A vehicle status data acquisition method based on multi-source data fusion

Abstract

The present application relates to the field of data processing technology, and in particular to a vehicle status data collection method based on multi-source data fusion. The method includes: obtaining the capacity coefficient of the road section where the vehicle is currently located, and obtaining the information gain of the driving control information and the driving state type under the capacity coefficient; determining the weight value of the driving control information under the capacity coefficient according to multiple historical information gains corresponding to the same driving control information under the capacity coefficient, so as to determine N types of target driving control information from M types of driving control information in combination with the first driving state type of the vehicle; obtaining N types of target driving control information of the vehicle in the current time period, and using the N types of target driving control information and the first driving state type to determine the second driving state type of the vehicle. Through the above technical scheme, a more accurate driving state type of the vehicle can be determined to ensure the driving safety of the driver.

Description

Technical Field

[0001] The present application relates to the field of data processing technology, and in particular to a vehicle status data acquisition method based on multi-source data fusion. Background Art

[0002] During the driving process, the vehicle may be in various driving states, such as normal driving, emergency braking, and abnormal driving. The acceleration sensor and speed sensor installed on the vehicle can be used to determine the type of driving state the vehicle is in. By identifying the abnormal driving state of the vehicle (such as skidding, emergency braking, and sudden acceleration), the vehicle can trigger a safety warning mechanism and remind the driver to take measures or automatically intervene to control the vehicle to avoid potential accidents.

[0003] For example, a method for determining a vehicle's driving state is provided in a Chinese patent application document with publication number CN107871392A, including: obtaining vehicle driving information of the vehicle, the vehicle driving information including the vehicle's angular velocity and / or the vehicle's linear velocity; determining the vehicle's driving state according to the vehicle driving information, the vehicle driving state including a normal vehicle driving state and an abnormal vehicle driving state.

[0004] However, when a vehicle passes through an uneven road surface such as a bumpy road or a ramp, the sensor data monitored by the sensors installed on the vehicle may be affected, making the acquired sensor data unable to effectively reflect the actual driving state of the vehicle, thereby causing the vehicle's ECU (Electronic Control Unit) to misjudge the type of driving state.

[0005] For example, since the sensor data monitored by the sensors installed on the vehicle are affected, restrictions on the engine may be triggered when the vehicle is actually in a stable state, affecting the user's driving safety. Therefore, it is difficult for related technologies to obtain more accurate vehicle status data to ensure the user's driving safety. Summary of the invention

[0006] In order to overcome the problem that it is difficult to obtain relatively accurate vehicle status data in related technologies, the present application provides a vehicle status data collection method based on multi-source data fusion, including: obtaining a capacity coefficient of a road section where the vehicle is currently located, and obtaining information gain of driving control information and driving state type under the capacity coefficient; the capacity coefficient is used to characterize the degree of smooth passage on the road section, and the information gain is used to characterize the degree of mutual dependence between the driving state type and the driving control information; according to multiple historical information gains corresponding to the same driving control information under the capacity coefficient, determining a weight value of the driving control information under the capacity coefficient; obtaining a first driving state type of the vehicle, and according to the weight value of the driving control information under the capacity coefficient, multiplying the information gain of the driving control information corresponding to the first driving state type, determining N types of target driving control information with the largest product from M types of driving control information; obtaining N types of target driving control information of the vehicle in the current time period, and using the N types of target driving control information and the first driving state type, determining a second driving state type of the vehicle, the second driving state type is used to control the driving of the vehicle.

[0007] In this way, by obtaining the first driving state type of the vehicle and the capacity coefficient of the current road section, it is possible to determine a plurality of driving control information with the best stability from a plurality of driving control information based on the product of the weight value of the driving control information under the capacity coefficient and the information gain, thereby determining the second driving state type that can better reflect the actual driving state of the vehicle, thereby ensuring the driving safety of the driver; and because the accuracy of the determined driving state type is improved, it helps to reduce the number of false interventions of the vehicle's electronic control unit, which can improve the driver's driving experience.

[0008] Optionally, the information gain of the driving control information and the driving state type under the capacity coefficient is determined in the following manner: obtaining the information entropy of the driving state type of a specified vehicle when it is driving under the capacity coefficient; obtaining the conditional entropy of the driving state type of the specified vehicle when it is driving according to the driving control information under the capacity coefficient; and taking the difference between the information entropy and the conditional entropy of the corresponding driving control information as the information gain of the driving control information and the driving state type under the capacity coefficient.

[0009] In this way, by determining the information gain of the driving control information and the driving state type under the traffic capacity coefficient, the degree of mutual dependence between the driving state type and the driving control information of the vehicle under the traffic capacity coefficient can be better characterized.

