A security analysis system and method based on big data

Abstract

The present invention relates to the technical field of safety analysis, and discloses a safety analysis system and method based on big data, including a main control unit, which calculates a safe following distance according to the currently obtained final adhesion coefficient by the information processing module, and adjusts the speed of the host vehicle according to the distance between the host vehicle and the vehicle in front to control the distance between the host vehicle and the vehicle in front not to exceed the safe following distance. The present invention takes road photos, selects the initial road adhesion coefficient of the current road by comparing the road photos and sample photos using the similarity of image fusion, and compares the initial road adhesion coefficient with the verified adhesion coefficient to narrow the verification calculation range of the initial road adhesion coefficient, thereby shortening the response time of the main control unit, further improving the response speed of the braking system of the vehicle, and strengthening the braking performance during driving.

Description

Technical Field

[0001] The present invention relates to the field of security analysis technology, and in particular to a security analysis system and method based on big data. Background Art

[0002] An autonomous vehicle, also known as a driverless car, a computer-driven car or a wheeled mobile robot, is a smart car that achieves unmanned driving through a computer system. In the vehicle's autonomous driving technology, vehicle following is a core element in the current autonomous driving vehicles.

[0003] When a vehicle is driving automatically, it usually retrieves the driving route from the existing road map (Amap, Baidu Map, etc.) data package based on the starting point and destination. During the driving process, the road conditions will change, and some sections will be modified due to urban construction, resulting in changes in road conditions. In addition, for roads with positive and negative sides (such as mountain roads), if it snows, the snow on the positive side of the road will melt, but there will still be snow on the negative side of the road, and the road conditions will be inconsistent.

[0004] CN201810336352.2 A method, device and system for adjusting the following vehicle status, the method comprising: when it is monitored that the current vehicle starts the adaptive cruise mode, obtaining the weather condition parameters when the current vehicle is traveling at a preset time interval; correcting the adhesion coefficient of the current vehicle according to the weather condition parameters; calculating the braking distance of the current vehicle based on the corrected adhesion coefficient; and adjusting the following vehicle status of the current vehicle in the adaptive cruise mode according to the braking distance.

[0005] However, when selecting the adhesion coefficient based on weather factors, there may be situations where the road surface smoothness is the same under different weather conditions, such as sunny and cloudy days, but the humidity of the road surface is the same. If the coefficient is changed, it will interfere with the safe following distance setting. Therefore, weather factors cannot accurately select the road adhesion coefficient to ensure safe following driving.

[0006] Currently, no effective solution has been proposed for the problems in the related technologies. Summary of the invention

[0007] 1. Technical issues to be resolved

[0008] In view of the shortcomings of the prior art, the present invention provides a safety analysis system and method based on big data, which has the advantages of selecting the adhesion coefficient in combination with the road material and performing secondary verification and selection, thereby solving the problem that the road adhesion coefficient cannot be accurately selected due to weather factors to ensure safe following driving.

[0009] (II) Technical solution

[0010] To solve the technical problem that the above weather factors cannot accurately select the road adhesion coefficient to ensure safe following driving, the present invention provides the following technical solutions:

[0011] A big data-based safety analysis system and method, including:

[0012] S1: Obtain the current road photo, and preliminarily match the current road photo with the sample photos in the sample photo set to obtain the sample photo with the optimal matching degree;

[0013] S2: Traverse the initial adhesion coefficients in the initial adhesion coefficient database, and select the initial adhesion coefficient that maps to the current vehicle speed V, the road name feedback by the sample photo with the optimal matching degree obtained in S1, and the current road humidity d as the current road initial adhesion coefficient;

[0014] Among them, the method for obtaining the initial adhesion coefficient database:

[0015] Obtain the standard initial adhesion coefficient range table, and average each range value in the standard initial adhesion coefficient range table to obtain the initial adhesion coefficient table;

[0016] Take the sample photos corresponding to the road conditions in the standard initial adhesion coefficient range table, and map the sample photos to the road conditions in the standard initial adhesion coefficient range table and the initial adhesion coefficient table to form the initial adhesion coefficient database;

[0017] S3. Obtain the current vehicle tire pressure and temperature, obtain the current verification adhesion coefficient through their relational formula, and judge the correlation between the current road initial adhesion coefficient and the current verification adhesion coefficient μ; if they are relevant, enter step S4; if not, enter S1, and at the same time, the first count value n1 is incremented by 1. When n1 exceeds the first count threshold, the emergency adhesion coefficient is used as the current final adhesion coefficient And enter S5;

[0018] S4: Select the corresponding alternative road conditions from the standard initial adhesion coefficient range table according to μ combined with V and d; perform image recognition on the current road photo, and judge the relevance between the road conditions feedback by the current road photo after image recognition and the corresponding alternative road conditions;

[0019] If there is a correlation, use the associated alternative road conditions as the final road conditions, and select the initial adhesion coefficient that maps to V, d, and the name of the final road conditions as

[0020] If there is no correlation, enter S1, and at the same time, the second count value n2 is incremented by 1. When n2 exceeds the second count threshold, the emergency adhesion coefficient is used as And enter S5;

[0021] S5. According to calculate the current safe following distance of the host vehicle.

