Vehicle behavior detection method, device and system based on vehicle AI platform

Abstract

The invention relates to the technical field of road vehicle control systems, and provides a vehicle behavior detection method, device and system based on a vehicle AI platform, and the method comprises the steps: obtaining the multi-type vehicle driving information of a vehicle, the weather information of the location of the vehicle and the road map information, and recording a plurality of traffic accident data; obtaining the vehicle density of each vehicle based on the similar performance of the vehicle driving information of the plurality of vehicles; quantifying road environment parameters of each vehicle; quantifying weather environment parameters of the vehicle; obtaining a driving environment coefficient of the vehicle; constructing a plurality of driving environment information groups of the vehicle to obtain a plurality of related accident information groups of the vehicle; obtaining the accident correlation between various vehicle driving information and various environment information of each vehicle; building an accident space diagram of each vehicle; and obtaining a driving risk coefficient of the vehicle and carrying out dangerous behavior early warning. The objective of the invention is to solve the problem that vehicle behavior detection lacks dynamic risk quantification capability based on information such as road environment and traffic flow.

Description

Technical Field

[0001] The present invention relates to the technical field of road vehicle control systems, and in particular to a vehicle behavior detection method, device, and system based on a vehicle AI platform. Background Art

[0002] Since a large proportion of current road traffic accidents are related to dangerous driving behaviors, that is, vehicles fail to comply with traffic regulations and driving requirements when driving on the road, resulting in traffic accidents, in order to prevent traffic accidents and reduce the risk of accidents, it is necessary to detect abnormal behaviors of vehicles during driving. However, the current vehicle behavior detection method only relies on a single criterion, so that the same driving behavior may be judged as abnormal in different scenarios. That is, since the same driving behavior may correspond to completely different risk levels in highway and urban road scenarios, and the existing technology has not effectively solved the "scenario-behavior-risk" mapping problem, the current vehicle behavior detection mostly uses static risk assessment models, and lacks dynamic risk quantification capabilities based on contextual information such as road environment and traffic flow. Summary of the Invention

[0003] The present invention provides a vehicle behavior detection method, device, and system based on a vehicle AI platform to address the problem that existing vehicle behavior detection lacks dynamic risk quantification based on information such as road environment and traffic flow. The technical solutions adopted are as follows: The present invention proposes a vehicle behavior detection method based on a vehicle AI platform, which includes the following steps: Obtain multiple types of vehicle driving information and weather information and road map information of the location of the vehicle, transmit them to the vehicle AI platform, and record certain traffic accident data; Based on the similarity between the driving information of several vehicles, the vehicle density of each vehicle is obtained, and the weight of the road type at the vehicle's location is recorded; the road environment parameters of each vehicle are quantified by combining the driving information of multiple types of vehicles and the highway map information at the location; the weather information at the vehicle's location is analyzed to quantify the vehicle's weather environment parameters; and the vehicle's driving environment coefficient is obtained by combining the vehicle's road environment parameters; Based on the vehicle's driving information, as well as weather information and road map information at the vehicle's location as environmental information, several vehicle driving environment information groups are constructed. Combined with similar relationships between the same types of information in traffic accident data, several vehicle-related accident information groups are obtained. The correlation between the environmental information and vehicle driving information in the relevant accident information groups and their proportion in the traffic accident data are analyzed to obtain the accident correlation between each type of vehicle driving information and each type of environmental information for each vehicle. Based on the vehicle driving information and environmental information, as well as the accident correlation, an accident space map of each vehicle is constructed; combined with the vehicle's driving environment coefficient, the vehicle's driving risk coefficient is obtained and a dangerous behavior warning is issued.

[0004] Optionally, the vehicle density of each vehicle is obtained by: Take any vehicle as the target vehicle, and construct the target vehicle's driving vector using the real-time positioning information and driving direction in the target vehicle's driving information; obtain the driving vectors of all vehicles except the target vehicle, and obtain the cosine similarity between the driving vectors of all other vehicles and the driving vector of the target vehicle as the driving similarity factor between the other vehicles and the target vehicle; Perform DBSCAN clustering on all vehicles except the target vehicle. The distance metric is the absolute value of the difference between the driving similarity factors of each vehicle and the target vehicle, resulting in several clusters. The mean driving similarity factor between the vehicles in any cluster and the target vehicle is obtained, and the cluster with the largest mean driving similarity factor is selected as the peer cluster of the target vehicle. The product of the number of lanes on the road where the target vehicle is located and the minimum value of the driving similarity factor between the vehicles in the same cluster and the target vehicle is obtained, and the ratio of the number of vehicles in the same cluster to the product is used as the vehicle density of the target vehicle.

[0005] Optionally, the road environment parameters of each vehicle are obtained in the following specific method: Based on the absolute value of the road slope, road curvature, road type weight, and road speed limit value at the target vehicle's location, the target vehicle's road driving factor is obtained; the average driving speed of all vehicles in the target vehicle's peer cluster is used as the target vehicle's peer speed; Based on the vehicle density, peer speed and road driving factor of the target vehicle, the road environment parameters of the target vehicle are obtained; The road environment parameter is positively correlated with the road driving factor and the vehicle density, and negatively correlated with the traveling speed; the road driving factor is positively correlated with the absolute value of the road slope and the road curvature, and negatively correlated with the road type weight and the road speed limit.

