Low-adhesion road surface recognition method and device, medium and electronic equipment

Abstract

The invention discloses a low-adhesion road surface recognition method and device, a medium and electronic equipment, and the method comprises the steps: obtaining an early warning probability that a road surface where a vehicle is located is a low-adhesion road surface caused by the road surface recognition data for each kind of road surface recognition data, and carrying out the early warning of the vehicle in combination with each early warning probability, the total early warning probability that the road surface where the vehicle is located is a low-adhesion road surface is determined; and under the condition that the total early warning probability is greater than or equal to a preset probability threshold value, acquiring a preset weighting coefficient of the road surface identification data for each type of road surface identification data, determining whether the road surface where the vehicle is located is a low-adhesion road surface according to each type of road surface identification data and the corresponding preset weighting coefficient, and if yes, performing early warning on the road surface where the vehicle is located. The road surface where the vehicle is located is a low-adhesion road surface. According to the invention, the accuracy of low-adhesion road surface recognition can be improved.

Description

Technical Field

[0001] This application relates to the technical field of road surface recognition, and in particular, to a method, device, medium, and electronic device for recognizing a low-adhesion road surface. Background Art

[0002] Currently, mainly through a single method such as a wheel speed sensor, weather forecast, rain sensor, or front probe, it is recognized whether a vehicle is on a low-adhesion road surface, and its accuracy is not high, and misjudgment is likely to occur in some cases. Summary of the Invention

[0003] Embodiments of this application provide a method, device, medium, and electronic device for recognizing a low-adhesion road surface, which are used to solve the technical problem of low accuracy of existing low-adhesion road surface recognition methods.

[0004] Other features and advantages of this application will become apparent through the following detailed description, or will be partially learned through the practice of this application.

[0005] According to a first aspect of this application, a method for recognizing a low-adhesion road surface is provided. The method includes:

[0006] Obtain various road surface recognition data of the vehicle;

[0007] For each of the road surface recognition data, obtain the warning probability that the road surface recognition data causes the road surface where the vehicle is located to be a low-adhesion road surface. Combine each of the warning probabilities to determine the total warning probability that the road surface where the vehicle is located is a low-adhesion road surface;

[0008] In the case where the total warning probability is greater than or equal to a preset probability threshold, for each of the road surface recognition data, obtain the preset weighting coefficient of the road surface recognition data, and determine whether the road surface where the vehicle is located is a low-adhesion road surface according to each of the road surface recognition data and the corresponding preset weighting coefficient. If so, the road surface where the vehicle is located is a low-adhesion road surface.

[0009] In some embodiments, based on the foregoing solution, the obtaining of the warning probability that the road surface recognition data causes the road surface where the vehicle is located to be a low-adhesion road surface includes:

[0010] Obtain a preset probability condition;

[0011] If the road surface recognition data meets the preset probability condition, the warning probability of the road surface recognition data is a first probability value. If it does not meet, the warning probability of the road surface recognition data is a second probability value. Both the first probability value and the second probability value are greater than or equal to 0. The first probability value is greater than the second probability value, and the sum of the first probability value and the second probability value is 100%;

[0012] Determining the total warning probability that the road surface where the vehicle is located is a low - adhesion road surface by combining each of the foregoing warning probabilities includes:

[0013] Obtaining the prior probability that the road surface where the vehicle is located is a low - adhesion road surface;

[0014] Determining the posterior probability that the road surface where the vehicle is located is a low - adhesion road surface according to each of the warning probabilities and the prior probability, and using the posterior probability as the total warning probability.

[0015] In some embodiments, based on the foregoing solution, the obtaining of the preset probability condition includes:

[0016] When the road surface recognition data is slip ratio and acceleration, taking the slip ratio being greater than the first slip ratio threshold and the acceleration being less than the first acceleration threshold as the preset probability condition;

[0017] When the road surface recognition data is rainfall data, taking the rainfall data being greater than the first rainfall threshold as the preset probability condition;

[0018] When the road surface recognition data is the road surface image data and the vehicle image data in front of the vehicle, determining the first road surface recognition result of the road surface image data and the vehicle image data for the road surface where the vehicle is located, and taking the first road surface recognition result being a low - adhesion road surface risk as the preset probability condition;

[0019] When the road surface recognition data is weather forecast data, taking the weather forecast data being an icing forecast, a rain forecast or a snow forecast as the preset probability condition.

[0020] In some embodiments, based on the foregoing solution, the determining whether the road surface where the vehicle is located is a low - adhesion road surface according to each of the road surface recognition data and the corresponding preset weighting coefficients includes:

[0021] Obtaining a preset numerical threshold;

[0022] For each type of the road surface recognition data, obtaining a preset recognition condition. If the road surface recognition data meets the preset recognition condition, taking a first value as an intermediate value; if not, taking a second value as the intermediate value, where the first value is greater than the second value. Weighting the intermediate value based on the preset weighting coefficient to obtain a target value;

[0023] Summing up the respective target values to obtain a total target value;

[0024] When the total target value is greater than the preset numerical threshold, the road surface where the vehicle is located is a low - adhesion road surface.

[0025] In some embodiments, based on the foregoing solution, the obtaining of the preset recognition condition includes:

[0026] When the road surface recognition data is the slip ratio and the acceleration, determining a first duration for which the slip ratio is greater than a second slip ratio threshold, and taking the first duration being greater than a first duration threshold and the vehicle acceleration being less than a second acceleration threshold as the preset recognition condition;

[0027] When the road surface recognition data is rainfall data, determining a second duration for which the rainfall data is greater than a second rainfall threshold, and taking the second duration being greater than a second duration threshold as the preset recognition condition;

[0028] When the road surface recognition data is road surface image data and vehicle image data in front of the vehicle, determining a second road surface recognition result of the road surface where the vehicle is located from the road surface image data and the vehicle image data, and taking the second road surface recognition result being a low - adhesion road surface and the confidence level being greater than a confidence level threshold as the preset recognition condition;

[0029] When the road surface recognition data is weather forecast data, taking the weather forecast data being an icing forecast, a rain forecast or a snow forecast as the preset recognition condition.

[0030] In some embodiments, based on the foregoing solution, the road surface image data includes first road surface data collected by a camera and second road surface data collected by a radar, and the determining of the second recognition result of the road surface where the vehicle is located from the road surface image data and the vehicle image data includes:

[0031] Identifying the characteristics of the road surface where the vehicle is located according to the first road surface data to obtain a first recognition result;

[0032] Determining the lateral shift standard deviation and the brake light flashing frequency of the target vehicle according to the vehicle image data;

[0033] Identifying the road surface where the vehicle is located according to the magnitude of the lateral shift standard deviation and / or the magnitude of the brake light flashing frequency to obtain a second recognition result;

[0034] Determining the signal reflection condition of the radar signal on the road surface where the vehicle is located according to the second road surface data;

[0035] Identifying the road surface where the vehicle is located according to the signal reflection condition to obtain a third recognition result;

[0036] Combining the first recognition result, the second recognition result and the third recognition result to obtain the second road surface recognition result.

[0037] In some embodiments, based on the foregoing solution, it further includes:

[0038] Obtain a variety of environmental parameters of the vehicle, where the variety of environmental parameters includes environmental temperature, environmental time, navigation system positioning error, surrounding vehicles, and regional environment;

[0039] Adjust the preset weighting coefficient according to the variety of environmental parameters.