[0022] Preferably, determine whether the sample photo with the optimal matching degree reaches a preset standard;

[0023] If it reaches the preset standard, the initial adhesion coefficient of the current road is used as and enter S5;

[0024] If it does not reach the preset standard, enter S3.

[0025] Preferably, the method for obtaining the distance between the current host vehicle and the vehicle ahead is as follows: install an ultrasonic ranging sensor on the host vehicle, and measure the distance between the host vehicle and the vehicle ahead in real time through the ultrasonic ranging sensor;

[0026] The method for obtaining the speed of the host vehicle is as follows: obtain the speed of the host vehicle by installing a speed sensor on the host vehicle;

[0027] The method for obtaining the current road photo and the humidity of the road surface is as follows: obtain the current road photo and the temperature and humidity of the road surface through the in-vehicle camera and the in-vehicle humidity sensor of the host vehicle respectively.

[0028] Preferably, the method for establishing the initial adhesion coefficient database is as follows:

[0029] Conduct on-site tests on the road coefficients of roads with different road conditions, road humidities and host vehicle speeds to form a standard initial adhesion coefficient range table, and take the average of the range values in the standard initial adhesion coefficient range table as the initial adhesion coefficient in the initial adhesion coefficient table;

[0030] Take photos of the roads of each road condition to form sample photos, and the sample photos of the roads with different road conditions form a sample photo set;

[0031] Map the road conditions corresponding to the sample photos in the sample photo set to the initial adhesion coefficient table and the standard initial adhesion coefficient range table to form an initial adhesion coefficient database.

[0032] Preferably, the method for obtaining the initial adhesion coefficient of the current road is as follows:

[0033] Step 1: Obtain the current road photo through the in-vehicle camera of the host vehicle;

[0034] Step 2: Calculate the fusion similarity between the current road photo and each sample photo in the sample photo set;

[0035] Step 3: Select the sample photo with the maximum fusion similarity to the current road photo, and obtain the sample photo name;

[0036] Step 4: Obtain the road condition mapped by the sample photo with the highest fusion similarity to the current road photo in the initial adhesion coefficient table;

[0037] Step 5: Obtain the vehicle speed of the own vehicle and the humidity of the current road surface through a speed sensor and an in-vehicle humidity sensor respectively;

[0038] Step 6: Select the initial adhesion coefficient from the initial adhesion coefficient table according to the road condition mapped by the sample photo with the highest fusion similarity to the current road photo, the vehicle speed of the own vehicle, and the humidity of the current road surface.

[0039] Preferably, the calculation method of the safe following distance is:

[0040] where S is the safe following distance, V is the vehicle speed of the own vehicle, t is the response time of the own vehicle system, is the final road adhesion coefficient.

[0041] Preferably, the method for obtaining the verification adhesion coefficient at the current moment:

[0042] Obtain the current verification adhesion coefficient through the relationship between the vehicle tire pressure, tire temperature, and the verification adhesion coefficient at the current moment, where the relationship between the two is: μ = 0.1πAB + sinAcosB, μ is the verification adhesion coefficient, A is the vehicle tire pressure, B is the tire temperature, and both the vehicle tire pressure A and the tire temperature B are obtained through real-time monitoring by a TPMS tire pressure monitoring sensor.

[0043] Preferably, the method for performing image recognition on the current road photo is:

[0044] Establish neural network models for each road condition, where each neural network model for road condition is used to identify each road condition one by one; and train each neural network model for road condition to obtain each trained neural network model for road condition respectively;

[0045] Use the trained neural network model for road condition corresponding to the alternative road condition as the verification neural network model;

[0046] Input the current road photo into each verification neural network model for recognition respectively, and output each recognized current road photo.

[0047] A safety analysis system based on big data, comprising:

[0048] A distance measuring device, including an ultrasonic ranging sensor, and obtain the distance between the current own vehicle and the vehicle in front through the ultrasonic ranging sensor;

[0049] A speed measuring device, including a speed sensor, and obtain the vehicle speed of the own vehicle through the speed sensor;

[0050] An environment acquisition device, including a camera device and a temperature and humidity measurement device, respectively acquires a current road photo and the temperature and humidity of the road surface through the camera device and the temperature and humidity measurement device;

[0051] A data storage module for storing an initial adhesion coefficient database;

[0052] An information processing module for selecting an initial adhesion coefficient of the current road from the data storage module according to the distance between the host vehicle and the vehicle ahead, the speed of the host vehicle, the current road photo, and the humidity of the road surface, judging the initial adhesion coefficient, obtaining the current final adhesion coefficient, and sending it to the main control unit;

[0053] The main control unit is used to calculate a safe following distance according to the current final adhesion coefficient obtained by the information processing module, and adjust the speed of the host vehicle according to the current distance between the host vehicle and the vehicle ahead to control the distance between the host vehicle and the vehicle ahead not to exceed the safe following distance.