[0006] Optionally, the quantified vehicle weather environment parameters are combined with the vehicle road environment parameters to obtain the vehicle driving environment coefficient, including the following specific methods: Set corresponding precipitation type weights for each precipitation type; set corresponding road condition weights based on road conditions; The sum of the precipitation type weight and the road condition weight at the target vehicle's location and the ratio of the visibility at the target vehicle's location are normalized and used as the weather environment parameter of the target vehicle; The product of the target vehicle's weather environment parameters and road environment parameters and the normalized result is used as the target vehicle's driving environment coefficient.

[0007] Optionally, the specific method of obtaining several relevant accident information groups of the vehicle includes: Combine multiple types of weather information and road map information at the vehicle's location as multiple types of vehicle environmental information; combine any type of vehicle driving information and any type of environmental information as a driving environment information group, and obtain multiple driving environment information groups for the target vehicle; For any driving environment information group of the target vehicle, obtaining the element corresponding to the driving environment information group under each traffic accident data from a plurality of traffic accident data as the accident information group of each traffic accident data in the driving environment information group; Obtain the cosine similarity between each accident information group and the target vehicle's driving environment information group. If the cosine similarity is greater than or equal to a relevant threshold, use the corresponding accident information group as the relevant accident information group of the target vehicle's driving environment information group; obtain the relevant accident information group of each driving environment information group of the target vehicle.

[0008] Optionally, the specific method of obtaining the accident correlation between the driving information of each vehicle and each type of environmental information includes: For any environmental information and any vehicle driving information, based on all relevant accident information groups of the target vehicle, the ratio of the number of relevant accident information groups containing the environmental information and the vehicle driving information to the number of relevant accident information groups containing the environmental information is used as the conditional probability of the vehicle driving information occurring under the environmental information; The ratio of the number of relevant accident information groups including the environmental information and the vehicle driving information to the total number of traffic accident data is used as the relative frequency of the environmental information and the vehicle driving information; The product of the conditional probability of the vehicle driving information occurring under the environmental information and the relative frequency of the environmental information and the vehicle driving information is used as the accident correlation factor of the vehicle driving information and the environmental information of the target vehicle; Obtain accident correlation factors of various types of vehicle driving information and various types of environmental information of the target vehicle, perform linear normalization on all accident correlation factors, and obtain the results as accident correlations of various types of vehicle driving information and various types of environmental information of the target vehicle.

[0009] Optionally, the construction of the accident space graph of each vehicle includes the following specific methods: The target vehicle is taken as the central node, the driving information of various types of vehicles is taken as driving nodes, and the various types of environmental information is taken as environmental nodes. The central node is connected to each driving node. For any driving node and any environmental node, the accident correlation between the vehicle driving information corresponding to the driving node of the target vehicle and the environmental information corresponding to the environmental node is obtained. If the accident correlation is greater than or equal to the association threshold, the driving node and the environmental node are connected, and the accident correlation is used as the edge value of the connecting edge. All driving nodes and environmental nodes are connected and judged based on the corresponding accident correlation, and the connection between each node is completed to form the accident space diagram of the target vehicle.

[0010] Optionally, the driving risk coefficient of the vehicle is obtained in the following specific method: The mean of the accident correlation between the vehicle driving information corresponding to each driving node and the environmental information corresponding to each environmental node in the accident space diagram of the target vehicle is obtained, and the product of the mean and the driving environment coefficient of the target vehicle is used as the driving risk coefficient of the target vehicle.

[0011] The present invention also proposes a vehicle behavior detection system based on a vehicle AI platform, which includes: The vehicle information collection module is used to obtain various types of vehicle driving information and weather information and road map information of the location, transmit it to the vehicle AI platform, and record certain traffic accident data; The vehicle information analysis module is used to obtain the vehicle density of each vehicle based on the similar performance of the vehicle driving information of multiple vehicles, and record the road type weight of the vehicle's location; combine the driving information of multiple types of vehicles with the local highway map information to quantify the road environment parameters of each vehicle; analyze the weather information of the vehicle's location to quantify the vehicle's weather environment parameters; and combine the vehicle's road environment parameters to obtain the vehicle's driving environment coefficient; Based on the vehicle's driving information, as well as weather information and road map information at the vehicle's location as environmental information, several vehicle driving environment information groups are constructed. Combined with similar relationships between the same types of information in traffic accident data, several vehicle-related accident information groups are obtained. The correlation between the environmental information and vehicle driving information in the relevant accident information groups and their proportion in the traffic accident data are analyzed to obtain the accident correlation between each type of vehicle driving information and each type of environmental information for each vehicle. The vehicle behavior detection module is used to construct an accident space map for each vehicle based on vehicle driving information and environmental information, as well as the accident correlation; combined with the vehicle's driving environment coefficient, it obtains the vehicle's driving risk coefficient and issues dangerous behavior warnings.

[0012] The present invention also proposes a vehicle behavior detection device based on a vehicle AI platform, which includes a memory, a processor, and a computer program stored in the memory and running on the processor. When the processor executes the computer program, the steps of the above method are implemented.