[0010] Optionally, the weight value of the driving control information under the traffic capacity coefficient is determined by: ,in, is the weight value of the driving control information under the capacity coefficient, T is the number of historical information gains corresponding to the same driving control information under the capacity coefficient, is the i-th historical information gain corresponding to the driving control information under the traffic capacity coefficient, is the mean of the historical information gain corresponding to the driving control information under the capacity coefficient.

[0011] In this way, according to the differences between multiple historical information gains corresponding to the same driving control information under the capacity coefficient, the weight value of the vehicle's driving control information under the capacity coefficient can be adaptively determined, the weight value of the driving control information with stronger stability under the capacity coefficient can be increased, and the weight value of the driving control information with worse stability under the capacity coefficient can be reduced.

[0012] Optionally, the traffic capacity coefficient of the road section where the vehicle is currently located is obtained by: obtaining a road surface image of the current road section; the traffic capacity coefficient ,in, is the normalization function, is the area of ​​the flat road surface in the road image, is the area of ​​the potholes in the road image, and P is the contribution of the weather information of the current road section to the traffic capacity.

[0013] In this way, by taking into account the potholes on the road section where the vehicle is located and the weather information on the road section where the vehicle is located, a more accurate capacity coefficient can be determined.

[0014] Optionally, the weather information of the current road section is determined in the following manner: converting the road surface image of the road section where the vehicle is currently located into an HSV image, and obtaining three probability density curves of the pixel points in the HSV image in the hue, saturation and brightness channels respectively; inputting the three probability density curves corresponding to the road surface image into a pre-trained weather classification model to obtain the weather information output by the weather classification model; the weather classification model is used to output corresponding weather information according to the three probability density curves input.

[0015] In this way, by converting the road surface image of the road section where the vehicle is currently located into an HSV image, and obtaining three probability density curves of the pixel points in the HSV image in the hue, saturation and brightness channels respectively, more accurate weather information of the current road section can be obtained based on the three probability density curves in the hue, saturation and brightness channels respectively.

[0016] Optionally, the weather information of the current road section is determined in the following manner: obtaining sensor data obtained by a target sensor, and determining the weather information of the current road section based on the sensor data; the target sensor includes at least one of a temperature sensor, a humidity sensor, and an image sensor for capturing images of the vehicle's environment.

[0017] Optionally, the second driving state type of the vehicle is determined using N types of target driving control information and the first driving state type, including: inputting the N types of target driving control information and the first driving state type of the vehicle in the current time period into a pre-trained driving state classification model, and using the driving state type output by the driving state classification model as the second driving state type of the vehicle in the current time period; wherein the driving state classification model is used to output the corresponding second driving state type based on the input first driving state type and at least one driving control information.

[0018] Optionally, the driving state classification model is trained in the following manner: obtaining multiple driving control information of a specified vehicle within a preset time period, and a third driving state type of the specified vehicle within a preset time period, and a fourth driving state type obtained after pre-adjusting the third driving state type; using the multiple driving control information and the third driving state type as inputs of a pre-constructed classification model, and using the fourth driving state type as output of the classification model, and training the classification model to obtain the driving state classification model.

[0019] The technical solution provided by the embodiments of the present application may include the following beneficial effects: obtaining the first driving state type of the vehicle and the capacity coefficient of the current road section, and being able to determine a plurality of driving control information with the best stability from a plurality of driving control information based on the product of the weight value of the driving control information under the capacity coefficient and the information gain, thereby determining a second driving state type that can better reflect the actual driving state of the vehicle, thereby ensuring the driving safety of the driver; and since the accuracy of the determined driving state type is improved, it helps to reduce the number of false interventions of the vehicle's electronic control unit, which can improve the driver's driving experience.

[0020] It should be understood that the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the present application. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present application and, together with the description, serve to explain the principles of the present application.

[0022] Figure 1The present invention is a flow chart showing a method for collecting vehicle status data based on multi-source data fusion according to an exemplary embodiment. DETAILED DESCRIPTION

[0023] Here, exemplary embodiments are described in detail, and examples thereof are shown in the accompanying drawings. When the following description refers to the drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The implementations described in the following exemplary embodiments do not represent all implementations consistent with the present application.

[0024] First, the application scenario of the embodiment of the present application is briefly introduced. In the application scenario of the present application, the sensor data monitored by the sensors installed on the vehicle may be affected because the vehicle passes through uneven roads such as bumps or ramps, thereby obtaining sensor data that is difficult to reflect the actual driving state of the vehicle, causing the vehicle to misjudge the driving state type.

[0025] For example, due to an incorrect judgment of the driving state type, the vehicle may mistakenly believe that the vehicle is in a dangerous driving state when it is actually driving smoothly, thereby limiting the vehicle's driving speed and causing the vehicle to be rear-ended; or, the vehicle may mistakenly believe that the vehicle is in a safe driving state when it is actually in a dangerous driving state and fail to take braking measures, threatening the user's driving safety.