[0040] According to a second aspect of the present application, there is provided a low adhesion road surface recognition device, and the device includes:

[0041] A first acquisition unit configured to acquire a variety of road surface recognition data of the vehicle;

[0042] A first determination unit configured to, for each of the road surface recognition data, obtain a warning probability that the road surface recognition data causes the road surface where the vehicle is located to be a low adhesion road surface, and combine the warning probabilities to determine a total warning probability that the road surface where the vehicle is located is a low adhesion road surface;

[0043] A second determination unit configured to, when the total warning probability is greater than or equal to a preset probability threshold, for each of the road surface recognition data, obtain a preset weighting coefficient of the road surface recognition data, and determine whether the road surface where the vehicle is located is a low adhesion road surface according to each of the road surface recognition data and the corresponding preset weighting coefficient. If so, the road surface where the vehicle is located is a low adhesion road surface.

[0044] In some embodiments, based on the foregoing solution, the first determination unit includes a second acquisition unit and a third determination unit. The second acquisition unit is configured to acquire a warning probability that the road surface recognition data causes the road surface where the vehicle is located to be a low adhesion road surface, and the third determination unit is configured to combine the warning probabilities to determine a total warning probability that the road surface where the vehicle is located is a low adhesion road surface. The second acquisition unit is configured as follows:

[0045] A third acquisition unit configured to acquire a preset probability condition;

[0046] A fourth determination unit configured to, if the road surface recognition data meets the preset probability condition, the warning probability of the road surface recognition data is a first probability value, and if not, the warning probability of the road surface recognition data is a second probability value. Both the first probability value and the second probability value are greater than or equal to 0, the first probability value is greater than the second probability value, and the sum of the first probability value and the second probability value is 100%;

[0047] The third determination unit is configured as follows:

[0048] A fourth acquisition unit configured to acquire a prior probability that the road surface where the vehicle is located is a low adhesion road surface;

[0049] A fifth determination unit determines a posterior probability that the road surface where the vehicle is located is a low-adhesion road surface according to each of the warning probabilities and the prior probability, and uses the posterior probability as the total warning probability.

[0050] In some embodiments, based on the foregoing solution, the third acquisition unit is configured as follows:

[0051] A first acting unit, when the road surface recognition data is a slip ratio and an acceleration, uses the slip ratio being greater than a first slip ratio threshold and the acceleration being less than a first acceleration threshold as the preset probability condition;

[0052] A second acting unit, when the road surface recognition data is rainfall data, uses the rainfall data being greater than a first rainfall threshold as the preset probability condition;

[0053] A third acting unit, when the road surface recognition data is road surface image data and vehicle image data in front of the vehicle, determines a first road surface recognition result of the road surface where the vehicle is located for the road surface image data and the vehicle image data, and uses the first road surface recognition result being a low-adhesion road surface risk as the preset probability condition;

[0054] A fourth acting unit, when the road surface recognition data is weather forecast data, uses the weather forecast data being an ice warning, a rain warning or a snow warning as the preset probability condition.

[0055] In some embodiments, based on the foregoing solution, the second determination unit includes a sixth determination unit, and the sixth determination unit is used to determine whether the road surface where the vehicle is located is a low-adhesion road surface according to each of the road surface recognition data and the corresponding preset weighting coefficient. The sixth determination unit is configured as follows:

[0056] A fifth acquisition unit acquires a preset numerical threshold;

[0057] A fifth acting unit, for each type of the road surface recognition data, acquires a preset recognition condition. If the road surface recognition data meets the preset recognition condition, a first numerical value is used as an intermediate numerical value. If not, a second numerical value is used as the intermediate numerical value. The first numerical value is greater than the second numerical value. A target numerical value is obtained by weighting the intermediate numerical value based on the preset weighting coefficient;

[0058] A first obtaining unit sums up each of the target numerical values to obtain a total target numerical value;

[0059] A sixth acting unit, when the total target numerical value is greater than the preset numerical threshold, the road surface where the vehicle is located is a low-adhesion road surface.

[0060] In some embodiments, based on the foregoing solution, the fifth acting unit includes a sixth obtaining unit, and the sixth obtaining unit is configured to obtain a preset recognition condition, and the sixth obtaining unit is configured as follows:

[0061] A seventh acting unit, when the road surface recognition data is the slip ratio and the acceleration, determines a first duration for which the slip ratio is greater than a second slip ratio threshold, and takes the first duration being greater than a first duration threshold and the vehicle acceleration being less than a second acceleration threshold as the preset recognition condition;

[0062] An eighth acting unit, when the road surface recognition data is rainfall data, determines a second duration for which the rainfall data is greater than a second rainfall threshold, and takes the second duration being greater than a second duration threshold as the preset recognition condition;

[0063] A ninth acting unit, when the road surface recognition data is the road surface image data and the vehicle image data in front of the vehicle, determines a second road surface recognition result of the road surface where the vehicle is located from the road surface image data and the vehicle image data, and takes the second road surface recognition result being a low-adhesion road surface and the confidence level being greater than a confidence level threshold as the preset recognition condition;

[0064] A tenth acting unit, when the road surface recognition data is weather forecast data, takes the weather forecast data being an icing forecast, a rain forecast or a snow forecast as the preset recognition condition.

[0065] In some embodiments, based on the foregoing solution, the road surface image data includes first road surface data collected by a camera and second road surface data collected by a radar, and the ninth acting unit includes a seventh determination unit, and the seventh determination unit is configured to determine a second recognition result of the road surface where the vehicle is located from the road surface image data and the vehicle image data, and the seventh determination unit is configured as follows:

[0066] A second obtaining unit, recognizes the characteristics of the road surface where the vehicle is located according to the first road surface data, and obtains a first recognition result;

[0067] An eighth determination unit, determines the lateral movement standard deviation and the brake light flashing frequency of the target vehicle according to the vehicle image data;

[0068] A third obtaining unit, recognizes the road surface where the vehicle is located according to the magnitude of the lateral movement standard deviation and / or the magnitude of the brake light flashing frequency, and obtains a second recognition result;

[0069] A ninth determination unit, determines the signal reflection condition of the radar signal on the road surface where the vehicle is located according to the second road surface data;

[0070] A fourth obtaining unit, which identifies the road surface where the vehicle is located according to the signal reflection condition, and obtains a third identification result;

[0071] A fifth obtaining unit, which combines the first identification result, the second identification result and the third identification result to obtain the second road surface identification result.

[0072] In some embodiments, based on the foregoing solution, it further includes:

[0073] A seventh obtaining unit, which obtains various environmental parameters of the vehicle, and the various environmental parameters include environmental temperature, environmental time, positioning error of the navigation system, surrounding vehicles and regional environment;

[0074] A first adjusting unit, which adjusts the preset weighting coefficient according to the various environmental parameters.

[0075] According to a third aspect of the present application, there is provided a computer-readable storage medium, on which a computer program is stored, and the computer program includes executable instructions, and when the executable instructions are executed by a processor, the method described in any embodiment of the first aspect of the present application is implemented.

[0076] According to a fourth aspect of the present application, it includes: one or more processors; a memory for storing executable instructions of the processor, and when the executable instructions are executed by the one or more processors, the one or more processors implement the method described in any embodiment of the first aspect of the present application.

[0077] The beneficial effects of the present application are as follows:

[0078] First, determine the warning probability that each road surface identification data causes the road surface where the vehicle is located to be a low-adhesion road surface, and combine the warning probabilities to determine the total warning probability that the road surface where the vehicle is located is a low-adhesion road surface. Secondly, when the total warning probability is greater than or equal to a preset probability threshold, determine whether the road surface where the vehicle is located is a low-adhesion road surface according to each road surface identification data and the corresponding preset weighting coefficient. On the one hand, identify the road surface where the vehicle is located by combining multiple road surface identification data. On the other hand, through two road surface identifications, the accuracy of road surface identification is improved.