[0013] The beneficial effects of the present invention are as follows: the present invention monitors the vehicle's driving information in real time through a vehicle AI platform, and dynamically assesses the risk of driving behavior in combination with factors such as weather information, road type and traffic flow; the real-time vehicle density of the vehicle is determined by combining the similar performance of the driving information of different vehicles with the number of lanes, and the driving environment coefficient is quantified based on this combined with the impact of weather information on road conditions; the correlation between driving information and weather information and highway map information as environmental information under historical traffic accident data is further considered, and the accident-related performance between driving behavior and environmental information is quantified to construct an accident space map and analyze it to obtain a driving risk coefficient; compared with traditional methods, it can provide more accurate driving risk prediction. If high-risk driving behavior is detected, the system will automatically issue a warning to help the driver reduce the possibility of an accident; through quantitative analysis of weather and road environment, a judgment is made on whether the driver's behavior is reasonable; for example, speeding when driving in rainy days may lead to a higher accident risk. The system will adjust the risk assessment criteria according to weather conditions. By introducing dynamic mapping of scene-behavior-risk, more accurate driving behavior detection is achieved. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0015] Figure 1 A schematic flow chart of a vehicle behavior detection method based on a vehicle AI platform provided by one embodiment of the present invention; Figure 2 A structural block diagram of a vehicle behavior detection system based on a vehicle AI platform provided by another embodiment of the present invention; Figure 3 This is a structural diagram of the accident space diagram of the target vehicle of the present invention. DETAILED DESCRIPTION

[0016] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0017] See also Figure 1 , which shows a flow chart of a vehicle behavior detection method based on a vehicle AI platform provided by one embodiment of the present invention, the method comprising the following steps: Step S001: Obtain multiple types of vehicle driving information and weather information and road map information of the vehicle's location, transmit them to the vehicle AI platform, and record some traffic accident data.

[0018] It should be noted that in order for the vehicle AI platform to effectively understand the vehicle's status and conduct driving behavior detection and analysis, it is necessary to obtain various information about the vehicle and transmit it to the vehicle AI platform to provide a data basis for the vehicle AI platform; the vehicle status is directly related to the geographical environment or road environment in which the vehicle is located during driving and the driver's driving behavior. In addition, different weather environments also affect the driving of the vehicle during driving. For example, in freezing rain, ice and snow, vehicles driving at unreasonable speeds or other driving behaviors on the highway may cause serious traffic accidents. Therefore, it is also necessary to obtain weather information where the vehicle is located.

[0019] Specifically, the vehicle driving information of several vehicles, as well as the weather information and road map information of the vehicle's location, are obtained respectively through the corresponding API and the on-board module unit. Since the vehicle AI platform is required for transmission, a large number of vehicles are required for real-time collection; the vehicle driving information includes the vehicle's real-time positioning information (longitude, latitude, altitude), driving speed (unit: km / h), driving direction (unit: degree) and acceleration data (unit: m / s²). The vehicle driving information is obtained through the vehicle's built-in global positioning system (GPS) module and the on-board inertial measurement unit (IMU), and various types of vehicle driving information are obtained.

[0020] Furthermore, real-time weather information at the vehicle's location is obtained through a third-party weather service application programming interface (i.e., API, such as the National Meteorological Administration OpenAPI or commercial services such as the AccuWeather API) integrated into the vehicle AI platform. The weather information includes temperature (real-time air temperature, unit: ℃), humidity (relative humidity, unit: %), precipitation type (enumerated values: no precipitation / light rain / moderate rain / heavy rain / freezing rain / snow / hail), visibility (unit: meter) and road conditions (enumerated values: dry / wet / water accumulation / snow accumulation / ice accumulation), thereby obtaining various types of weather information.

[0021] Furthermore, highway map information is obtained through the map data interface of the vehicle AI platform (the map data interface is connected to a high-precision digital map service (such as Here Maps or AutoNavi Map API)). The highway map information includes road type (enumerated values: expressway, urban expressway, ordinary urban road, rural road), speed limit value of the current road section (unit: km / h), number of lanes, road curvature (unit: 1 / meter, representing the radius of the curve, the larger the value, the sharper the curve) and road slope (unit: %), thereby obtaining highway map information of various types.

[0022] Furthermore, the collected multi-type vehicle driving information and the weather information and highway map information of the vehicles' locations are encapsulated into a unified data packet through the 5G communication module and transmitted to the vehicle AI platform via the 5G network; at the same time, a large amount of traffic accident data is collected and stored in the vehicle AI platform, where the traffic accident data includes the vehicle driving information, weather information and highway map information of each historical accident at the time of occurrence.

[0023] Step S002: Based on the similar performance of vehicle driving information of several vehicles, the vehicle density of each vehicle is obtained, and the road type weight of the vehicle's location is recorded; the road environment parameters of each vehicle are quantified by combining the driving information of multiple types of vehicles and the highway map information of the location; the weather information of the vehicle's location is analyzed and the weather environment parameters of the vehicle are quantified; and the driving environment coefficient of the vehicle is obtained by combining the vehicle's road environment parameters.

[0024] It should be noted that whether a vehicle's behavior constitutes dangerous driving needs to be judged based on the driving environment in which the vehicle is located. For example, regarding driving speed, when a vehicle is driving at high speed on roads of different levels, if the vehicle's driving speed is within the speed limit of the highway, then the vehicle is not considered dangerous driving, while driving at high speed on non-highways is dangerous driving. Therefore, in order to accurately detect vehicle behavior, the vehicle AI platform needs to comprehensively analyze the local weather information, vehicle driving information, and road map information to determine whether the vehicle's driving behavior constitutes dangerous driving under the vehicle's driving environment.

