A sensor error correction platform for an inertial navigation system, characterized in that: include: A connected vehicle control system and a vehicle inertial sensor, the vehicle inertial sensor is used to sample the vehicle's inertial navigation data and transmit it to the vehicle control system, the vehicle control system is used to perform navigation planning for the vehicle based on the collected vehicle's inertial navigation data, the vehicle's inertial navigation data includes acceleration value, inclination, vibration amount and rotation angle; The modules running on the vehicle control system include: An acquisition module is used to collect the inertial navigation data of the vehicle sampled by the vehicle inertial sensor and transmitted to obtain the vehicle's pre-estimated inclination and obtain the vehicle's safety inclination queue; A classification module, which is used to classify the status of the vehicle safety inclination according to the arrangement of the inclination at each time in the vehicle safety inclination queue, identify the potential distortion factor of the vehicle inclination, and obtain the potential distortion inclination queue; A recognition module, which is used to recognize the corrected inclination at each time point in the vehicle safety inclination queue according to the associated arrangement status of the inclination at each time point in each potential distortion inclination queue and the potential distortion factor of the vehicle inclination, and obtain the vehicle driving inclination estimation queue; The estimation module is used to estimate the corrected inclination at each time point in the queue based on the vehicle driving inclination, determine the vehicle safety determination factor, and perform estimation on the vehicle driving safety.

Images