[0054] (III) Beneficial effects

[0055] Compared with the prior art, the present invention provides a safety analysis system and method based on big data, having the following beneficial effects:

[0056] 1. The present invention takes a road photo, selects the initial road adhesion coefficient of the current road by comparing the image fusion similarity between the road photo and the sample photo, and compares the initial road adhesion coefficient with the verification adhesion coefficient to narrow the verification calculation range of the initial road adhesion coefficient, thereby shortening the response time of the main control unit, further improving the response speed of the braking system of the vehicle, and strengthening the braking performance during driving.

[0057] 2. The present invention selects multiple road conditions where the verification adhesion coefficient is located from the standard initial road adhesion coefficient range table through the verification adhesion coefficient, retrieves the road condition neural network model related to the road condition, inputs the taken road photo into the retrieved road condition neural network model related to the road condition, removes impurities from the photo, so that the road condition neural network model outputs the recognized current road photo with only substances of the road condition, calculates the fusion similarity between the recognized current road photo and the sample photo respectively, selects the road condition corresponding to the maximum fusion similarity reaching the set value as the final road condition, and selects the current final adhesion coefficient from the initial road adhesion coefficients according to the final road condition, the current speed of the host vehicle, and the current humidity of the road surface, thereby realizing the secondary verification selection of the road adhesion coefficient. Through verification, the authenticity of the association between the verification adhesion coefficient and the alternative road conditions is improved, effectively avoiding the situation that the verification adhesion coefficient obtained due to accidental factors (such as calculation errors or testing of individual abnormal points on the road surface) is abnormal and the associated alternative road conditions are inaccurate, and further improving the accuracy of the selection of the road adhesion coefficient.

[0058] 3. The present invention calculates the optimal braking distance based on the current ultimate adhesion coefficient and the current vehicle speed of the host vehicle, and adjusts the vehicle speed of the host vehicle according to the current actual distance, thereby ensuring the safety of the vehicle's autonomous driving. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] Figure 1 is the overall control flow chart of the present invention;

[0060] Figure 2 is the flow chart of the initial road coefficient selection method of the present invention;

[0061] Figure 3 is the flow chart of the fusion similarity calculation process method of the present invention;

[0062] Figure 4 is the system hardware structure diagram of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0063] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts fall within the protection scope of the present invention.

[0064] As introduced in the background art, there are deficiencies in the prior art. To solve the above technical problems, the present application proposes a big data-based security analysis system and method.

[0065] Embodiment 1:

[0066] Please refer to Figure 4 , a big data-based security analysis system, including:

[0067] A distance measurement device, including an ultrasonic ranging sensor, to obtain the distance between the current host vehicle and the vehicle in front through the ultrasonic ranging sensor; where the host vehicle is the vehicle itself (i.e., the vehicle for calculating the following distance), and the vehicle in front is the vehicle followed by the host vehicle;

[0068] A speed measurement device, including a speed sensor, to obtain the current vehicle speed of the host vehicle through the speed sensor; by installing a speed sensor on the host vehicle, the speed sensor adopts a single-pulse switch-type Hall effect vehicle speed sensor, which is connected to the driving shaft of the vehicle odometer, and this shaft is connected to the vehicle speed odometer driving turbine shaft at the rear end of the second shaft of the vehicle's gearbox through a flexible shaft. When the vehicle is running, every time the second shaft rotates one week, the speed sensor outputs an electrical pulse, and the pulse voltage signal output by the speed sensor is sent to the information processing module, and the information processing module calculates to obtain the vehicle speed of the host vehicle according to the formula.

[0069] An environment acquisition device, including a camera device and a humidity measurement device, respectively acquires a current road photo and the humidity of the road surface through the camera device and the temperature and humidity measurement device;

[0070] A data storage module for storing an initial adhesion coefficient database;

[0071] An information processing module for matching the current distance between the host vehicle and the vehicle ahead, the current speed of the host vehicle, the current road photo, and the humidity of the current road surface with the initial adhesion coefficients in the initial adhesion coefficient database, selecting the optimally matched initial adhesion coefficient as the current road initial adhesion coefficient, verifying the initial adhesion coefficient to obtain the current final adhesion coefficient, and sending the current final adhesion coefficient to the main control unit;

[0072] A main control unit for calculating a safe following distance according to the current final adhesion coefficient obtained by the information processing module, and adjusting the speed of the host vehicle according to the distance between the current host vehicle and the vehicle ahead to control the distance between the host vehicle and the vehicle ahead not to exceed the safe following distance.