[0026] In view of the above technical problems, the present application provides a vehicle status data collection method based on multi-source data fusion. Figure 1 is a flow chart of a vehicle status data collection method based on multi-source data fusion according to an exemplary embodiment. Figure 1 As shown, the method includes the following steps.

[0027] In step S101, the capacity coefficient of the road section where the vehicle is currently located is obtained, and the information gain of the driving control information and the driving state type under the capacity coefficient is obtained.

[0028] The driving control information may include at least one of the following: lateral acceleration, longitudinal acceleration, steering speed, steering angle, accelerator pedal pressure, brake pedal pressure, throttle opening, transmission gear, tire pressure, driver concentration, and engine speed.

[0029] By setting corresponding sensors on the vehicle, it is possible to obtain corresponding driving control information of the vehicle. The sensor may be, for example, at least one of an acceleration sensor, a speed sensor, an inertial measurement unit, a pressure sensor, an infrared sensor, and an image sensor.

[0030] The driving state type may include at least one of the following: a normal acceleration driving state, a normal deceleration driving state, a constant speed driving state, an emergency braking state, an emergency acceleration driving state, a sharp turn state, and a drifting state.

[0031] The driving state type of the vehicle refers to the type corresponding to the driving state of the vehicle, and the driving state type of the vehicle can be determined based on at least one of the angular velocity, speed and acceleration of the vehicle in different directions.

[0032] The different directions of a vehicle may refer to the lateral, longitudinal and vertical directions of the vehicle; the lateral direction of the vehicle refers to the left and right directions of the vehicle; the longitudinal direction of the vehicle refers to the forward and backward directions of the vehicle; the vertical direction of the vehicle refers to the direction perpendicular to the plane where the vehicle's body is located.

[0033] The capacity coefficient of a road section is used to characterize the degree of smooth passage on the road section; the larger the capacity coefficient of a road section, the less impact the road conditions of the road section have on the passage of vehicles, and the more smoothly vehicles can pass on the road section, for example, the road section where the vehicle is currently located is a flat section.

[0034] On the contrary, the smaller the traffic capacity coefficient of the road section, the greater the impact of the road conditions of the road section on the traffic of vehicles, and the more difficult it is for vehicles to pass smoothly on the road section, for example, the road section where the vehicle is currently located is a bumpy section.

[0035] The information gain of driving control information and driving state type under the capacity coefficient can characterize the degree of mutual dependence between the driving state type and the driving control information when the vehicle travels on the road under the capacity coefficient.

[0036] The larger the value of information gain is, the higher the degree of mutual dependence between driving control information and driving state type is. When the vehicle is driving on a road with a traffic capacity coefficient, the higher the influence of driving control information on driving state type is. On the contrary, the smaller the value of information gain is, the lower the degree of mutual dependence between driving control information and driving state type is.

[0037] The information gain of driving control information and driving state type under the capacity coefficient can be determined by the driving control information obtained when the specified vehicle is driving on the road under the capacity coefficient and the driving state type of the vehicle; the specified vehicle can be the own vehicle for which vehicle state data is to be collected, or it can be other vehicles except the own vehicle.

[0038] Obtaining the information gain of driving control information and driving state type under the capacity coefficient is helpful to determine the degree of association between different driving control information and driving state type when the vehicle is traveling on a road with the capacity coefficient.

[0039] In step S102, a weight value of the driving control information under the traffic capacity coefficient is determined according to a plurality of historical information gains corresponding to the same driving control information under the traffic capacity coefficient.

[0040] The weight value of the driving control information under the capacity coefficient can characterize the consistency between multiple historical information gains corresponding to the same driving control information under the capacity coefficient, thereby characterizing the stability of the same driving control information under the capacity coefficient.

[0041] The better the consistency between multiple historical information gains corresponding to the same driving control information under the capacity coefficient, the higher the stability of the same driving control information under the capacity coefficient, and the stronger the anti-interference ability of the driving control information under the capacity coefficient. The more stable the driving control information is, the more accurate the driving state type of the vehicle can be determined.

[0042] In one embodiment, the weight value of the driving control information under the traffic capacity coefficient is determined by: ,in, is the weight value of the driving control information under the capacity coefficient, T is the number of historical information gains corresponding to the same driving control information under the capacity coefficient, is the i-th historical information gain corresponding to the driving control information under the traffic capacity coefficient, is the mean of the historical information gain corresponding to the driving control information under the capacity coefficient.

[0043] In the calculation formula for the weight value of driving control information under the capacity coefficient, the greater the difference between multiple historical information gains corresponding to the same driving control information under the capacity coefficient, the smaller the obtained weight value of the driving control information under the capacity coefficient, and a lower weight value can be assigned to driving control information with lower stability under the capacity coefficient.