[0025] It should be further explained that the vehicle's driving environment refers to the vehicle's external environment, which is related to the road environment and weather environment in which it is located. Generally speaking, whether a vehicle's behavior is dangerous depends on whether the environment allows the existence of its corresponding behavior. Therefore, it is necessary to analyze the vehicle's driving environment.

[0026] Preferably, in one embodiment of the present invention, based on the similar performance of vehicle driving information of a plurality of vehicles, the vehicle density of each vehicle is obtained, and the road type weight of the vehicle location is recorded, including the specific method of: It should be noted that the road environment on which the vehicle is traveling usually involves the current road conditions and congestion situation. The road conditions involve the road type, speed limit, number of lanes, curvature and slope of the road on which the vehicle is currently located; and the congestion situation involves the density of vehicles on the road on which the vehicle is currently located and the average travel speed. Therefore, in order to understand the environment in which the vehicle is currently located on the road, it is necessary to quantify the road environment parameters.

[0027] Specifically, any vehicle is taken as the target vehicle, and the real-time positioning information and driving direction in the vehicle driving information of the target vehicle are constructed as the driving vector of the target vehicle; the driving vectors of all vehicles except the target vehicle are obtained, and the cosine similarity between the driving vectors of all other vehicles and the driving vector of the target vehicle is obtained as the driving similarity factor between all other vehicles and the target vehicle; DBSCAN clustering is performed on all vehicles except the target vehicle, and the distance measurement uses the absolute value of the difference between the driving similarity factors of all other vehicles and the target vehicle to obtain several clusters; the mean of the driving similarity factor between the vehicles in any cluster and the target vehicle is obtained, and the cluster with the largest mean of the driving similarity factor is taken as the same-cluster of the target vehicle; the number of lanes on the road where the target vehicle is located and the product of the minimum value of the driving similarity factor between the vehicles in the same-cluster and the target vehicle are obtained, and the ratio of the number of vehicles in the same-cluster to the product is taken as the vehicle density of the target vehicle.

[0028] It should be noted that the more other vehicles there are in the same cluster and the smaller the minimum value of the driving similarity factor, the more other vehicles with similar driving vectors there are to the target vehicle, and they are more likely to be distributed on the same road section. The smaller the number of corresponding lanes, the greater the vehicle density at the location of the target vehicle.

[0029] Furthermore, the road types in the highway map information of the vehicle's location include expressways, urban expressways, ordinary urban roads, and rural roads. This embodiment sets corresponding road type weights for the road types, specifically: expressways are 1, urban expressways are 2, ordinary urban roads are 3, and rural roads are 4; it should be noted that the road type weights of expressways, urban expressways, ordinary urban roads, and rural roads in this embodiment are preset according to the use requirements of the roads, so the specific weight values ​​can be adjusted according to actual needs or actual conditions, but the order of size should be ensured to be expressways < urban expressways < ordinary urban roads < rural roads to ensure subsequent calculation results.

[0030] Preferably, in one embodiment of the present invention, the road environment parameters of each vehicle are quantified by combining multiple types of vehicle driving information and local road map information, including the following specific methods: Based on the absolute value of the road slope, road curvature, road type weight and road speed limit value at the location of the target vehicle, the road driving factor of the target vehicle is obtained; the average driving speed of all vehicles in the target vehicle's peer cluster is used as the target vehicle's peer speed; based on the target vehicle's vehicle density, peer speed and road driving factor, the road environment parameters of the target vehicle are obtained; the road environment parameters are positively correlated with the road driving factor and the vehicle density, and negatively correlated with the peer speed; the road driving factor is positively correlated with the absolute value of the road slope and the road curvature, and negatively correlated with the road type weight and the road speed limit value.

[0031] As an example, the specific calculation method of the road environment parameters of the target vehicle is: in, Indicates the road environment parameters of the target vehicle; represents the vehicle density of the target vehicle; Indicates the target vehicle's speed; Indicates the curvature of the road where the target vehicle is located; Indicates the road slope where the target vehicle is located; Indicates the road type weight where the target vehicle is located; Indicates the speed limit of the road where the target vehicle is located; represents the road travel factor of the target vehicle; represents the absolute value function; represents a sigmoid function, which is used for normalization processing in this embodiment.

[0032] It should be noted that the road environment parameter is used to describe the degree of restriction on the vehicle's driving in the road environment on the road it is currently traveling. The larger the value of the road environment parameter, the higher the degree of restriction. That is, when the road environment is poor, the road type, speed limit, number of lanes, road curvature and slope restrict the vehicle from driving normally. At the same time, due to the high vehicle density and low average travel speed on the road, the vehicle cannot drive normally, which further affects the degree of restriction on vehicle driving, resulting in an increase in the degree of restriction on vehicle driving, and the corresponding road environment parameter is larger.