[0073] The main control unit controls the environment acquisition device to regularly acquire the current road photo and the temperature and humidity information of the road surface, and measures the speed of the host vehicle in real time through a speed measurement device. The main control unit compares the acquired current road photo with the initial adhesion coefficient database to obtain the current road condition, and determines the adhesion coefficient of the current road in combination with the humidity of the road surface and the speed of the host vehicle, so as to effectively select an appropriate adhesion coefficient according to the current road condition and the current vehicle speed and other information to improve the accuracy of the parameters related to the road adhesion coefficient during autonomous driving;

[0074] In addition, the main control unit selects an appropriate adhesion coefficient according to the current road condition and the current vehicle speed and other information, calculates the safe following distance of the current host vehicle through the adhesion coefficient and the current speed of the host vehicle, and at the same time obtains the distance between the current host vehicle and the vehicle ahead through an ultrasonic ranging sensor, and adjusts the speed of the host vehicle in real time according to this distance to maintain a safe following distance and prevent the situation of the vehicle behind rear-ending the vehicle ahead due to the vehicle ahead suddenly braking.

[0075] Further, if the distance between the current host vehicle and the vehicle ahead does not exceed the safe following distance, the main control unit controls the host vehicle to decelerate so that the distance between the host vehicle and the vehicle ahead exceeds the safe following distance; if the distance between the current host vehicle and the vehicle ahead exceeds the safe following distance, the main control unit maintains the normal driving operation of the host vehicle.

[0076] The normal driving operation includes accelerating, decelerating, or changing lanes, etc.;

[0077] Further, the camera device and the temperature and humidity measurement device are respectively an in-vehicle camera of the host vehicle and an in-vehicle humidity sensor.

[0078] The camera device may also be other devices connected to the main control unit, such as an external driving recorder, etc. The temperature and humidity measuring device may also be other devices connected to the main control unit. For example, to improve the measurement accuracy, an external sensor is installed on the vehicle chassis to measure the humidity.

[0079] Embodiment 2:

[0080] Please refer to Figures 1-3 , a security analysis method based on big data,

[0081] S1: Obtain the distance between the current vehicle itself and the vehicle in front, the speed of the vehicle itself, the current road photo, and the humidity respectively; where the vehicle itself is the own vehicle, and the vehicle in front is the vehicle being followed.

[0082] Preferably, the way to obtain the distance between the current vehicle itself and the vehicle in front is: install an ultrasonic ranging sensor on the vehicle itself, and measure the distance between the vehicle itself and the vehicle in front in real time through the ultrasonic ranging sensor;

[0083] The way to obtain the speed of the vehicle itself is: obtain the speed of the vehicle itself by installing a speed sensor on the vehicle itself;

[0084] Among them: The specific calculation method of the speed of the vehicle itself is as follows:

[0085]

[0086] Among them, π is the pi, taking 3.14, r is the rolling radius of the tire, n is the number of pulse signals obtained in the time period of ΔT; i is the reduction ratio of the main reducer, and π is a constant, ΔT is the time difference between the current moment and the previous moment, and ΔT takes 1s;

[0087] The way to obtain the current road photo and the temperature and humidity of the road surface is: obtain the current road photo and the humidity of the road surface through the in-vehicle camera and the in-vehicle humidity sensor of the vehicle itself respectively;

[0088] Among them, when driving according to the path in the road information data packet, whenever changing the road or when the road on which the current vehicle is traveling obtained according to GPRS switches from the sunny side to the shady side or from the shady side to the sunny side (such as turning), take a photo of the current road for the selection of the adhesion coefficient. That is, the update of the final road adhesion coefficient occurs when changing the road or when the road switches from the sunny side to the shady side or from the shady side to the sunny side (such as turning).

[0089] S2: Obtain the initial adhesion coefficient database;

[0090] S21: Obtain the standard initial adhesion coefficient range table, and take the average of each range in the standard initial adhesion coefficient range table, or select the smaller value or the minimum end value in each range as the initial adhesion coefficient to obtain the initial adhesion coefficient table;

[0091] Conduct on-site tests on the road coefficients for roads with different road conditions, road humidities, and vehicle speeds of the vehicle itself to form a standard initial adhesion coefficient range table, and take the average of the range values in the standard initial adhesion coefficient range table, or select the smaller value or the minimum end value in each range as the initial adhesion coefficient in the initial adhesion coefficient table;

[0092] Usage method of the initial adhesion coefficient database:

[0093] Obtain the standard initial adhesion coefficient range table (as shown in Table 1), and take the average of each range value in the standard initial adhesion coefficient range table to obtain the initial adhesion coefficient table (as shown in Table 2). For the humidity measured by the temperature and humidity sensor, if the humidity is lower than the set value, the road surface is dry, otherwise it is wet, and the set value is preferably RH30%; measure the vehicle speed of the vehicle itself through the speed sensor, and the vehicle speed is divided at 48 km / h.