[0044] In this way, according to the differences between multiple historical information gains corresponding to the same driving control information under the capacity coefficient, the weight value of the vehicle's driving control information under the capacity coefficient can be adaptively determined, the weight value of the driving control information with stronger stability under the capacity coefficient can be increased, and the weight value of the driving control information with worse stability under the capacity coefficient can be reduced.

[0045] In step S103, the first driving state type of the vehicle is obtained, and based on the weight value of the driving control information under the traffic capacity coefficient and the product of the information gain of the driving control information corresponding to the first driving state type, N types of target driving control information with the largest product are determined from M types of driving control information and based on the driving control information.

[0046] The first driving state type of the vehicle may be determined based on a variety of driving control information of the vehicle within a certain period of time before the current moment, or information such as acceleration, angular velocity, and speed of the vehicle in different directions.

[0047] Under the influence of the road conditions of the section where the vehicle is located, the monitoring data monitored by some sensors installed on the vehicle may be difficult to reflect the actual driving state of the vehicle, so that the first driving state type of the vehicle may deviate from the actual driving state type of the vehicle. Therefore, it is necessary to determine the actual driving state type of the vehicle based on obtaining the first driving state type.

[0048] The weight value of the same driving control information under the capacity coefficient can reflect the stability of the same driving control information when the vehicle is traveling on the road with the capacity coefficient; for example, it can reflect the stability of the same driving control information when different vehicles are traveling on the road with the capacity coefficient, or, it can reflect the stability of the same driving control information at different times when the same vehicle is traveling on the road with the capacity coefficient.

[0049] The information gain of the driving control information corresponding to the first driving state type under the capacity coefficient can reflect the degree of mutual dependence between the driving control information of the vehicle and the first driving state type when the vehicle is traveling on the road with the capacity coefficient.

[0050] Here, by multiplying the weight value of the driving control information under the capacity coefficient by the information gain of the first driving state type corresponding to the driving control information, the corresponding weight value and information gain under the capacity coefficient of the road section where the vehicle is located can be combined, so that the product can more comprehensively reflect the reliability of the driving control information of the vehicle on the current road section.

[0051] M is a positive integer greater than the positive integer N. The specific values ​​of M and N can be determined according to actual needs. For example, M can be 10 and N can be 5.

[0052] Since the product can more comprehensively reflect the reliability of the vehicle's driving control information on the current road section, by determining the N types of target driving control information with the largest product from the M types of driving control information, it is possible to avoid interference from other driving control information with lower reliability while retaining the N types of target driving control information with the highest reliability, thereby helping to obtain a more accurate driving state type of the vehicle.

[0053] For example, driving control information includes lateral acceleration, longitudinal acceleration, steering speed, steering angle, accelerator pedal pressure, brake pedal pressure, throttle opening, transmission gear, tire pressure, driver's concentration and engine speed.

[0054] When a vehicle passes through a bumpy road section, a variety of driving control information that is less affected by the bumps of the road section where the vehicle is located and is more closely related to the driving state type may be engine speed, brake pedal pressure, throttle opening, transmission gear, and tire pressure.

[0055] In step S104, N types of target driving control information of the vehicle in the current time period are obtained, and the second driving state type of the vehicle is determined using the N types of target driving control information and the first driving state type, where the second driving state type is used to control the driving of the vehicle.

[0056] After screening the M types of driving control information of the vehicle, when the vehicle is traveling on a road section with a traffic capacity coefficient, the N types of target driving control information are more stable, and the N types of target driving control information and the driving state type are more interdependent. Therefore, obtaining the N types of target driving control information of the vehicle in the current time period can determine a more accurate driving state type of the vehicle.

[0057] Although there may be certain differences between the first driving state type and the actual driving state type, there is still a certain consistency between the first driving state type and the actual driving state type of the vehicle. For example, when it is determined that the first driving state type of the vehicle is braking, the actual driving state type of the vehicle may be emergency braking, and the actual driving state type of the vehicle is not actually accelerating. Therefore, the first driving state type can provide a basic reference for determining a more accurate driving state type of the vehicle.

[0058] Since the N types of target driving control information can more accurately reflect the actual control state of the vehicle, the first driving state type can provide a basic reference for determining a more accurate driving state type of the vehicle. Therefore, combining the N types of target driving control information and the first driving state type can determine a second driving state type that is more in line with the actual driving state of the vehicle.

[0059] The second driving state type is used to control the driving of the vehicle; for example, when the second driving state type represents that the vehicle is in a sharp turning state, the damping coefficient of the shock absorber of the vehicle's suspension can be increased to provide stronger lateral support for the vehicle; or, when the second driving state type represents that the vehicle is in a normal turning state, the damping coefficient of the shock absorber of the vehicle's suspension can be appropriately reduced to improve the comfort of the driver of the vehicle; or, when the second driving state type represents that the vehicle is in a dangerous driving state, the vehicle's driving speed can be reduced to ensure the driving safety of the vehicle while ensuring the front and rear vehicle distance, thereby ensuring the safety of the driver and the vehicle.