[0079] It should be understood that the above general description and the following detailed description are only exemplary and explanatory, and cannot limit the present application. Description of the Drawings

[0080] Figure 1 Shows a flowchart of a method for identifying a low-adhesion road surface in an embodiment of the present application;

[0081] Figure 2 Shows a block diagram of a device for identifying a low-adhesion road surface in an embodiment of the present application;

[0082] Figure 3 Shows a schematic diagram of a computer-readable storage medium in an embodiment of the present application;

[0083] Figure 4 Shows a schematic diagram of the system structure of an electronic device in an embodiment of the present application. Detailed implementation manners

[0084] Next, the technical solutions in the embodiments of the present application will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present application.

[0085] Figure 1 Shows a flowchart of a low-adhesion road surface recognition method in an embodiment of the present application. Refer to Figure 1 , provides a low-adhesion road surface recognition method, which at least includes S1 to S3, and is introduced in detail as follows:

[0086] In step S1, various road surface recognition data of the vehicle are obtained, and each road surface recognition data can be used to identify the road surface where the vehicle is located respectively.

[0087] In step S2, for each of the road surface recognition data, the warning probability that the road surface recognition data causes the road surface where the vehicle is located to be a low-adhesion road surface is obtained, and in combination with each of the warning probabilities, the total warning probability that the road surface where the vehicle is located is a low-adhesion road surface is determined.

[0088] In step S3, when the total warning probability is greater than or equal to a preset probability threshold, for each of the road surface recognition data, the preset weighting coefficient of the road surface recognition data is obtained, and based on each of the road surface recognition data and the corresponding preset weighting coefficient, it is determined whether the road surface where the vehicle is located is a low-adhesion road surface. If so, the road surface where the vehicle is located is a low-adhesion road surface.

[0089] In some implementation manners, after determining whether the road surface where the vehicle is located is a low-adhesion road surface based on each of the road surface recognition data and the corresponding preset weighting coefficient, the method further includes: if not, the road surface where the vehicle is located is not a low-adhesion road surface.

[0090] In some embodiments, obtaining the road surface recognition data to cause a warning probability that the road surface where the vehicle is located is a low - adhesion road surface includes: obtaining a preset probability condition; if the road surface recognition data meets the preset probability condition, the warning probability of the road surface recognition data is a first probability value, and if it does not meet, the warning probability of the road surface recognition data is a second probability value. Both the first probability value and the second probability value are greater than or equal to 0, the first probability value is greater than the second probability value, and the sum of the first probability value and the second probability value is 100%; combining each of the warning probabilities to determine the total warning probability that the road surface where the vehicle is located is a low - adhesion road surface includes: obtaining a prior probability that the road surface where the vehicle is located is a low - adhesion road surface; determining a posterior probability that the road surface where the vehicle is located is a low - adhesion road surface according to each of the warning probabilities and the prior probability, and using the posterior probability as the total warning probability.

[0091] In some embodiments, determining the posterior probability that the road surface where the vehicle is located is a low - adhesion road surface according to each of the warning probabilities and the prior probability includes: multiplying each of the warning probabilities and the prior probability to obtain a first target total probability; obtaining a preset total probability; using the ratio of the first target total probability and the preset total probability as the posterior probability.

[0092] In some embodiments, obtaining the preset total probability includes: for each type of the road surface recognition data, obtaining a set probability that meets the preset probability condition when the road surface where the vehicle is located is a non - low - adhesion road surface; using the difference between 1 and the prior probability as a target probability; multiplying each of the set probabilities and the target probability to obtain a second target total probability; adding the first target total probability and the second target total probability to obtain the preset total probability.

[0093] Exemplarily, the preset probability threshold is 0.8. When the road surface recognition data is the slip ratio, the preset probability condition is that the slip ratio is greater than 20%, the first probability value is 0.75, the second probability value is 0.25. If the slip ratio is greater than 20%, the set probability is 0.1. If the slip ratio is less than or equal to 20%, the set probability is 0.9; when the road surface recognition data is image data, the preset probability condition is that a wet road surface is detected based on the image data. The image data can be collected by a camera. The first probability value is 0.65, the second probability value is 0.35. If a wet road surface is detected based on the image data, the set probability is 0.05. If a wet road surface is not detected based on the image data, the set probability is 0.95; when the road surface recognition data is rainfall data, the preset probability condition is that the rainfall data is greater than 5 mm / h. The first probability value is 0.6, the second probability value is 0.4. If the rainfall data is greater than 5 mm / h, the set probability is 0.1. If the rainfall data is less than or equal to 5 mm / h, the set probability is 0.9.

[0094] When the slip ratio is greater than 20%, a wet road surface is detected based on the image data, and the rainfall data is greater than 5 mm / h, the first target total probability is 0.75×0.65×0.6×0.05 = 0.14625, the second target total probability is 0.1×0.05×0.1×0.95 = 0.000475, the preset total probability is 0.14625 + 0.000475 = 0.146725, and the posterior probability is 0.14625÷0.146725 = 0.9968, that is, the posterior probability is greater than the preset probability threshold 0.8.

[0095] When the slip ratio is less than or equal to 20%, a wet road surface is detected based on the image data, and the rainfall data is greater than 5 mm / h, the first target total probability is 0.25×0.65×0.6×0.05 = 0.004875, the second target total probability is 0.9×0.05×0.1×0.95 = 0.004275, the preset total probability is 0.004875 + 0.004275 = 0.00915, and the posterior probability is 0.004875÷0.00915 = 0.5328, that is, the posterior probability is less than the preset probability threshold 0.8.

[0096] When the slip rate is less than or equal to 20%, no wet road surface is detected according to the image data, and the rainfall data is greater than 5 mm / h, the first target total probability is 0.25×0.35×0.6×0.05 = 0.002625, the second target total probability is 0.9×0.95×0.1×0.95 = 0.081225, the preset total probability is 0.081225 + 0.002625 = 0.08385, and the posterior probability is 0.002625÷0.08385 = 0.0313, that is, the posterior probability is less than the preset probability threshold of 0.8.

[0097] In the present application, obtaining the prior probability that the road surface where the vehicle is located is a low - adhesion road surface; determining the posterior probability that the road surface where the vehicle is located is a low - adhesion road surface according to each of the warning probabilities and the prior probability can be understood as calculating the posterior probability according to Bayes' theorem.

[0098] In some embodiments, obtaining the preset probability conditions includes: when the road surface recognition data is the slip rate and acceleration, taking the slip rate being greater than the first slip rate threshold and the acceleration being less than the first acceleration threshold as the preset probability conditions; when the road surface recognition data is rainfall data, taking the rainfall data being greater than the first rainfall threshold as the preset probability conditions; when the road surface recognition data is the road surface image data and vehicle image data in front of the vehicle, determining the first road surface recognition result of the road surface image data and the vehicle image data for the road surface where the vehicle is located, and taking the first road surface recognition result being at risk of a low - adhesion road surface as the preset probability conditions; when the road surface recognition data is weather forecast data, taking the weather forecast data being an ice warning, a rain warning or a snow warning as the preset probability conditions. The acceleration can be detected by a longitudinal acceleration sensor of the vehicle. Combining the acceleration and the slip rate to identify the road surface where the vehicle is located. On the one hand, if the slip rate is high but the acceleration is low, it indicates that the driving force is not effectively transmitted, supporting the determination of low adhesion, that is, supporting the determination that the road surface where the vehicle is located is a low - adhesion road surface. On the other hand, if the slip rate is greater than the first slip rate threshold but the acceleration is greater than the first acceleration threshold, it is determined that the vehicle is driving normally, that is, the road surface where the vehicle is located is not a low - adhesion road surface, thereby eliminating misjudgment and improving the judgment accuracy.