[0094] Take multiple photos of each road condition to form multiple sample photos of the road condition, and name the photos of the road condition with the name of the road condition to achieve the mapping of the initial adhesion coefficient table, the standard initial adhesion coefficient range table, and the sample photos.

[0095] Table 1: Standard initial adhesion coefficient range table

[0096]

[0097] Table 2: Initial adhesion coefficient table

[0098]

[0099] S22: Take photos of the road conditions in the standard initial adhesion coefficient range table to form sample photos, and map the sample photos to the road conditions in the standard initial adhesion coefficient range table and the initial adhesion coefficient table to form the initial adhesion coefficient database;

[0100] The sample photos of multiple roads with different road conditions form a sample photo set;

[0101] S3. Take a photo of the current road, preliminarily match the photo of the current road with the sample photos in the sample photo set to obtain the sample photo with the optimal matching degree; judge whether the matching degree of the sample photo with the optimal matching degree reaches the preset standard;

[0102] If the preset standard is met, it is determined that the road condition feedback by the sample photo with the optimal matching degree is the final road condition. Under the conditions of the final road condition, the current vehicle speed of the vehicle itself, and the current road surface humidity, the adhesion coefficient corresponding to the initial road adhesion coefficient table is selected as the current final adhesion coefficient, and then proceed to S8;

[0103] If the preset standard is not met, proceed to S4;

[0104] Among them: The specific method for obtaining the sample photo with the optimal matching degree is as follows,

[0105] By calculating the fusion similarity between the current road photo and each sample photo in the sample photo set, select the sample photo with the largest fusion similarity with the current road photo;

[0106] This embodiment discloses a specific method for obtaining the sample photo with the optimal matching degree. Of course, other methods in the prior art for calculating the similarity between images are also within the protection scope of the present invention:

[0107] S321: Extract the GIST feature value, LBP feature value, HSV feature value, and neural network feature value of the current road photo; extract the GIST feature value, LBP feature value, HSV feature value, and neural network feature value of each sample photo;

[0108] S322: Calculate the similarity between the GIST feature value of the current road photo and the GIST feature value of the sample photo to obtain the GIST feature similarity; calculate the similarity between the LBP feature value of the current road photo and the LBP feature value of the sample photo to obtain the LBP feature similarity; calculate the similarity between the HSV feature value of the current road photo and the HSV feature value of the sample photo to obtain the HSV feature similarity; calculate the similarity between the neural network feature value of the current road photo and the neural network feature value of the sample photo to obtain the neural network feature similarity;

[0109] S323: According to the formula:

[0110] Fusion similarity = GIST feature similarity × M1 + LBP feature similarity × M2 + HSV feature similarity × M3 + neural network feature similarity × M4,

[0111] Calculate the fusion similarity between the current road photo and the sample photo;

[0112] Where 0.1 ≤ M1 ≤ 0.5, 0.1 ≤ M2 ≤ 0.4, 0.1 ≤ M3 ≤ 0.4, 0.1 ≤ M4 ≤ 0.5, and M1 + M2 + M3 + M4 = 1

[0113] S324: Calculate the fusion similarity between each sample photo and the current road photo through steps 2 and 3;

[0114] Select the sample photo with the maximum similarity as the sample photo with the optimal matching degree.

[0115] Further, reaching the preset standard is set as the fusion similarity being not less than 90%.

[0116] S4. Match the current vehicle speed, the sample photo with the optimal matching degree, and the humidity of the current road surface with the initial adhesion coefficients in the initial adhesion coefficient table in the initial adhesion coefficient database to obtain the current initial road adhesion coefficient;

[0117] S5: Verify the reliability of the initial adhesion coefficient;

[0118] Obtain the current vehicle tire pressure A and tire temperature B, and obtain the current verification adhesion coefficient through their relational formula, where the relational formula between the two is: μ = 0.1πAB + sinAcosB, μ is the current verification adhesion coefficient, A is the vehicle tire pressure, B is the tire temperature, and both the vehicle tire pressure A and the tire temperature B are obtained through real-time monitoring by the TPMS tire pressure monitoring sensor.