[0060] In one embodiment, a second driving state type of a vehicle is determined using N types of target driving control information and a first driving state type, including: inputting the N types of target driving control information and the first driving state type of the vehicle in a current time period into a pre-trained driving state classification model, and using the driving state type output by the driving state classification model as the second driving state type of the vehicle in the current time period; wherein the driving state classification model is used to output a corresponding second driving state type based on the input first driving state type and at least one driving control information.

[0061] Compared with all the driving control information of the vehicle in the current time period, since the N types of target driving control information are driving control information obtained after screening multiple driving control information, the N types of target driving control information can better match the driving status or road conditions of the vehicle, so that a more accurate driving status type of the vehicle can be determined through the N types of target driving control information.

[0062] The pre-trained driving state classification model realizes the learning of the relationship between the driving control information of the vehicle at historical moments and the corresponding driving state type. It can output a driving state type that better reflects the actual situation of the vehicle based on the input driving control information and the initial driving state type.

[0063] There is a certain correlation between the first driving state type and the actual driving state type of the vehicle. Using the first driving state type as a driving state classification model can help the model output a second driving state type that matches the actual driving condition of the vehicle more quickly.

[0064] When the vehicle's sensor is not affected by factors such as road bumps or weather, the vehicle's first driving state type and the second driving state type output by the driving state classification model may be the same driving state type.

[0065] Alternatively, when the vehicle's sensors are affected by factors such as road bumps or weather, the vehicle's first driving state type and the second driving state type output by the driving state classification model may be different driving state types, or may be different driving state sub-types under the same driving state type; for example, the emergency braking driving state and the normal braking driving state may be two different driving state sub-types under the braking driving state type.

[0066] In this way, by inputting N types of target driving control information and the first driving state type into a pre-trained driving state classification model, the second driving state type of the vehicle can be adaptively determined so as to control the driving of the vehicle according to the second driving state type that matches the actual driving condition of the vehicle.

[0067] In one embodiment, the driving state classification model is trained in the following manner: obtaining a variety of driving control information of a specified vehicle within a preset time period, a third driving state type of a specified vehicle within a preset time period, and a fourth driving state type obtained after pre-adjusting the third driving state type; using a variety of driving control information and the third driving state type as input of a pre-constructed classification model, and using the fourth driving state type as output of the classification model, and training the classification model to obtain a driving state classification model.

[0068] By acquiring a variety of driving control information of the designated vehicle within a preset time period, the third driving state type of the designated vehicle can reflect the driving state of the vehicle to a certain extent. The implementer can determine the actual driving state of the vehicle based on the driving environment of the designated vehicle and the control state of the vehicle.

[0069] When the driving state represented by the third driving state type is inconsistent with the actual driving state of the vehicle, the third driving state type can be adjusted to obtain the fourth driving state type; or, when the driving state represented by the third driving state type is consistent with the actual driving state of the vehicle, the third driving state type can be used as the fourth driving state type obtained after adjusting the third driving state.

[0070] By taking a variety of driving control information and the third driving state type as the input of a pre-constructed classification model, and taking the fourth driving state type as the output of the classification model, the classification model can be trained to obtain a driving state classification model. When the training process meets preset conditions, such as the training process meets a predetermined duration or a predetermined number of times, or the loss function is less than a specified loss value, it can be determined that the training of the classification model is completed, and the trained classification model can be used as the driving state classification model.

[0071] In this way, by training the pre-built classification model to obtain the driving state classification model, it is possible to facilitate the use of the driving state classification model to determine the driving state type of the vehicle.

[0072] Through the vehicle status data collection method based on multi-source data fusion of the embodiment of the present application, the first driving state type of the vehicle and the capacity coefficient of the current road section are obtained, and according to the product of the weight value of the driving control information under the capacity coefficient and the information gain, a plurality of driving control information with the best stability can be determined from a plurality of driving control information, thereby determining a second driving state type that can better reflect the actual driving state of the vehicle, thereby ensuring the driving safety of the driver; and because the accuracy of the determined driving state type is improved, it helps to reduce the number of false interventions of the vehicle's electronic control unit, which can improve the driver's driving experience.

[0073] In one embodiment, the traffic capacity coefficient of the road section where the vehicle is currently located is obtained by: obtaining a road surface image of the current road section; the traffic capacity coefficient ,in, is the normalization function, is the area of ​​the flat road surface in the road image, is the area of ​​the potholes in the road image, and P is the contribution of the weather information of the current road section to the traffic capacity.