[0099] Exemplarily, the first slip rate threshold is 20%, and the first acceleration threshold is 0.3g, where g is the acceleration due to gravity. If the throttle opening of the vehicle is greater than 50% (corresponding to a slip rate greater than 20%) and the acceleration is less than 0.3g, it indicates that the driving force is not effectively transmitted, supporting the determination of low adhesion. When the vehicle accelerates rapidly, if the acceleration is greater than 0.5g, that is, greater than the first acceleration threshold, even if the slip rate is greater than 20%, it is determined that the vehicle is driving normally.

[0100] In some embodiments, the first slip rate threshold is 20%. On a dry road surface, if the vehicle does not accelerate rapidly, the slip rate is less than 5%. If the vehicle accelerates rapidly, the slip rate can reach 10%. Based on the ISO 21994 standard and the actual vehicle test data, on a low-adhesion road surface, the slip rate is greater than 20%. Therefore, the first slip rate threshold is set to 20%.

[0101] In some embodiments, it further includes: adjusting the first slip rate threshold according to the slope of the road surface where the vehicle is located and the load of the vehicle. Among them, the greater the slope or the load, the greater the first slip rate threshold, that is, the magnitude of the slope is positively correlated with the magnitude of the first slip rate threshold, and the magnitude of the load is positively correlated with the magnitude of the first slip rate threshold. That is to say, the first slip rate threshold can be finely adjusted according to the slope or the load, such as increasing or decreasing by 0.1%. The load can be detected by the suspension height sensor of the vehicle.

[0102] In some embodiments, the slip rate is determined in the following manner: obtaining the first wheel speed and the second wheel speed of the vehicle, where the first wheel speed is greater than the second wheel speed; taking the difference between the first wheel speed and the second wheel speed as the wheel speed difference; and taking the ratio of the wheel speed difference to the second wheel speed as the slip rate.

[0103] In some embodiments, when the vehicle is a four-wheel drive vehicle, if the front axle wheel speed is greater than the rear axle wheel speed, the first wheel speed is the front axle wheel speed and the second wheel speed is the rear axle wheel speed. If the front axle wheel speed is less than the rear axle wheel speed, the first wheel speed is the rear axle wheel speed and the second wheel speed is the front axle wheel speed. When the vehicle is a two-wheel drive vehicle, the first wheel speed is the wheel speed of the driving wheel and the second wheel speed is the wheel speed of the non-driving wheel.

[0104] In some embodiments, the second wheel speed is replaced by the actual vehicle speed, which is the average wheel speed of the non-driving wheels or the vehicle navigation system speed. The vehicle navigation system speed is the vehicle driving speed measured and displayed by the vehicle navigation system. The vehicle navigation system can be a GPS system or a Beidou satellite navigation system. When the actual vehicle speed is the vehicle navigation system speed, it is not affected by vehicle skidding and has high accuracy. That is to say, in order to improve the accuracy of the actual vehicle speed, the vehicle navigation system speed is preferably used as the actual vehicle speed.

[0105] Exemplarily, the non-driving wheels of the vehicle are the rear wheels, and the average wheel speed of the non-driving wheels is the average of the wheel speeds of the left rear wheel and the right rear wheel.

[0106] In some embodiments, the rainfall data is obtained in the following manner: the rainfall sensor of the vehicle is used to detect the rainfall to obtain the rainfall data.

[0107] In some embodiments, determining the first road surface recognition result of the road surface where the vehicle is located based on the road surface image data and the vehicle image data includes: determining the first road surface type of the road surface where the vehicle is located according to the road surface image data and the vehicle image data. If the first road surface type is a wet road surface or an icy road surface, the first road surface recognition result is a low adhesion road surface risk. If the first road surface type is a dry road surface, the first road surface recognition result is not a low adhesion road surface risk.

[0108] In some embodiments, the weather forecast data is obtained in the following manner: when the vehicle is in an online state, the vehicle's in-vehicle T-Box is used to obtain the real-time weather API, such as the hourly forecast of the China Meteorological Administration; when the vehicle is in an offline state, historical weather data is obtained, and the weather forecast data is determined according to the historical weather data; the real-time position of the vehicle is obtained, and the real-time weather data of the real-time position is determined according to the real-time position, and the real-time weather data is used as the weather forecast data. The vehicle being in an online state can be understood as the vehicle being in a communication online state, and the vehicle being in an offline state can be understood as the vehicle being in an offline mode. The real-time position can be obtained through the vehicle's navigation system, such as through GPS. Determining the real-time weather data of the real-time position according to the real-time position can be understood as matching the local real-time weather data to improve the forecast accuracy.

[0109] It should be noted that when determining the weather forecast data according to the historical weather data, the accuracy decreases, and the confidence level should be marked to improve the accuracy of the weather forecast data.

[0110] In some embodiments, determining whether the road surface where the vehicle is located is a low-adhesion road surface according to each of the road surface recognition data and the corresponding preset weighting coefficient includes: obtaining a preset numerical threshold; for each type of the road surface recognition data, obtaining a preset recognition condition. If the road surface recognition data meets the preset recognition condition, taking a first numerical value as an intermediate numerical value; if not, taking a second numerical value as the intermediate numerical value, where the first numerical value is greater than the second numerical value. Weighting the intermediate numerical value based on the preset weighting coefficient to obtain a target numerical value; summing up the target numerical values to obtain a total target numerical value; when the total target numerical value is greater than the preset numerical threshold, the road surface where the vehicle is located is a low-adhesion road surface.

[0111] Exemplarily, when the road surface recognition data is the slip ratio and the acceleration, the road surface recognition data meets the preset recognition condition, and the preset weighting coefficient is 0.4; when the road surface recognition data is the rainfall data, the road surface recognition data meets the preset recognition condition, and the preset weighting coefficient is 0.3; when the road surface recognition data is the road surface image data and the vehicle image data, the road surface recognition data does not meet the preset recognition condition, and the preset weighting coefficient is 0.2; when the road surface recognition data is the weather forecast data, the road surface recognition data does not meet the preset recognition condition, and the preset weighting coefficient is 0.1. The first numerical value is 100, the second numerical value is 0, the preset numerical threshold is 60, and the total target numerical value = 0.4×100 + 0×100 + 0.2×0 + 0.1×0 = 70. Since the total target numerical value is greater than 60, the road surface where the vehicle is located is a low-adhesion road surface.

[0112] In some embodiments, when the total target numerical value is greater than the preset numerical threshold, the method further includes: when at least two types of the road surface recognition data meet the preset recognition condition, the road surface where the vehicle is located is a low-adhesion road surface; when at most one type of the road surface recognition data meets the preset recognition condition, the road surface where the vehicle is located is not a low-adhesion road surface. When the number of types of the road surface recognition data that meet the preset recognition condition is greater than or equal to 2, the road surface where the vehicle is located is a low-adhesion road surface. When the number of types of the road surface recognition data that meet the preset recognition condition is less than or equal to 1, that is, when a single sensor is triggered, even if the total target numerical value is greater than the preset numerical threshold, the road surface where the vehicle is located is not a low-adhesion road surface, thereby reducing the possibility of misjudgment.

[0113] Exemplarily, the rainfall data satisfies the preset identification condition, but the slip ratio does not satisfy the preset identification condition. For example, the rainfall sensor gives a false alarm but the wheel speed is normal, and the road surface where the vehicle is located is not a low-adhesion road surface. Among them, normal wheel speed indicates normal slip; the road surface image data and the vehicle image data do not satisfy the preset identification condition, but the slip ratio satisfies the preset identification condition. For example, the camera detects a dry road surface with a confidence level of 90%, and the slip ratio is 25%. It is possible to check whether the radar detects a water film to confirm whether there is splashing interference. If there is no water film, it is judged that the wheel speed sensor gives a false alarm, such as a tire leak.