[0119] If the error between the current verification adhesion coefficient calculated by A and B and the current initial road coefficient is within the set range, the set range is preferably within 10% - 30%, further, the set range is preferably within 25%. Then it is determined that the current initial road adhesion coefficient is reliable, and enter S6; if the error between the current verification adhesion coefficient and the current initial road coefficient is not within the set range, then return to S1 or use the emergency adhesion coefficient as the current final adhesion coefficient, and enter S8;

[0120] In some embodiments, if the error between the current verification adhesion coefficient and the current initial road coefficient is not within the set range, then do not perform the operation of returning to S1, directly use the emergency adhesion coefficient as the current final adhesion coefficient, and enter S8.

[0121] In some embodiments, if the operation of returning to S1 is performed 2 times, and the errors between the 3 times of the current verification adhesion coefficients obtained in total and the current initial road coefficient are not within the set range, then use the emergency adhesion coefficient as the current final adhesion coefficient, and enter S8.

[0122] The error between the current verification adhesion coefficient and the current initial road coefficient = (the current verification adhesion coefficient - the current initial road coefficient) / the absolute value of the current initial road coefficient;

[0123] Among them, the emergency adhesion coefficient is preferably the minimum value of the coefficients in the initial road adhesion coefficient table.

[0124] S6: Select the corresponding alternative road conditions from the standard initial adhesion coefficient range table according to the current verification adhesion coefficient, the current vehicle speed, and the current road surface humidity;

[0125] S7. Identify the current road photo through the road condition neural network model; determine the final road condition corresponding to the current road photo;

[0126] It includes the following steps:

[0127] S71. Establish various road condition neural network models, where one road condition neural network model is used to identify one road condition; and train each road condition neural network model to obtain each trained road condition neural network model respectively;

[0128] S72. Use the trained road condition neural network model corresponding to the alternative road condition as the verification neural network model;

[0129] S73. Input the current road photo into each verification neural network model for identification respectively, and output each identified current road photo;

[0130] S74. Calculate the similarity between each identified current road photo and the sample photo, and determine whether the maximum similarity reaches the set value. If it reaches the set value, it means there is an association between the road condition feedback by the identified current road photo and the corresponding alternative road condition; this set value is preferably 90%; the maximum similarity reaching 90% and above is considered to reach the set value, otherwise, it is not reached.

[0131] To determine whether the maximum similarity reaches the set value, the method of fusing similarity mentioned in S3 can be used for calculation.

[0132] If the maximum similarity reaches the set value, the road condition feedback by the sample photo corresponding to this maximum similarity is the final road condition; enter S8; if the maximum similarity does not reach the set value, return to S1, or use the emergency adhesion coefficient as the current final adhesion coefficient and enter S8;

[0133] In some embodiments, if the maximum similarity does not reach the set value, the operation of returning to S1 is not performed, and the emergency adhesion coefficient is directly used as the current final adhesion coefficient and enter S8.

[0134] In some embodiments, if the operation of returning to S1 is performed 2 times and the maximum similarities obtained in total 3 times do not reach the set value, the emergency adhesion coefficient is used as the current final adhesion coefficient and enter S8.

[0135] S8. Calculate the current safe following distance of the vehicle according to the current final adhesion coefficient;

[0136] The safe following distance is calculated as follows: Where S is the safe following distance, V is the vehicle speed, and t is the vehicle system response time (a fixed parameter of the vehicle itself), is the final road adhesion coefficient, which represents the actual adhesion coefficient between the current tire and the ground; C is 0 or a set value; when the current vehicle brakes emergently, C is 0; when the current vehicle brakes suddenly, C is a set value, preferably 10m or 15m or 30m or 60m.

[0137] Furthermore, the true adhesion coefficient of the current road surface can also be directly measured by setting a friction coefficient sensor as the verification adhesion coefficient at the previous moment.

[0138] Furthermore, the road condition is the road type, including asphalt, gravel, cement, bare soil, etc.

[0139] Furthermore, the training of the neural network model is as follows:

[0140] Taking the asphalt road condition as an example, a set amount of asphalt road condition pictures are obtained as training set photos, and the training set photos are used to train the asphalt neural network model until the loss function converges, that is, the loss function obtains the minimum value, and the training of the asphalt neural network model is completed.

[0141] Similarly, other road condition neural network models in the standard initial adhesion coefficient range table or the initial adhesion coefficient table are trained, and the trained road condition neural network models are stored in the main control unit for waiting to be retrieved.

[0142] Furthermore, when selecting the coefficient according to the road condition in the initial adhesion coefficient table, under the conditions of the current vehicle speed and the current road surface humidity of the current vehicle, select the adhesion coefficient corresponding to the road condition.

[0143] When selecting the alternative road condition in the standard initial road adhesion coefficient range table, in the standard initial road adhesion coefficient range table, taking the current vehicle speed and the current road surface humidity as the preconditions, judge the road condition corresponding to the standard initial road adhesion coefficient range where the current verification adhesion coefficient is located, and take this road condition as the alternative road condition.