[0074] The contribution of a road section to the traffic capacity is different in different weather conditions, and different weather conditions correspond to different contribution values; for example, in cloudy and sunny weather conditions, the road section has a stronger traffic capacity, and the probability or degree of the vehicle's sensor being affected is lower.

[0075] In rainy, snowy and icy weather conditions, the traffic capacity of a road section is obstructed. When a vehicle slips due to snow or ice on the road, it will affect the vehicle's wheel speed sensor's monitoring of the vehicle's wheel speed, causing it to mistakenly think that the vehicle is driving at high speed, affecting the vehicle's electronic control unit's control of the vehicle.

[0076] In the road image of the current road section of the vehicle, the pothole road and the flat road usually have different characteristics, and the pothole road area and the flat road area in the road image of the vehicle can be segmented according to the different characteristics of the pothole road and the flat road.

[0077] For example, segmentation can be achieved through the color features and aggregation features of the pixels in the image, or the road surface image can be input into a pre-trained segmentation model to obtain the segmentation of the road surface image of the current road section to obtain the pothole road surface area and the flat road surface area; the embodiment of the present application does not limit the segmentation process of the road surface image of the current road section.

[0078] By comparing the area of ​​the flat road surface region in the road surface image with the area of ​​the pothole road surface region in the road surface image, the degree of potholes on the road surface where the vehicle is located can be determined, thereby determining the extent of the possible impact on the passage of the vehicle.

[0079] In this way, by taking into account the potholes on the road section where the vehicle is located and the weather information on the road section where the vehicle is located, a more accurate capacity coefficient can be determined.

[0080] In one embodiment, the weather information of the current road section is determined in the following manner: converting the road surface image of the road section where the vehicle is currently located into an HSV image, and obtaining three probability density curves of the pixel points in the HSV image in the hue, saturation and brightness channels respectively; inputting the three probability density curves corresponding to the road surface image into a pre-trained weather classification model to obtain the weather information output by the weather classification model; the weather classification model is used to output corresponding weather information according to the three probability density curves input.

[0081] HSV image is a color model that decomposes the color of an image into three components: hue, saturation, and value. HSV image is closer to the way humans perceive color, so it is easy to reflect the weather conditions of the road where the vehicle is located.

[0082] By analyzing the probability density curve of the HSV image, the model can more accurately identify different weather conditions, such as low brightness on rainy days, specific hues and high saturation on snowy days, etc.; and because the brightness information is separated in the HSV image, the impact of lighting changes on weather recognition results can be reduced, making the weather recognition results more stable.

[0083] Furthermore, the three probability density curves of the hue, saturation, and brightness channels of the HSV image reflect the statistical characteristics of the image instead of relying on specific pixel values, which helps the model remain robust in the face of different road conditions and environmental changes.

[0084] In this way, by converting the road surface image of the road section where the vehicle is currently located into an HSV image, and obtaining three probability density curves of the pixel points in the HSV image in the hue, saturation and brightness channels respectively, more accurate weather information of the current road section can be obtained based on the three probability density curves in the hue, saturation and brightness channels respectively.

[0085] In one embodiment, the weather classification model can be obtained by training in the following manner: converting the sample road image into an HSV image, and obtaining three probability density curves of the pixels in the HSV image in the hue, saturation, and brightness channels respectively; taking the three probability density curves corresponding to the sample road image as input, and taking the weather information label of the sample road image as output, and training the pre-constructed network model to obtain the weather classification model.

[0086] Multiple sample road images corresponding to different weather information labels can be obtained, the sample road images can be converted from RGB images to HSV images, and the three probability density curves of the pixel points in the HSV image in the hue, saturation and brightness channels can be determined respectively, which can better realize the training of the network model, so as to realize the classification of weather information through the weather classification model obtained by training.

[0087] In another embodiment, the weather information of the current road section is determined by: acquiring sensor data acquired by a target sensor, and determining the weather information of the current road section based on the sensor data; the target sensor includes at least one of a temperature sensor, a humidity sensor, and an image sensor for capturing images of the vehicle's environment.

[0088] In different weather conditions, the environmental parameters of the vehicle's environment may be different, such as temperature, humidity, wind speed, particle concentration, or environmental image characteristics. This allows the acquisition of sensor data about the vehicle's environment by sensors to effectively determine weather information.

[0089] The image sensor used to capture images of the vehicle's environment can capture images of the vehicle's surroundings. By comparing and analyzing multiple consecutive frames of images of the surroundings, the weather information of the vehicle's environment can be determined; for example, in rainy and snowy weather, there may be falling raindrops or snowflakes in the vehicle's environment.

[0090] In this way, the sensor data obtained by the vehicle's target sensor can be used to determine the weather information of the current road section, thereby achieving diversified acquisition of weather information on the road where the vehicle is located.