[0033] Preferably, in one embodiment of the present invention, weather information at the location of the vehicle is analyzed to quantify the vehicle's weather environment parameters; and the vehicle's driving environment coefficient is obtained by combining the vehicle's road environment parameters, including the following specific methods: In the weather information at the vehicle's location, this embodiment sets corresponding precipitation type weights for precipitation types according to the degree of precipitation, specifically: no precipitation is 1, light rain is 2, moderate rain is 3, heavy rain is 4, freezing rain is 5, snow is 6, and hail is 7; at the same time, corresponding road surface condition weights are set based on the road surface conditions, specifically: dry is 1, wet is 2, water accumulation is 3, snow accumulation is 4, and ice is 5; it should be noted that, similar to the road type weight, the precipitation type weight and the road surface condition weight in this embodiment are set according to the influence relationship on the slipperiness of the road. Therefore, the size of the precipitation type weight and the road surface condition weight can be adjusted according to actual needs or actual conditions, but the size relationship of no precipitation < light rain < moderate rain < heavy rain < freezing rain < snow < hail, and dry < wet < water accumulation < snow < ice should be ensured.

[0034] Furthermore, the sum of the precipitation type weight and the road condition weight at the target vehicle's location and the ratio of the visibility at the target vehicle's location are normalized and used as the weather environment parameter of the target vehicle; this embodiment uses the sigmoid function for normalization.

[0035] It should be noted that the weather environment parameter is used to describe the degree of restriction of the vehicle's driving process due to the weather environment at its current location. The larger the value of the weather environment parameter, the higher the degree of restriction. That is, when the weather conditions are poor, such as low visibility, slippery roads and other severe weather conditions that lead to vehicle driving restrictions, the corresponding degree of restriction on the vehicle's driving process will be higher, and the corresponding weather environment parameter will be larger.

[0036] Furthermore, the product of the weather environment parameter of the target vehicle and the road environment parameter is normalized and used as the driving environment coefficient of the target vehicle. In this embodiment, the sigmoid function is used for normalization.

[0037] It should be noted that the driving environment coefficient is used to describe the degree of restriction of the target vehicle under the combined restrictions of the road traffic conditions and weather conditions in which it is driving. The larger the value of the driving environment coefficient, the more the target vehicle's driving process is restricted by the current environmental conditions. When the driving behavior of the target vehicle is subsequently analyzed, when the vehicle's driving behavior exceeds the restriction range of the current environmental conditions, the vehicle behavior becomes more abnormal.

[0038] At this point, the vehicle's driving environment coefficient is obtained.

[0039] Step S003: Based on the vehicle driving information of the vehicle, as well as the weather information and road map information of the vehicle's location as environmental information, construct several vehicle driving environment information groups, and combine the similarity between the same type of information in the traffic accident data to obtain several vehicle-related accident information groups; analyze the correlation between the environmental information and the vehicle driving information in the relevant accident information groups and their proportion in the traffic accident data to obtain the accident correlation between each type of vehicle driving information and each type of environmental information of each vehicle.

[0040] It should be noted that in order to accurately obtain vehicle behavior detection results, it is necessary to further use the vehicle's driving information to perform behavior analysis, and analyze the vehicle behavior based on the vehicle's driving environment as a criterion to determine whether the behavior shown in the vehicle's driving information is within the range allowed by the vehicle's current driving environment.

[0041] Preferably, in one embodiment of the present invention, based on the vehicle driving information, and the weather information and road map information of the vehicle's location as environmental information, several vehicle driving environment information groups are constructed, and combined with the similarity relationship between the same type of information in the traffic accident data, several vehicle-related accident information groups are obtained, including the specific method of: The multi-type weather information and highway map information at the vehicle's location are collectively used as the multi-type environmental information of the vehicle. Any type of vehicle driving information and any type of environmental information are combined into a driving environment information group (the vehicle driving information is placed first, and the environmental information is placed second, forming a vector, with the values ​​of each dimension of the vector representing the element values ​​of the corresponding type; in particular, for example, precipitation type, road surface condition, road type, etc., the corresponding element values ​​are weighted accordingly). Several driving environment information groups are obtained for the target vehicle (element values ​​under the corresponding type of information of the target vehicle). For any driving environment information group of the target vehicle, the elements corresponding to the driving environment information group under each traffic accident data are obtained from the multiple traffic accident data as the accident information group of each traffic accident data in the driving environment information group. The cosine similarity between each accident information group and the driving environment information group of the target vehicle is obtained (cosine similarity is calculated using vectors). A relevant threshold is preset, and in this embodiment, the relevant threshold is 0.8. If the cosine similarity is greater than or equal to the relevant threshold, the corresponding accident information group is used as the relevant accident information group of the driving environment information group of the target vehicle. The relevant accident information group of each driving environment information group of the target vehicle is obtained according to the above method.

[0042] Preferably, in one embodiment of the present invention, the correlation between the environmental information and the vehicle driving information in the relevant accident information group and their proportion in the traffic accident data are analyzed to obtain the accident correlation between the various types of vehicle driving information and various types of environmental information, including the following specific methods: For any environmental information and any vehicle driving information, based on all relevant accident information groups of the target vehicle, the ratio (percentage) of the number of relevant accident information groups containing the environmental information and the vehicle driving information (corresponding type information, element values ​​of non-corresponding types) to the number of relevant accident information groups containing the environmental information shall be used as the conditional probability of the vehicle driving information occurring under the environmental information; the ratio of the number of relevant accident information groups containing the environmental information and the vehicle driving information to the total number of traffic accident data shall be used as the relative frequency of the environmental information and the vehicle driving information; the product of the conditional probability of the vehicle driving information occurring under the environmental information and the relative frequency of the environmental information and the vehicle driving information shall be used as the accident correlation factor of the vehicle driving information and the environmental information of the target vehicle; the accident correlation factors of various types of vehicle driving information and various types of environmental information of the target vehicle are obtained according to the above method, all accident correlation factors are linearly normalized, and the results obtained are used as the accident correlations of various types of vehicle driving information and various types of environmental information of the target vehicle.