[0144] Furthermore, the method for obtaining the initial adhesion coefficient of the current road by matching the current vehicle speed, the sample photo with the optimal matching degree, and the current road surface humidity with the initial adhesion coefficient table in the initial adhesion coefficient database is as follows:

[0145] Step 1: Obtain the current road photo through the in-vehicle camera of the vehicle;

[0146] Step 2: Calculate the fusion similarity between the current road photo and each sample photo in the sample photo set;

[0147] Step 3: Select the sample photo with the largest fusion similarity to the current road photo and obtain the sample photo name;

[0148] Step 4: Obtain the road condition mapped by the sample photo with the highest fusion similarity to the current road photo in the initial adhesion coefficient table;

[0149] Step 5: Obtain the vehicle speed of the vehicle itself and the humidity of the current road surface through a speed sensor and an in-vehicle humidity sensor respectively;

[0150] Step 6: Select the initial adhesion coefficient from the initial adhesion coefficient table according to the road condition mapped by the sample photo with the highest fusion similarity to the current road photo, the vehicle speed of the vehicle itself, and the humidity of the current road surface.

[0151] The present invention takes a road photo, compares the road photo and the sample photo by image fusion similarity to select the initial road adhesion coefficient of the current road, and compares the initial road adhesion coefficient with the verification adhesion coefficient to narrow the verification calculation range of the initial road adhesion coefficient, thereby shortening the response time of the main control unit, further improving the response speed of the vehicle's braking system, and strengthening the braking performance during driving;

[0152] Select multiple road conditions where the verification adhesion coefficient is located from the standard initial road adhesion coefficient range table according to the verification adhesion coefficient, retrieve the road condition neural network model related to this road condition, input the taken road photo into the retrieved road condition neural network model related to this road condition, remove impurities from the photo, so that the road condition neural network model outputs the recognized current road photo with only the substances of this road condition, calculate the fusion similarity between the recognized current road photo and the sample photo respectively, select the road condition corresponding to the maximum fusion similarity reaching the set value as the final road condition, and select the current final adhesion coefficient from the initial road adhesion coefficient according to the final road condition, the current vehicle speed of the vehicle itself, and the humidity of the current road surface, thereby realizing the secondary verification selection of the road adhesion coefficient and improving the accuracy of the road adhesion coefficient;

[0153] Calculate the optimal braking distance according to the current final adhesion coefficient and the current vehicle speed of the vehicle itself, and adjust the vehicle speed of the vehicle itself according to the current actual distance, thereby ensuring the safety of vehicle autonomous driving.

[0154] Although the embodiments of the present invention have been shown and described, for those of ordinary skill in the art, it can be understood that various changes, modifications, substitutions, and variations can be made to these embodiments without departing from the principle and spirit of the present invention. The scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A security analysis method based on big data, characterized in that: S1: Obtain the current road photo, and preliminarily match the current road photo with the sample photos in the sample photo set to obtain the sample photo with the optimal matching degree; S2: Traverse the initial adhesion coefficient in the initial adhesion coefficient database, and select the initial adhesion coefficient mapped to the current vehicle speed V, the road name reflected by the sample photo with the optimal matching degree obtained in S1, and the current road humidity d as the current road initial adhesion coefficient; Among them, the method for obtaining the initial adhesion coefficient database: Obtain the standard initial adhesion coefficient range table, and average each range value in the standard initial adhesion coefficient range table to obtain the initial adhesion coefficient table; Take the sample photos corresponding to the road conditions in the standard initial adhesion coefficient range table, and map the sample photos to the road conditions in the standard initial adhesion coefficient range table and the initial adhesion coefficient table to form the initial adhesion coefficient database; S3. Obtain the current tire pressure and temperature of the vehicle, and obtain the current verified adhesion coefficient through the relationship between the two and the current verified adhesion coefficient. Determine the correlation between the current initial road adhesion coefficient and the current verified adhesion coefficient μ. If they are correlated, go to step S4; if not, go to S1, and at the same time, increment the first count value n1 by 1. When n1 exceeds the first count threshold, use the emergency adhesion coefficient as the current final adhesion coefficient φ , and enter S5; S4: According to μ combined with V and d, select the corresponding alternative road conditions from the standard initial adhesion coefficient range table; perform image recognition on the current road photo, and judge the relevance between the road conditions reflected by the current road photo after image recognition and the corresponding alternative road conditions; φ If there is an association, the associated alternative road condition is taken as the final road condition, and the initial adhesion coefficient mapped to V, d, and the name of the final road condition is selected as φ ; If there is no association, enter S1, and at the same time increment the second count value n2 by 1. When n2 exceeds the second count threshold, use the emergency adhesion coefficient as φ , and enter S5; S5. According to φ Calculate the current safe following distance of the host vehicle.