[0114] In some embodiments, when at least two of the road surface identification data satisfy the preset identification condition, the method further includes: obtaining the target time when the road surface identification data satisfies the preset identification condition; when at least two of the target times are in the same time period, the road surface where the vehicle is located is a low-adhesion road surface; when at most one of the target times is in the same time period, the road surface where the vehicle is located is not a low-adhesion road surface. In the same time period, such as within 2 s, if the same time period includes at least two of the target times, it is confirmed that the road surface where the vehicle is located is a low-adhesion road surface. If the same time period includes at most one of the target times, even if the total target value is greater than the preset value threshold and at least two of the road surface identification data satisfy the preset identification condition, the road surface where the vehicle is located is not a low-adhesion road surface. This can be understood as a time window verification method, which requires multiple sensors to be triggered simultaneously within a short time window to reduce the possibility of instantaneous interference, thereby further reducing the possibility of misjudgment.

[0115] In some embodiments, obtaining the preset identification condition includes: when the road surface identification data is the slip ratio and the acceleration, determining the first duration for which the slip ratio is greater than the second slip ratio threshold, and taking the first duration being greater than the first duration threshold and the vehicle acceleration being less than the second acceleration threshold as the preset identification condition; when the road surface identification data is the rainfall data, determining the second duration for which the rainfall data is greater than the second rainfall threshold, and taking the second duration being greater than the second duration threshold as the preset identification condition; when the road surface identification data is the road surface image data in front of the vehicle and the vehicle image data, determining the second road surface identification result of the road surface where the vehicle is located for the road surface image data and the vehicle image data, and taking the second road surface identification result being a low-adhesion road surface and the confidence level being greater than the confidence level threshold as the preset identification condition; when the road surface identification data is the weather forecast data, taking the weather forecast data being an ice warning, a rain warning or a snow warning as the preset identification condition.

[0116] Exemplarily, the second slip rate threshold is 20%, the first duration threshold is 1 s, the second acceleration threshold is 0.3 g, where g is the acceleration due to gravity, the second rainfall threshold is 2 mm / h, the second duration threshold is 5 min, and the confidence threshold is 70%. When the road surface recognition data is slip rate and acceleration, the continuous duration with the slip rate greater than 20% being greater than 1 s and the acceleration being less than 0.3 g is used as the preset recognition condition; when the road surface recognition data is rainfall data, the continuous duration with the rainfall data greater than 2 mm / h being greater than 5 min is used as the preset recognition condition; when the road surface recognition data is the road surface image data and vehicle image data in front of the vehicle, the second road surface recognition result being a low adhesion road surface and the confidence being greater than 70% is used as the preset recognition condition.

[0117] In some embodiments, the first slip rate threshold is equal to the second slip rate threshold.

[0118] In some embodiments, the method further includes: when the weather is rainy, increasing the first slip rate threshold or the second slip rate threshold, that is, increasing the tolerance of wheel speed difference in rainy weather.

[0119] In some embodiments, the road surface image data includes first road surface data collected by a camera and second road surface data collected by a radar. Determining the second recognition result of the road surface where the vehicle is located from the road surface image data and the vehicle image data includes: identifying the characteristics of the road surface where the vehicle is located according to the first road surface data to obtain a first recognition result; determining the lateral shift standard deviation and the brake light flashing frequency of the target vehicle according to the vehicle image data; identifying the road surface where the vehicle is located according to the magnitude of the lateral shift standard deviation and / or the magnitude of the brake light flashing frequency to obtain a second recognition result; determining the signal reflection condition of the radar signal on the road surface where the vehicle is located according to the second road surface data; identifying the road surface where the vehicle is located according to the signal reflection condition to obtain a third recognition result; combining the first recognition result, the second recognition result, and the third recognition result to obtain the second road surface recognition result.

[0120] In some embodiments, identifying the characteristics of the road surface where the vehicle is located based on the first road surface data to obtain a first identification result includes: identifying the first road surface data according to a pre-established machine learning model to obtain the second road surface type of the road surface where the vehicle is located and the confidence level of the second road surface type. If the confidence level that the second road surface type is the target road surface is higher than the confidence level that the second road surface type is a dry road surface, and the target road surface is one of a wet road surface, an icy road surface, and a snow-covered road surface, the first identification result is a low-adhesion road surface and the confidence level is the maximum value of the confidence levels of each of the target road surfaces. If the second road surface type is a dry road surface, the first identification result is not a low-adhesion road surface.

[0121] In some embodiments, when training the machine learning model, a network public dataset of road surface images and in-vehicle collected data are obtained, and dry road surfaces, wet road surfaces, snow-covered road surfaces, or icy road surfaces are labeled according to the network public data and the in-vehicle collected data. When the road surface image is the dry road surface, it has the characteristics of clear texture and no specular reflection. When the road surface image is the wet road surface, it has the characteristics of reflective patches and tire splashing. When the road surface image is the icy road surface, it has the characteristics of low contrast and a white uniform area.

[0122] In some embodiments, the in-vehicle collected data is the collected data of the following scenarios: a heavy rain and water accumulation scenario where the vehicle speed is 80 km / h and the water film thickness is 3 mm; a black ice road surface where the vehicle is at a temperature of -5°C and a humidity of 90%; the vehicle performs a rapid acceleration on a gravel road, that is, the driving wheels slip but the road surface where the vehicle is located is not a low-adhesion road surface.

[0123] In some embodiments, the machine learning model is optimized through false alarm data or missed alarm data to achieve data closed-loop. The false alarm data or the missed alarm data can be obtained by uploading from the users of the vehicle.

[0124] In some embodiments, the machine learning model is continuously iterated with actual road condition data to optimize the image recognition model and the fusion algorithm of various road surface recognition data.

[0125] In some embodiments, identifying the road surface where the vehicle is located according to the magnitude of the lateral shift standard deviation and / or the magnitude of the brake light flashing frequency to obtain a second identification result includes: when the lateral shift standard deviation is greater than the standard deviation threshold and / or the brake light flashing frequency is greater than the frequency threshold, the second identification result is a low-adhesion road surface; when the lateral shift standard deviation is less than or equal to the standard deviation threshold and the brake light flashing frequency is less than or equal to the frequency threshold, the second identification result is not a low-adhesion road surface.

[0126] It should be noted that identifying the road surface where the vehicle is located based on the magnitude of the lateral displacement standard deviation and / or the magnitude of the brake light flashing frequency can be understood as detecting whether the target vehicle has abnormal braking, skidding, or ESP triggering, which serves as an indirect basis for judging the road surface where the vehicle is located. For example, if the target vehicle has abnormal braking, skidding, or ESP triggering, the road surface where the vehicle is located may be a low-adhesion road surface. If the vehicle in front of the target vehicle has abnormal braking, skidding, or no ESP triggering, the road surface where the vehicle is located may not be a low-adhesion road surface, and the target vehicle is the vehicle in front of the vehicle.

[0127] Exemplarily, when the lateral displacement standard deviation is less than 0.2 m / s, the target vehicle is driving normally. The standard deviation threshold is 0.5 m / s. When the lateral displacement standard deviation is greater than 0.5 m / s, it indicates that the target vehicle has a risk of losing control, indirectly indicating that the road surface where the vehicle is located is a low-adhesion road surface. The frequency threshold is 0.3 times / s, for example, 3 times in 10 s. When the brake light flashing frequency is greater than 0.3 times / s, it indicates that the target vehicle brakes frequently, indirectly indicating that the road surface where the vehicle is located is a low-adhesion road surface.

[0128] In some embodiments, the signal reflection condition includes reflection intensity and phase difference. Identifying the road surface where the vehicle is located according to the signal reflection condition to obtain a third identification result includes: when the reflection intensity is greater than the intensity threshold and / or the phase difference is greater than the phase threshold, the third identification result is a low-adhesion road surface; when the reflection intensity is less than or equal to the intensity threshold and the phase difference is less than or equal to the phase threshold, the third identification result is not a low-adhesion road surface.