[0043] It should be noted that by obtaining accident correlation, the strength of the association between specific driving behaviors and traffic accidents under specific environmental conditions is quantified. The conditional probability reflects the accident probability of the target vehicle's driving behavior in the corresponding environment based on previous accidents. The higher the probability, the less safe the behavior, the more abnormal the driving behavior, and the higher the probability of an accident. The relative frequency indicates the proportion of vehicle driving information and environmental information combinations that occur in all historical accidents. For example, the greater the number of relevant accident information groups (such as the frequent occurrence of "heavy rain + speeding" in accidents), the greater the relative frequency, indicating that this combination is a common cause of accidents. Physically, it reflects the representativeness of this behavior-environment combination in historical accidents. The higher the value, the more typical the combination is for high-risk scenarios. Therefore, the product of the conditional probability and the relative frequency combines the probability of the behavior in the environment and the historical representativeness of the combination in accidents, reflecting the dual dependence of risk. That is, whether a vehicle behavior is dangerous depends not only on the behavior itself, but also on whether the environment "amplifies" the risk of the behavior. For example, snowy and icy conditions can make speeding more deadly.

[0044] At this point, the accident correlation between the driving information of each vehicle and each type of environmental information is obtained.

[0045] Step S004: construct an accident space map for each vehicle based on the vehicle driving information and environmental information, as well as the accident correlation; and obtain the vehicle's driving risk coefficient in combination with the vehicle's driving environment coefficient and issue a dangerous behavior warning.

[0046] It should be noted that after obtaining the accident correlation between various types of vehicle driving information and environmental information, the accident correlation is compared through a threshold value, and this is used as an edge value to connect the driving nodes corresponding to various types of vehicle driving information and the environmental nodes corresponding to various types of environmental information. Then, an accident space graph is constructed by the driving nodes, environmental nodes, and connected edges. The accident space graph intuitively reflects the correlation between various types of vehicle driving information and environmental information, providing a basis for the subsequent comprehensive quantification of the driving risk coefficient based on the correlation performance combined with the driving environment coefficient.

[0047] Preferably, in one embodiment of the present invention, an accident space map of each vehicle is constructed based on the vehicle driving information and environmental information, as well as the accident correlation, including the specific method of: The target vehicle is taken as the central node, the driving information of various types of vehicles is taken as driving nodes, and the various types of environmental information is taken as environmental nodes. The central node and each driving node are connected. For any driving node and any environmental node, the accident correlation between the vehicle driving information corresponding to the driving node of the target vehicle and the environmental information corresponding to the environmental node is obtained. A correlation threshold is preset. In this embodiment, the correlation threshold is described as 0.7. If the accident correlation is greater than or equal to the correlation threshold, the driving node and the environmental node are connected, and the accident correlation is used as the edge value of the connecting edge; the connection judgment is performed on all driving nodes and environmental nodes based on the corresponding accident correlation, and the connection between each node is completed to form an accident space diagram of the target vehicle, as shown in FIG. Figure 3 As shown, it is a schematic diagram of the accident space graph structure of the vehicle, where, in order to facilitate the display of unlabeled edge values, the center of the accident space graph is the target vehicle, the first-layer nodes are driving nodes, and the second-layer nodes are environmental nodes; it should be noted that the structural schematic diagram of the accident space graph is only used to display the connection method between nodes in an intuitive way, and is not a limiting result of the connection relationship between nodes, that is, the connection relationship between nodes is constructed according to the specific method in this embodiment.

[0048] It should be noted that the corresponding edges between the driving nodes and the environmental nodes in the accident space graph reflect the correlation between the corresponding nodes in the possibility of accidents. Through topological processing, the vehicle AI platform can understand the risk of vehicle behavior based on the environmental context in the subsequent process based on the accident space graph, thereby effectively performing vehicle behavior detection.

[0049] It should be further explained that the accident relationship diagram intuitively reflects the relationship between vehicle driving information and environmental information. In the process of quantifying vehicle driving risks, the overall performance of the accident correlation between vehicle driving information and environmental information is used, and combined with the driving environment coefficient for comprehensive quantification. That is, in vehicle behavior detection, the environment is the basis of risk, and behavior is the core logic of risk triggering conditions. By utilizing the topological structure of the accident space diagram, historical traffic accident data is combined with the real-time environment and driving behavior of the target vehicle, avoiding the problem of unsatisfactory detection results caused by the solidification of traditional vehicle behavior detection methods.

[0050] Preferably, in one embodiment of the present invention, the driving risk coefficient of the vehicle is obtained and a dangerous behavior warning is performed in combination with the driving environment coefficient of the vehicle, including the specific method of: Obtain the mean of the accident correlation between the vehicle driving information corresponding to each driving node and the environmental information corresponding to each environmental node in the accident space diagram of the target vehicle, and multiply the mean by the driving environment coefficient of the target vehicle as the driving risk coefficient of the target vehicle; preset a risk threshold, and in this embodiment, the risk threshold is described as 0.7. If the driving risk coefficient of the target vehicle is greater than or equal to the risk threshold, the target vehicle has dangerous behavior, and the vehicle AI platform sends a dangerous behavior warning information to the target vehicle.