2. The security analysis method based on big data according to claim 1, characterized in that: Judge whether the sample photo with the optimal matching degree reaches the preset standard; If the preset standard is met, the initial adhesion coefficient of the current road is used as φ , and it enters S5; if the preset standard is not met, it enters S3.

3. The security analysis method based on big data according to claim 1, wherein: The method for obtaining the distance between the current vehicle and the vehicle in front: Install an ultrasonic ranging sensor on the vehicle, and measure the distance between the vehicle and the vehicle in front in real time through the ultrasonic ranging sensor; The method for obtaining the vehicle speed of the vehicle itself: Obtain the vehicle speed of the vehicle itself by installing a speed sensor on the vehicle; The method for obtaining the current road photo and the humidity of the road surface: Obtain the current road photo and the humidity of the road surface through the vehicle-mounted camera and the vehicle-mounted humidity sensor of the vehicle respectively.

4. A security analysis method based on big data according to claim 1, characterized in that: The method for establishing the initial adhesion coefficient database: Conduct on-site tests on the road coefficients of roads with different road conditions, road humidities and vehicle speeds to form a standard initial adhesion coefficient range table, and average the range values in the standard initial adhesion coefficient range table as the initial adhesion coefficient in the initial adhesion coefficient table; Take photos of the roads with each road condition to form sample photos, and the sample photos of the roads with different road conditions form a sample photo set; Map the road conditions corresponding to the sample photos in the sample photo set to the initial adhesion coefficient table and the standard initial adhesion coefficient range table to form the initial adhesion coefficient database.

5. A security analysis method based on big data according to claim 1, characterized in that: The method for obtaining the current road initial adhesion coefficient is: Step 1: Obtain the current road photo through the vehicle-mounted camera of the vehicle; Step 2: Calculate the fusion similarity between the current road photo and each sample photo in the sample photo set; Step 3: Select the sample photo with the largest fusion similarity with the current road photo, and obtain the sample photo name; Step 4: Obtain the road condition mapped by the sample photo name with the largest fusion similarity with the current road photo in the initial adhesion coefficient table; Step 5: Obtain the vehicle speed of the vehicle itself and the humidity of the current road surface through the speed sensor and the vehicle-mounted humidity sensor respectively; Step 6: Select the initial adhesion coefficient from the initial adhesion coefficient table according to the road conditions, the speed of the host vehicle, and the humidity of the current road surface mapped by the sample photo with the highest fusion similarity to the current road photo.

6. A security analysis method based on big data according to claim 1, characterized in that: The calculation method of the safe following distance is as follows: , where S is the safe following distance, V is the speed of the host vehicle, t is the response time of the host vehicle system, ϕ is the final road adhesion coefficient.

7. A security analysis method based on big data according to claim 1, characterized in that: Method for obtaining the verified adhesion coefficient at the current moment: Obtain the current verified adhesion coefficient through the relationship formula of the vehicle tire pressure, tire temperature, and the verified adhesion coefficient at the current moment. The relationship formula between the two is: μ = 0.1πAB + sinAcosB, where μ is the verified adhesion coefficient, A is the vehicle tire pressure, and B is the tire temperature. Both the vehicle tire pressure A and the tire temperature B are obtained through real-time monitoring by the TPMS tire pressure monitoring sensor.

8. According to the method for safety analysis based on big data described in claim 1, wherein: The method for image recognition of the current road photo is as follows: Establish various road condition neural network models, where each road condition neural network model is used to identify various road conditions one by one; and train each road condition neural network model to obtain each trained road condition neural network model respectively. Use the trained road condition neural network model corresponding to the alternative road condition as the verification neural network model. Input the current road photo into each verification neural network model for recognition respectively, and output each recognized current road photo.

9. A system applying the method for safety analysis based on big data according to any one of claims 1-8, wherein: A distance measurement device, including an ultrasonic ranging sensor, which obtains the distance between the current host vehicle and the vehicle ahead through the ultrasonic ranging sensor. A speed measurement device, including a speed sensor, which obtains the speed of the host vehicle through the speed sensor. An environment acquisition device, including a camera device and a humidity measurement device, which obtains the current road photo and the humidity of the road surface through the camera device and the temperature and humidity measurement device respectively. A data storage module, which is used to store the initial adhesion coefficient database. An information processing module, which is used to select the initial adhesion coefficient of the current road from the data storage module according to the distance between the host vehicle and the vehicle ahead, the speed of the host vehicle, the current road photo, and the humidity of the road surface, and judge the initial adhesion coefficient to obtain the current final adhesion coefficient and send it to the main control unit. The main control unit is used to calculate the safe following distance according to the current final adhesion coefficient obtained by the information processing module, and adjust the speed of the host vehicle according to the current distance between the host vehicle and the vehicle ahead to control the distance between the host vehicle and the vehicle ahead not to exceed the safe following distance. ​

Images