[0129] It should be noted that when the radar is a millimeter-wave radar, the phase difference of its emitted wave reflected on the water surface is proportional to the water depth, which is used to detect the water film thickness or ice and snow coverage on the road surface. Flowing water will produce the Doppler effect of frequency shift. Considering the reflection signal difference between the static water film and the flowing water accumulation improves the reliability when judging whether the road surface where the vehicle is located is a low-adhesion road surface.

[0130] In some embodiments, obtaining the second road surface recognition result by combining the first recognition result, the second recognition result, and the third recognition result includes: for each of the first recognition result, the second recognition result, and the third recognition result, obtaining a preset recognition coefficient; if the recognition result is a low adhesion road surface, using a third value as the recognition value, and if the recognition result is not a low adhesion road surface, using a fourth value as the recognition value; obtaining a target recognition value by weighting the recognition value based on the preset recognition coefficient; summing up the target recognition values to obtain a total target recognition value; if the total target recognition value is greater than the recognition value threshold, the second road surface recognition result is a low adhesion road surface, and using the confidence level of the first recognition result as the confidence level of the second road surface recognition result, and if the total target recognition value is less than or equal to the recognition value threshold, the second road surface recognition result is not a low adhesion road surface.

[0131] In some embodiments, the preset recognition coefficient of the first recognition result is 0.5, the preset recognition coefficient of the second recognition result is 0.3, and the preset recognition coefficient of the third recognition result is 0.2.

[0132] In some embodiments, it further includes: obtaining various environmental parameters of the vehicle, where the various environmental parameters include environmental temperature, environmental time, navigation system positioning error, surrounding vehicles, and regional environment; adjusting the preset weighting coefficient according to the various environmental parameters so as to improve the environmental adaptability.

[0133] In some embodiments, adjusting the preset weighting coefficient according to the multiple environmental parameters includes: when the road surface recognition data is the slip ratio and acceleration, if the environmental time is night, increasing the preset weighting coefficient; when the road surface recognition data is rainfall data, if the surrounding vehicle is a sprinkler truck, reducing the preset weighting coefficient, and if the second duration is greater than the second duration threshold, increasing the preset weighting coefficient; when the road surface recognition data is the road surface image data and vehicle image data in front of the vehicle, if the environmental temperature is less than the temperature threshold, increasing the preset weighting coefficient, and if the environmental time is night, reducing the preset weighting coefficient; when the road surface recognition data is weather forecast data, if the regional environment is an area prone to icing, increasing the preset weighting coefficient, and if the positioning error of the navigation system is greater than the preset distance, reducing the preset weighting coefficient; when the road surface recognition data is the slip ratio and acceleration, if the environmental time is night, increasing the preset weighting coefficient and when the road surface recognition data is the road surface image data and vehicle image data in front of the vehicle, if the environmental time is night, reducing the preset weighting coefficient can be understood as the insufficient fill light of the camera at night, which requires reducing the weight of the road surface image data and vehicle image data and increasing the weight of the slip ratio and acceleration. For example, reducing the camera sensitivity by 20% and increasing the weight of the slip ratio and acceleration to 50%.

[0134] In some embodiments, the method further includes: when there is an icing prone point on the road surface in front of the vehicle, triggering a local warning in advance at a target distance. The icing prone point can be a bridge surface, and the target distance can be 200m.

[0135] In some embodiments, the method further includes: obtaining the point cloud data of the radar, determining the radar signal reflection intensity of the road surface in front of the vehicle based on the point cloud data, and if the radar signal reflection intensity of the road surface in front of the vehicle changes suddenly, the road surface where the vehicle is located is a low adhesion road surface. For example, if the radar signal intensity within 50m in front of the vehicle changes suddenly, it indicates that the road surface where the vehicle is located is a low adhesion road surface because the radar reflectivity of the ice surface is 30% lower than that of the asphalt road surface.

[0136] In this application, if the second duration is greater than the second duration threshold, increasing the preset weighting coefficient can be understood as the mixing of road surface oil stains and rainwater, and the adhesion coefficient of the road surface where the vehicle is located drops to the lowest, and the highest warning level needs to be triggered; if the surrounding vehicle is a sprinkler truck, reducing the preset weighting coefficient can be understood as detecting the rainfall data and also reducing the warning level of the low-adhesion road surface where the vehicle is located; if the regional environment is an area prone to icing, increasing the preset weighting coefficient can be understood as based on the map fence technology, when the vehicle enters an area prone to icing such as the exits of bridges and tunnels, automatically improving the monitoring sensitivity of weather forecasts; if the environmental temperature is less than the temperature threshold, increasing the preset weighting coefficient can be understood as the lower the temperature, the higher the icing risk of the road surface where the vehicle is located. Because when the environmental temperature is less than 3°C, for every 1°C decrease, the icing probability increases by 5%, so it is necessary to increase the weights of the road surface image data and the vehicle image data, that is, increase the icing risk weight, which can also be understood as a temperature compensation method; if the positioning error of the navigation system is greater than the preset distance, reducing the preset weighting coefficient can be understood as an error processing method. If the positioning error of the navigation system is large, the rainfall data, the road surface image data and the vehicle image data are preferentially used to identify the road surface where the vehicle is located.

[0137] In some embodiments, it further includes: identifying the surrounding vehicles through the vehicle image data.

[0138] In some embodiments, the adhesion coefficient of the road surface where the vehicle is located is determined by the following formula: μ(t) = μ 0 + 0.1×t, where μ(t) is the adhesion coefficient of the road surface where the vehicle is located t minutes after sprinkling, and μ 0 is the initial value of the adhesion coefficient of the road surface where the vehicle is located after sprinkling, and can be set to 0.4.

[0139] In this application, when the rainfall data is detected, it is inferred that the road surface where the vehicle is located may become slippery. In the initial stage of rainfall, due to the mixing of road surface oil stains and rainwater, the road surface where the vehicle is located may be even more slippery. Considering the second duration being greater than the second duration threshold as the preset recognition condition can be understood as judging the state of the road surface where the vehicle is located in combination with the rainfall duration. For example, continuous rainfall may cause water accumulation, and the road surface where the vehicle is located may become a low-adhesion road surface. Light rain may evaporate relatively quickly, and the road surface where the vehicle is located will not become a low-adhesion road surface. The rainfall data corresponds to the adhesion coefficient of the road surface where the vehicle is located. When the rainfall data is 0, that is, the weather is dry, the adhesion coefficient of the road surface where the vehicle is located is 0.8 - 1.0. When the rainfall data is 0 - 2 mm / h, that is, the weather is light rain, the adhesion coefficient of the road surface where the vehicle is located is 0.5 - 0.7. When the rainfall data is 2 - 10 mm / h, that is, the weather is moderate rain, the adhesion coefficient of the road surface where the vehicle is located is 0.3 - 0.5. When the rainfall data is greater than 10 mm / h, that is, the weather is heavy rain, the adhesion coefficient of the road surface where the vehicle is located is 0.2 - 0.3, which means the road surface where the vehicle is located has a risk of water accumulation.

[0140] In some embodiments, after the road surface where the vehicle is located becomes a low-adhesion road surface, the method further includes: adjusting the ABS, ESP, and TCS parameters of the vehicle to limit torque output or apply brakes in advance, or prompting "Slippery road surface, drive carefully" on the vehicle's instrument panel or HUD. Adjusting the ABS, ESP, and TCS parameters of the vehicle can be understood as the active safety system of the vehicle intervening after confirming that the road surface where the vehicle is located is a low-adhesion road surface. Prompting "Slippery road surface, drive carefully" on the vehicle's instrument panel or HUD can be understood as providing a warning prompt to the driver of the vehicle.