[0091] In one embodiment, the information gain of the driving control information and the driving state type under the capacity coefficient is determined in the following manner: obtaining the information entropy of the driving state type when the specified vehicle is driving under the capacity coefficient; obtaining the conditional entropy of the driving state type when the specified vehicle is driving according to the driving control information under the capacity coefficient; and using the difference between the information entropy and the conditional entropy of the corresponding driving control information as the information gain of the driving control information and the driving state type under the capacity coefficient.

[0092] Information entropy is a parameter value that measures the uncertainty of random variables. It specifies the information entropy of the driving state type of a vehicle when it is driving under the capacity coefficient. It can characterize the uncertainty of the driving state type of a vehicle when it is driving on a road with the capacity coefficient without considering the driving control information.

[0093] The information entropy of the driving state type when a specified vehicle is driving under the capacity coefficient is determined according to the type of driving state type and the frequency ratio of the occurrence when the vehicle is driving on the road of the capacity coefficient.

[0094] The information entropy of the driving state type of a specified vehicle when driving under the traffic capacity coefficient can be performed by referring to the calculation process of the information entropy, and the calculation process of the information entropy of the same variable belongs to the common knowledge of those skilled in the art and will not be repeated here.

[0095] Conditional entropy refers to the uncertainty of a random variable under a known condition. In the embodiment of the present application, the conditional entropy of the driving state type of a specified vehicle when driving according to driving control information under a capacity coefficient refers to the uncertainty of the driving state type of the vehicle when the vehicle is driving on a section of the capacity coefficient under the condition that the driving control information of the vehicle is known.

[0096] The information gain between two different variables, also known as the mutual information between two variables, can be used to measure the degree of mutual dependence between two variables; through the determined information gain, the contribution of driving control information to reducing the uncertainty of driving state type can be quantified.

[0097] When a vehicle is traveling on a road with a capacity coefficient, the greater the information gain between the driving control information and the driving state information, the more helpful the driving control information is in predicting and evaluating the driving state type of the vehicle; on the contrary, the smaller the information gain between the driving control information and the driving state information, the more difficult it is for the driving control information to predict and evaluate the driving state type of the vehicle.

[0098] The driving control information of a specified vehicle when traveling on a road with a capacity coefficient and the type of driving state of the vehicle can be statistically analyzed. Referring to the determination and steps of information gain in the embodiments of the present application, the values ​​of the information gain of each type of driving control information relative to each type of driving state can be obtained respectively.

[0099] For example, for driving control information x among the multiple types of driving control information and driving state type y among the multiple types of driving state, the information gain between the driving control information x and the driving state type y may be determined: , is the information gain between driving control information x and driving state type y, is the information entropy of driving state type y, is the conditional entropy of driving state type y when the driving control information x is known.

[0100] In this way, by determining the information gain of the driving control information and the driving state type under the traffic capacity coefficient, the degree of mutual dependence between the driving state type and the driving control information of the vehicle under the traffic capacity coefficient can be better characterized.

[0101] It should be understood that the features of some embodiments of the various applications described herein may be combined with each other unless specifically stated otherwise.

[0102] Although terms such as "first", "second", and "third" may be used herein to describe various components, parts, regions, layers, or sections, these components, parts, regions, layers, or sections are not limited to these terms. Instead, these terms are only used to distinguish one component, part, region, layer, or section from another component, part, region, layer, or section. Therefore, without departing from the teachings of the examples described herein, the first component, part, region, layer, or section mentioned in the examples may also be referred to as the second component, part, region, layer, or section.

[0103] In addition, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined as "first" and "second" may explicitly or implicitly include at least one of the features. In the description herein, the meaning of "plurality" is at least two, such as two, three, etc., unless otherwise clearly and specifically defined.

[0104] In addition, the word "exemplary" is used herein to mean serving as an example, instance, or diagram. Any aspect or design described herein as "exemplary" is not necessarily to be understood as being advantageous over other aspects or designs. Instead, the use of the word exemplary is intended to present concepts in a concrete manner. As used herein, the term "or" is intended to mean an inclusive "or" rather than an exclusive "or".

[0105] Likewise, although the present application has been shown and described with respect to one or more implementations, equivalent variations and modifications will occur to those skilled in the art upon reading and understanding this specification and the accompanying drawings. With particular regard to the various functions performed by the components (e.g., elements, resources, etc.) described above, unless otherwise noted, the terms used to describe such components are intended to correspond to any component (functionally equivalent) that performs the specific functions of the described components, even if not structurally equivalent to the disclosed structures.

[0106] Additionally, while particular features of the present application may have been disclosed with respect to only one of several implementations, such features may be combined with one or more other features of other implementations as may be desired and advantageous for any given or particular application.