[0051] At this point, this embodiment is completed.

[0052] See also Figure 2 , which shows a vehicle behavior detection system based on a vehicle AI platform provided by another embodiment of the present invention, the system comprising: Vehicle information collection module 101: acquires various types of vehicle driving information and weather information and road map information of the location of the vehicle, transmits it to the vehicle AI platform, and records certain traffic accident data; Vehicle information analysis module 102: Based on the similarity between the driving information of multiple vehicles, the vehicle density of each vehicle is obtained, and the road type weight of the vehicle location is recorded; the road environment parameters of each vehicle are quantified by combining the driving information of multiple types of vehicles and the road map information of the location; the weather information of the vehicle location is analyzed to quantify the weather environment parameters of the vehicle; and the driving environment coefficient of the vehicle is obtained by combining the road environment parameters of the vehicle. Based on the vehicle's driving information, as well as weather information and road map information at the vehicle's location as environmental information, several vehicle driving environment information groups are constructed. Combined with similar relationships between the same types of information in traffic accident data, several vehicle-related accident information groups are obtained. The correlation between the environmental information and vehicle driving information in the relevant accident information groups and their proportion in the traffic accident data are analyzed to obtain the accident correlation between each type of vehicle driving information and each type of environmental information for each vehicle. Vehicle behavior detection module 103: constructs an accident space map for each vehicle based on vehicle driving information and environmental information, as well as the accident correlation; obtains the vehicle's driving risk coefficient in combination with the vehicle's driving environment coefficient and issues a dangerous behavior warning.

[0053] Another embodiment of the present invention proposes a vehicle behavior detection device based on a vehicle AI platform, which includes a memory, a processor, and a computer program stored in the memory and running on the processor. When the processor executes the computer program, steps S001 to S004 of the above method are implemented.

[0054] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A vehicle behavior detection method based on a vehicle AI platform, characterized in that: The method comprises the following steps: Obtain multiple types of vehicle driving information and weather information and road map information of the location of the vehicle, transmit them to the vehicle AI platform, and record certain traffic accident data; Based on the similarity between the driving information of several vehicles, the vehicle density of each vehicle is obtained, and the weight of the road type at the vehicle's location is recorded; the road environment parameters of each vehicle are quantified by combining the driving information of multiple types of vehicles and the highway map information at the location; the weather information at the vehicle's location is analyzed to quantify the vehicle's weather environment parameters; and the vehicle's driving environment coefficient is obtained by combining the vehicle's road environment parameters; Based on the vehicle's driving information, as well as weather information and road map information at the vehicle's location as environmental information, several vehicle driving environment information groups are constructed. Combined with similar relationships between the same types of information in traffic accident data, several vehicle-related accident information groups are obtained. The correlation between the environmental information and vehicle driving information in the relevant accident information groups and their proportion in the traffic accident data are analyzed to obtain the accident correlation between each type of vehicle driving information and each type of environmental information for each vehicle. Based on the vehicle driving information and environmental information, as well as the accident correlation, an accident space map of each vehicle is constructed; combined with the vehicle's driving environment coefficient, the vehicle's driving risk coefficient is obtained and a dangerous behavior warning is issued; The vehicle density of each vehicle is specifically obtained as follows: Take any vehicle as the target vehicle, and construct the target vehicle's driving vector using the real-time positioning information and driving direction in the target vehicle's driving information; obtain the driving vectors of all vehicles except the target vehicle, and obtain the cosine similarity between the driving vectors of all other vehicles and the driving vector of the target vehicle as the driving similarity factor between the other vehicles and the target vehicle; Perform DBSCAN clustering on all vehicles except the target vehicle. The distance metric is the absolute value of the difference between the driving similarity factors of each vehicle and the target vehicle, resulting in several clusters. The mean driving similarity factor between the vehicles in any cluster and the target vehicle is obtained, and the cluster with the largest mean driving similarity factor is selected as the peer cluster of the target vehicle. The product of the number of lanes on the road where the target vehicle is located and the minimum value of the driving similarity factor between the vehicles in the same cluster and the target vehicle is obtained, and the ratio of the number of vehicles in the same cluster to the product is used as the vehicle density of the target vehicle.

2. The vehicle behavior detection method based on the vehicle AI platform according to claim 1 is characterized in that: The specific method for obtaining the road environment parameters of each vehicle is as follows: Based on the absolute value of the road slope, road curvature, road type weight, and road speed limit value at the target vehicle's location, the target vehicle's road driving factor is obtained; the average driving speed of all vehicles in the target vehicle's peer cluster is used as the target vehicle's peer speed; Based on the vehicle density, peer speed and road driving factor of the target vehicle, the road environment parameters of the target vehicle are obtained; The road environment parameter is positively correlated with the road driving factor and the vehicle density, and negatively correlated with the traveling speed; the road driving factor is positively correlated with the absolute value of the road slope and the road curvature, and negatively correlated with the road type weight and the road speed limit.