[0141] It should be noted that not being a low-adhesion road surface can be understood as a non-low-adhesion road surface.

[0142] In this application, first, determine the warning probability that each road surface recognition data causes the road surface where the vehicle is located to be a low-adhesion road surface, and combine the individual warning probabilities to determine the total warning probability that the road surface where the vehicle is located is a low-adhesion road surface. Second, when the total warning probability is greater than or equal to the preset probability threshold, determine whether the road surface where the vehicle is located is a low-adhesion road surface according to each road surface recognition data and the corresponding preset weighting coefficient. On the one hand, identify the road surface where the vehicle is located by combining multiple road surface recognition data. On the other hand, through two road surface identifications, the accuracy of road surface identification is improved. That is to say, this application conducts multi-dimensional data fusion and logical judgment based on multiple data such as slip rate, acceleration, weather forecast data, rainfall data, road surface image data, and vehicle image data, and can efficiently identify low-adhesion road surfaces, improving driving safety and the response accuracy of the control system.

[0143] According to a second aspect of the present application, a low adhesion road surface recognition device 100 is provided. The device includes:

[0144] A first acquisition unit 101, which acquires various road surface recognition data of the vehicle;

[0145] A first determination unit 102, for each of the road surface recognition data, acquires a warning probability that the road surface where the vehicle is located is a low adhesion road surface due to the road surface recognition data, and combines each of the warning probabilities to determine a total warning probability that the road surface where the vehicle is located is a low adhesion road surface;

[0146] A second determination unit 103, when the total warning probability is greater than or equal to a preset probability threshold, for each of the road surface recognition data, acquires a preset weighting coefficient of the road surface recognition data, and determines whether the road surface where the vehicle is located is a low adhesion road surface according to each of the road surface recognition data and the corresponding preset weighting coefficient. If so, the road surface where the vehicle is located is a low adhesion road surface.

[0147] In some embodiments, the first determination unit includes a second acquisition unit and a third determination unit. The second acquisition unit is configured to acquire a warning probability that the road surface where the vehicle is located is a low adhesion road surface due to the road surface recognition data. The third determination unit is configured to combine each of the warning probabilities to determine a total warning probability that the road surface where the vehicle is located is a low adhesion road surface. The second acquisition unit is configured as follows: a third acquisition unit, which acquires a preset probability condition; a fourth determination unit, if the road surface recognition data meets the preset probability condition, the warning probability of the road surface recognition data is a first probability value, if not, the warning probability of the road surface recognition data is a second probability value. Both the first probability value and the second probability value are greater than or equal to 0, the first probability value is greater than the second probability value, and the sum of the first probability value and the second probability value is 100%; the third determination unit is configured as follows: a fourth acquisition unit, which acquires a prior probability that the road surface where the vehicle is located is a low adhesion road surface; a fifth determination unit, which determines a posterior probability that the road surface where the vehicle is located is a low adhesion road surface according to each of the warning probabilities and the prior probability, and uses the posterior probability as the total warning probability.

[0148] In some embodiments, the third acquisition unit is configured as follows: First, as a unit, when the road surface recognition data is the slip ratio and acceleration, taking that the slip ratio is greater than a first slip ratio threshold and the acceleration is less than a first acceleration threshold as the preset probability condition; Second, as a unit, when the road surface recognition data is rainfall data, taking that the rainfall data is greater than a first rainfall threshold as the preset probability condition; Third, as a unit, when the road surface recognition data is the road surface image data and vehicle image data in front of the vehicle, determining a first road surface recognition result of the road surface where the vehicle is located from the road surface image data and the vehicle image data, and taking that the first road surface recognition result is a low adhesion road surface risk as the preset probability condition; Fourth, as a unit, when the road surface recognition data is weather forecast data, taking that the weather forecast data is an icing forecast, a rain forecast or a snow forecast as the preset probability condition.

[0149] In some embodiments, the second determination unit includes a sixth determination unit, and the sixth determination unit is used to determine whether the road surface where the vehicle is located is a low adhesion road surface according to each of the road surface recognition data and the corresponding preset weighting coefficient. The sixth determination unit is configured as follows: A fifth acquisition unit, which acquires a preset numerical threshold; Fifth, as a unit, for each type of road surface recognition data, acquires a preset recognition condition. If the road surface recognition data meets the preset recognition condition, taking a first numerical value as an intermediate numerical value, and if it does not meet, taking a second numerical value as the intermediate numerical value, where the first numerical value is greater than the second numerical value, and obtaining a target numerical value by weighting the intermediate numerical value based on the preset weighting coefficient; A first obtaining unit, which sums up each of the target numerical values to obtain a total target numerical value; Sixth, as a unit, when the total target numerical value is greater than the preset numerical threshold, the road surface where the vehicle is located is a low adhesion road surface.

[0150] In some embodiments, the fifth acting unit includes a sixth obtaining unit configured to obtain a preset recognition condition. The sixth obtaining unit is configured as follows: a seventh acting unit, when the road surface recognition data is slip ratio and acceleration, determines a first duration during which the slip ratio is greater than a second slip ratio threshold, and takes the first duration being greater than a first duration threshold and the vehicle acceleration being less than a second acceleration threshold as the preset recognition condition; an eighth acting unit, when the road surface recognition data is rainfall data, determines a second duration during which the rainfall data is greater than a second rainfall threshold, and takes the second duration being greater than a second duration threshold as the preset recognition condition; a ninth acting unit, when the road surface recognition data is road surface image data and vehicle image data in front of the vehicle, determines a second road surface recognition result of the road surface where the vehicle is located from the road surface image data and the vehicle image data, and takes the second road surface recognition result being a low adhesion road surface and the confidence level being greater than a confidence level threshold as the preset recognition condition; a tenth acting unit, when the road surface recognition data is weather forecast data, takes the weather forecast data being an icing forecast, a rain forecast or a snow forecast as the preset recognition condition.

[0151] In some embodiments, the road surface image data includes first road surface data collected by a camera and second road surface data collected by a radar. The ninth acting unit includes a seventh determination unit configured to determine a second recognition result of the road surface where the vehicle is located from the road surface image data and the vehicle image data. The seventh determination unit is configured as follows: a second obtaining unit, which identifies the characteristics of the road surface where the vehicle is located based on the first road surface data to obtain a first recognition result; an eighth determination unit, which determines the lateral movement standard deviation and the brake light flashing frequency of the target vehicle based on the vehicle image data; a third obtaining unit, which identifies the road surface where the vehicle is located based on the magnitude of the lateral movement standard deviation and / or the magnitude of the brake light flashing frequency to obtain a second recognition result; a ninth determination unit, which determines the signal reflection condition of the radar signal on the road surface where the vehicle is located based on the second road surface data; a fourth obtaining unit, which identifies the road surface where the vehicle is located based on the signal reflection condition to obtain a third recognition result; a fifth obtaining unit, which combines the first recognition result, the second recognition result and the third recognition result to obtain the second road surface recognition result.

[0152] In some embodiments, it further includes: a seventh obtaining unit configured to obtain various environmental parameters of the vehicle, where the various environmental parameters include environmental temperature, environmental time, navigation system positioning error, surrounding vehicles and regional environment; a first adjustment unit configured to adjust the preset weighting coefficient according to the various environmental parameters.

[0153] Based on the same inventive concept, as a third aspect, the present application also provides a computer-readable storage medium, on which a program product is stored that can implement the above-mentioned low-adhesion road surface recognition method in this specification.

[0154] Each aspect of the present application can also be implemented in the form of a program product, which includes program code. When the program product runs on a terminal device, the program code is used to cause the terminal device to execute the steps according to various exemplary embodiments of the present application described in the "Exemplary Method" section above in this specification.