[0107] Those skilled in the art will readily appreciate other embodiments of the present application after considering the specification and practicing the invention disclosed herein. The present application is intended to cover any modification, use or adaptation of the present application, which follows the general principles of the present application and includes common knowledge or customary technical means in the art that are not disclosed in the present application, and the specification and embodiments are only considered as exemplary.

[0108] It should be understood that the present application is not limited to the exact construction that has been described above and shown in the drawings, and that various modifications and changes may be made without departing from the scope thereof.

Claims

1. A vehicle status data acquisition method based on multi-source data fusion, characterized in that: include: Obtaining the capacity coefficient of the road section where the vehicle is currently located, and obtaining the information gain of the driving control information and the driving state type under the capacity coefficient; the capacity coefficient is used to characterize the degree of smooth passage on the road section, and the information gain is used to characterize the degree of mutual dependence between the driving state type and the driving control information; Determining a weight value of the driving control information under the capacity coefficient according to a plurality of historical information gains corresponding to the same driving control information under the capacity coefficient; Acquire a first driving state type of the vehicle, and determine N types of target driving control information with the largest product from M types of driving control information according to a weight value of the driving control information under a traffic capacity coefficient and a product of an information gain of the driving control information corresponding to the first driving state type; N types of target driving control information of the vehicle in a current time period are obtained, and a second driving state type of the vehicle is determined using the N types of target driving control information and the first driving state type, where the second driving state type is used to control the driving of the vehicle.

2. The vehicle status data acquisition method based on multi-source data fusion according to claim 1 is characterized in that: The information gain of driving control information and driving state type under the capacity coefficient is determined by the following method: Obtain the information entropy of the driving state type of the specified vehicle when driving under the traffic capacity coefficient; Obtain the conditional entropy of the driving state type of the specified vehicle when it is driving according to the driving control information under the traffic capacity coefficient; The difference between the information entropy and the conditional entropy of the corresponding driving control information is used as the information gain of the driving control information and the driving state type under the traffic capacity coefficient.

3. The vehicle status data acquisition method based on multi-source data fusion according to claim 1 is characterized in that: The weight value of driving control information under the traffic capacity coefficient is determined by the following method: ,in, is the weight value of the driving control information under the capacity coefficient, T is the number of historical information gains corresponding to the same driving control information under the capacity coefficient, is the i-th historical information gain corresponding to the driving control information under the traffic capacity coefficient, is the mean of the historical information gain corresponding to the driving control information under the capacity coefficient.

4. The vehicle status data acquisition method based on multi-source data fusion according to claim 1 is characterized in that: The capacity coefficient of the road section where the vehicle is currently located is obtained in the following way: Get the road surface image of the current road section; the traffic capacity coefficient ,in, is the normalization function, is the area of ​​the flat road surface in the road image, is the area of ​​the potholes in the road image, and P is the contribution of the weather information of the current road section to the traffic capacity.

5. The vehicle status data acquisition method based on multi-source data fusion according to claim 4 is characterized in that: The weather information of the current road section is determined in the following manner: Convert the road surface image of the road section where the vehicle is currently located into an HSV image, and obtain three probability density curves of the pixel points in the HSV image in the hue, saturation and brightness channels respectively; The three probability density curves corresponding to the road surface image are input into a pre-trained weather classification model to obtain weather information output by the weather classification model; the weather classification model is used to output corresponding weather information according to the three probability density curves input.

6. The vehicle status data acquisition method based on multi-source data fusion according to claim 4 is characterized in that: The weather information of the current road section is determined in the following manner: Acquire sensor data acquired by a target sensor, and determine weather information of a current road section according to the sensor data; the target sensor includes at least one of a temperature sensor, a humidity sensor, and an image sensor for capturing images of an environment in which a vehicle is located.

7. The vehicle status data acquisition method based on multi-source data fusion according to claim 1 is characterized in that: Determining a second driving state type of the vehicle using the N types of target driving control information and the first driving state type includes: Inputting N kinds of target driving control information of the vehicle in the current time period and the first driving state type into a pre-trained driving state classification model, and using the driving state type output by the driving state classification model as the second driving state type of the vehicle in the current time period; The driving state classification model is used to output a corresponding second driving state type according to an input first driving state type and at least one driving control information.

8. The vehicle status data acquisition method based on multi-source data fusion according to claim 7 is characterized in that: The driving state classification model is trained by the following method: Acquire a plurality of driving control information of a specified vehicle within a preset time period, a third driving state type of the specified vehicle within the preset time period, and a fourth driving state type obtained by pre-adjusting the third driving state type; The plurality of driving control information and the third driving state type are used as inputs of a pre-constructed classification model, and the fourth driving state type is used as an output of the classification model, and the classification model is trained to obtain a driving state classification model.

Images