3. The vehicle behavior detection method based on the vehicle AI platform according to claim 1 is characterized in that: The quantified vehicle weather environment parameters are combined with the vehicle road environment parameters to obtain the vehicle driving environment coefficient, including the specific method of: Set corresponding precipitation type weights for each precipitation type; set corresponding road condition weights based on road conditions; The sum of the precipitation type weight and the road condition weight at the target vehicle's location and the ratio of the visibility at the target vehicle's location are normalized and used as the weather environment parameter of the target vehicle; The product of the target vehicle's weather environment parameters and road environment parameters and the normalized result is used as the target vehicle's driving environment coefficient.

4. The vehicle behavior detection method based on the vehicle AI platform according to claim 1 is characterized in that: The specific method of obtaining several relevant accident information groups of the vehicle includes: Combine multiple types of weather information and road map information at the vehicle's location as multiple types of vehicle environmental information; combine any type of vehicle driving information and any type of environmental information as a driving environment information group, and obtain multiple driving environment information groups for the target vehicle; For any driving environment information group of the target vehicle, obtaining the element corresponding to the driving environment information group under each traffic accident data from a plurality of traffic accident data as the accident information group of each traffic accident data in the driving environment information group; Obtain the cosine similarity between each accident information group and the target vehicle's driving environment information group. If the cosine similarity is greater than or equal to a relevant threshold, use the corresponding accident information group as the relevant accident information group of the target vehicle's driving environment information group; obtain the relevant accident information group of each driving environment information group of the target vehicle.

5. The vehicle behavior detection method based on the vehicle AI platform according to claim 1 is characterized in that: The specific method for obtaining the accident correlation between the various types of vehicle driving information and various types of environmental information includes: For any environmental information and any vehicle driving information, based on all relevant accident information groups of the target vehicle, the ratio of the number of relevant accident information groups containing the environmental information and the vehicle driving information to the number of relevant accident information groups containing the environmental information is used as the conditional probability of the vehicle driving information occurring under the environmental information; The ratio of the number of relevant accident information groups including the environmental information and the vehicle driving information to the total number of traffic accident data is used as the relative frequency of the environmental information and the vehicle driving information; The product of the conditional probability of the vehicle driving information occurring under the environmental information and the relative frequency of the environmental information and the vehicle driving information is used as the accident correlation factor of the vehicle driving information and the environmental information of the target vehicle; Obtain accident correlation factors of various types of vehicle driving information and various types of environmental information of the target vehicle, perform linear normalization on all accident correlation factors, and obtain the results as accident correlations of various types of vehicle driving information and various types of environmental information of the target vehicle.

6. The vehicle behavior detection method based on the vehicle AI platform according to claim 1 is characterized in that: The specific method of constructing the accident space graph of each vehicle includes: The target vehicle is taken as the central node, the driving information of various types of vehicles is taken as driving nodes, and the various types of environmental information is taken as environmental nodes. The central node is connected to each driving node. For any driving node and any environmental node, the accident correlation between the vehicle driving information corresponding to the driving node of the target vehicle and the environmental information corresponding to the environmental node is obtained. If the accident correlation is greater than or equal to the association threshold, the driving node and the environmental node are connected, and the accident correlation is used as the edge value of the connecting edge. All driving nodes and environmental nodes are connected and judged based on the corresponding accident correlation, and the connection between each node is completed to form the accident space diagram of the target vehicle.

7. The vehicle behavior detection method based on the vehicle AI platform according to claim 1 is characterized in that: The specific method for obtaining the driving risk coefficient of the vehicle is as follows: The mean of the accident correlation between the vehicle driving information corresponding to each driving node and the environmental information corresponding to each environmental node in the accident space diagram of the target vehicle is obtained, and the product of the mean and the driving environment coefficient of the target vehicle is used as the driving risk coefficient of the target vehicle.

8. The vehicle behavior detection system based on the vehicle AI platform is characterized by: When the system is running, the steps of the vehicle behavior detection method based on the vehicle AI platform according to any one of claims 1 to 7 are implemented, and the system includes: The vehicle information collection module is used to obtain various types of vehicle driving information and weather information and road map information of the location, transmit it to the vehicle AI platform, and record certain traffic accident data; The vehicle information analysis module is used to obtain the vehicle density of each vehicle based on the similar performance of the vehicle driving information of multiple vehicles, and record the road type weight of the vehicle's location; combine the driving information of multiple types of vehicles with the local highway map information to quantify the road environment parameters of each vehicle; analyze the weather information of the vehicle's location to quantify the vehicle's weather environment parameters; and combine the vehicle's road environment parameters to obtain the vehicle's driving environment coefficient; Based on the vehicle's driving information, as well as weather information and road map information at the vehicle's location as environmental information, several vehicle driving environment information groups are constructed. Combined with similar relationships between the same types of information in traffic accident data, several vehicle-related accident information groups are obtained. The correlation between the environmental information and vehicle driving information in the relevant accident information groups and their proportion in the traffic accident data are analyzed to obtain the accident correlation between each type of vehicle driving information and each type of environmental information for each vehicle. The vehicle behavior detection module is used to construct an accident space map for each vehicle based on vehicle driving information and environmental information, as well as the accident correlation; combined with the vehicle's driving environment coefficient, it obtains the vehicle's driving risk coefficient and issues dangerous behavior warnings.

9. A vehicle behavior detection device based on a vehicle AI platform, comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that: When the processor executes the computer program, the steps of the vehicle behavior detection method based on the vehicle AI platform as described in any one of claims 1 to 7 are implemented.

Images