[0155] Reference Figure 3 As shown, a program product 200 for implementing the above method according to an embodiment of the present application is described. It can be a portable compact disc read-only memory (CD-ROM) and includes program code, and can run on a terminal device, such as a personal computer. However, the program product of the present application is not limited to this. In this document, the readable storage medium can be any tangible medium that contains or stores a program, and this program can be used by or in conjunction with an instruction execution system, apparatus, or device.

[0156] As another aspect, the present application also provides an electronic device that can implement the above method.

[0157] The electronic device 300 is a general-purpose computing device, and its key components include a processing unit 310, a storage unit 320, and a bus 330 connecting them. The storage unit 320 stores program code, and this program can be executed by the processing unit 310 to implement the method steps in the present application. The storage unit 320 includes volatile storage (such as RAM 321 and cache 322) and read-only storage (ROM 323), and may include program modules such as an operating system and application programs.

[0158] The electronic device 300 can communicate with an external device 400 through an I / O interface 350, and can also communicate with a network (such as a LAN, WAN, Internet) through a network adapter 360. In addition, the electronic device 300 may also include other hardware and software modules, such as microcode, device drivers, etc., although these are not explicitly shown in the figure.

[0159] The above are only embodiments of the present application and are not used to limit the present application. For those skilled in the art, the present application can have various changes and modifications. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included within the scope of the claims of the present application.

Claims

1. A method for identifying a low adhesion road surface, characterized in that: The method comprises: Obtain various road surface recognition data of the vehicle; For each of the road surface identification data, obtaining the warning probability that the road surface where the vehicle is located is a low-adhesion road surface caused by the road surface identification data, and combining the various warning probabilities to determine the total warning probability that the road surface where the vehicle is located is a low-adhesion road surface; When the total warning probability is greater than or equal to a preset probability threshold, for each type of road surface identification data, a preset weighting coefficient of the road surface identification data is obtained, and whether the road surface where the vehicle is located is a low-adhesion road surface is determined based on each type of road surface identification data and the corresponding preset weighting coefficient. If so, the road surface where the vehicle is located is a low-adhesion road surface.

2. A low adhesion road surface identification method according to claim 1, characterized in that: The obtaining of the road surface recognition data results in a warning probability that the road surface on which the vehicle is located is a low-adhesion road surface, including: Obtaining preset probability conditions; If the road surface recognition data meets the preset probability condition, the warning probability of the road surface recognition data is a first probability value; if not, the warning probability of the road surface recognition data is a second probability value, the first probability value and the second probability value are both greater than or equal to 0, the first probability value is greater than the second probability value, and the sum of the first probability value and the second probability value is 100%; The combining of the warning probabilities to determine the total warning probability that the road surface where the vehicle is located is a low adhesion road surface includes: Obtaining a priori probability that the road surface on which the vehicle is located is a low-adhesion road surface; The posterior probability that the road surface where the vehicle is located is a low-adhesion road surface is determined according to each of the warning probabilities and the prior probability, and the posterior probability is used as the total warning probability.

3. A low adhesion road surface identification method according to claim 2, characterized in that: The obtaining of the preset probability condition includes: In a case where the road surface identification data is a slip rate and an acceleration, the slip rate being greater than a first slip rate threshold and the acceleration being less than a first acceleration threshold are used as the preset probability condition; In the case where the road surface identification data is rainfall data, taking the rainfall data being greater than a first rainfall threshold as the preset probability condition; In a case where the road surface recognition data is road surface image data and vehicle image data in front of the vehicle, determining a first road surface recognition result of the road surface image data and the vehicle image data for the road surface on which the vehicle is located, and taking the first road surface recognition result as a low adhesion road surface risk as the preset probability condition; In the case where the road surface recognition data is weather forecast data, the weather forecast data is an icing forecast, a rain forecast, or a snow forecast as the preset probability condition.

4. The low adhesion road surface identification method according to claim 1, characterized in that: The determining whether the road surface on which the vehicle is located is a low-adhesion road surface according to each of the road surface recognition data and the corresponding preset weighting coefficients includes: Get the preset numerical threshold; For each type of road surface recognition data, a preset recognition condition is obtained. If the road surface recognition data meets the preset recognition condition, a first value is used as an intermediate value. If not, a second value is used as an intermediate value, the first value is greater than the second value, and the intermediate values ​​are weighted based on the preset weighting coefficient to obtain a target value. Summing each of the target values ​​to obtain a total target value; When the total target value is greater than the preset value threshold, the road surface on which the vehicle is located is a low-adhesion road surface.

5. A low adhesion road surface identification method according to claim 4, characterized in that: The obtaining of the preset identification condition includes: In a case where the road surface recognition data is a slip rate and an acceleration, determining a first duration during which the slip rate is greater than a second slip rate threshold, and taking the first duration greater than a first duration threshold and the vehicle acceleration less than a second acceleration threshold as the preset recognition condition; In the case where the road surface identification data is rainfall data, determining a second duration during which the rainfall data is greater than a second rainfall threshold, and taking the second duration being greater than the second duration threshold as the preset identification condition; In a case where the road surface recognition data is road surface image data and vehicle image data in front of the vehicle, determining a second road surface recognition result of the road surface image data and the vehicle image data for the road surface on which the vehicle is located, and taking the second road surface recognition result as a low-adhesion road surface and a confidence level greater than a confidence level threshold as the preset recognition condition; In the case where the road surface recognition data is weather forecast data, the weather forecast data is ice forecast, rain forecast or snow forecast as the preset recognition condition.

6. A low adhesion road surface identification method according to claim 5, characterized in that: The road surface image data includes first road surface data collected by a camera and second road surface data collected by a radar, and determining a second recognition result of the road surface where the vehicle is located by the road surface image data and the vehicle image data includes: Identify the characteristics of the road surface where the vehicle is located according to the first road surface data to obtain a first identification result; Determine the lateral movement standard deviation and brake light flashing frequency of the target vehicle according to the vehicle image data; Identify the road surface where the vehicle is located according to the size of the lateral displacement standard deviation and / or the size of the brake light flashing frequency, and obtain a second identification result; determining, according to the second road surface data, a signal reflection condition of the radar signal on the road surface where the vehicle is located; Identify the road surface where the vehicle is located according to the signal reflection condition to obtain a third identification result; The second road surface recognition result is obtained by combining the first recognition result, the second recognition result and the third recognition result.

7. The low adhesion road surface identification method according to claim 4, characterized in that: Also includes: Acquiring various environmental parameters of the vehicle, including ambient temperature, ambient time, positioning error of the navigation system, surrounding vehicles and regional environment; The preset weighting coefficient is adjusted according to the multiple environmental parameters.

8. A low adhesion road surface recognition device, characterized in that: The device comprises: A first acquisition unit, which acquires various road surface recognition data of the vehicle; A first determination unit, for each of the road surface identification data, obtains a warning probability that the road surface where the vehicle is located is a low-adhesion road surface due to the road surface identification data, and determines a total warning probability that the road surface where the vehicle is located is a low-adhesion road surface by combining the warning probabilities; The second determination unit obtains a preset weighting coefficient of the road surface identification data for each type of the road surface identification data when the total warning probability is greater than or equal to a preset probability threshold, and determines whether the road surface where the vehicle is located is a low-adhesion road surface based on each type of the road surface identification data and the corresponding preset weighting coefficient. If so, the road surface where the vehicle is located is a low-adhesion road surface.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: The computer program includes executable instructions, and when the executable instructions are executed by a processor, the method described in any one of claims 1 to 7 is implemented.

10. An electronic device, characterized in that: include: one or more processors; A memory for storing executable instructions of the processor, wherein when the executable instructions are executed by the one or more processors, the one or more processors implement the method according to any one of claims 1 to 7.

Images