Search

CN-121982896-A - Data skeleton matching method and system based on intelligent Internet of vehicles environment

CN121982896ACN 121982896 ACN121982896 ACN 121982896ACN-121982896-A

Abstract

The invention particularly discloses a data skeleton matching method and a system based on an intelligent vehicle networking environment, wherein the method comprises the steps of classifying vehicles according to automatic driving grades of the vehicles, calculating specific data demand intensity, scene emergency score and communication constraint degree of the vehicles, constructing a quantization model based on space vision grade demands, calculating vision demand factors and determining space vision grade demands; the method comprises the steps of carrying out global fusion on the automatic driving grade of the vehicle, the specific data demand intensity of the vehicle, the scene emergency grade, the communication constraint degree and the visual field demand, outputting the precision grade and the scale of a final data skeleton, and packaging the perceived data by a road side unit according to the precision grade and the scale of the final data skeleton and sending the perceived data to a requesting vehicle. By adopting the technical scheme, individuation and accurate matching of data response in the vehicle-road cooperative system are realized through a multi-level self-adaptive decision process, and the utilization rate of communication resources, the real-time performance of data transmission and the driving safety of vehicles are improved.

Inventors

  • ZENG LINGQIU
  • HUANG YONG
  • LEI JIANMEI
  • HAN QINGWEN
  • YE LEI
  • YANG HAO

Assignees

  • 重庆大学

Dates

Publication Date
20260505
Application Date
20260304

Claims (10)

  1. 1. The data skeleton matching method based on the intelligent Internet of vehicles environment is characterized by comprising the following steps of: according to the automatic driving grade of the vehicle, classifying the types of the vehicle; Aiming at an auxiliary driving vehicle and an intelligent driving vehicle in the vehicle types, constructing a demand intensity quantization model based on the perception capability of the vehicle, and calculating the specific data demand intensity of the vehicle by combining the behavior data and the equipment conditions carried in the request vehicle; constructing a quantitative model based on scene and communication environment requirements, and calculating scene urgency scores and communication constraint degrees; Constructing a quantitative model based on the space visual field level requirement, calculating a visual field requirement factor, and determining the space visual field level requirement according to a visual field division threshold; Based on a fuzzy reasoning model, carrying out global fusion on the automatic driving level of the vehicle, the specific data demand intensity of the vehicle, the scene emergency score, the communication constraint degree and the space vision level demand, and outputting the precision level and the scale of a final data skeleton; and packaging the perceived data by the road side unit according to the precision grade and the scale of the final data skeleton and sending the perceived data to a requesting vehicle.
  2. 2. The data skeleton matching method based on the intelligent internet of vehicles environment according to claim 1, wherein the types of vehicles are classified according to the automatic driving level of the vehicles, specifically: When a terminal vehicle sends a request Q for acquiring surrounding environment information to a road side unit, the road side unit firstly judges the automatic driving level of the vehicle, namely, reads an automatic_level field in the Q, takes values of 0 to 5 and respectively corresponds to L0 to L5; The road side units are divided into three categories according to the own grade information provided by the vehicle in Q: the low-grade vehicle, namely the common vehicle, corresponds to L0, is expressed as a whole-course manual driving vehicle, has no automatic auxiliary system, and has the requirements for environment information as basic safety warning and low-precision perception; the medium-level vehicle, namely the auxiliary driving vehicle, corresponds to L1-L2, has conditional automatic driving capability and needs medium-precision environmental data to support system decision; The vehicles of high grade, namely intelligent driving vehicles, correspond to L3-L5, belong to automatic driving vehicles, and high-precision environment skeleton data are needed to support full-automatic driving.
  3. 3. The data skeleton matching method based on the intelligent internet of vehicles environment according to claim 2, wherein the method is characterized by constructing a demand intensity quantization model based on the perception capability of the vehicle, and calculating the specific data demand intensity of the vehicle by combining the behavior data and the equipment condition carried in the request vehicle, and comprises the specific steps of: For driving behavior, the current instantaneous speed of the vehicle is defined as The current instantaneous acceleration is The current steering angular velocity is ; For the configuration of the vehicle-mounted equipment, defining a sensor configuration score S_score, wherein the score is a static score obtained by comprehensively calculating the type, the number and the performance factors of sensors carried by the vehicle, and the score is carried by a message Q sent by the vehicle; the urgent degree D_score of the current running state on the real-time performance and the precision of the environmental information is calculated as follows: , Wherein, the , , The normalization reference value is preset; , , Is a weight coefficient, satisfies The larger the D_score value is, the stronger the driving behavior is required for high-precision data; For the sensor equipment carried by the vehicle, the road side unit receives and analyzes the dynamic behavior parameters reported by the request vehicle in real time, and meanwhile, the sensor configuration score S_score calculated by the vehicle is obtained from the vehicle registration information or the request message Q, specifically: basic camera = 1 minute, millimeter wave radar = 2 minutes, laser radar = 3 minutes, take the sum of all vehicle sensor scores and normalize to [0,1], get s_score: , Wherein, the In order to set the types of sensors, Is the first The weight coefficient of the class sensor, In order to be able to do this in a practical number, At the upper limit of the effective amount, The lower the score, the weaker the self-perception capability of the vehicle is, and the stronger the dependence on external high-precision data is, whereas the higher the score is, the vehicle is provided with a complete perception system, and the road side only needs to provide light-weight cooperative information; the differential weighting strategy is adopted for the L1-L2 assisted driving vehicle and the L3-L5 advanced intelligent driving vehicle: For the auxiliary driving vehicles (L1-L2), the data demand intensity based on the perception capability is determined by the device configuration of the auxiliary driving vehicles, so the device configuration weight of the L1-L2 auxiliary driving vehicles is higher than that of the L3-L5 advanced intelligent driving vehicles: , wherein λ is the d_score and s_score weight factors, adjusted according to the actual situation; The integrated demand strength calculated for the L1 and L2 vehicles; For highly intelligent vehicles (L3-L5), the highly intelligent vehicle defaults to its sensor configuration to complete, so the highly intelligent vehicle will focus on its own driving state: , Wherein, the Is the integrated demand strength calculated for the L3-L5 vehicle.
  4. 4. The method for matching data skeleton based on intelligent internet of vehicles environment according to claim 1, wherein a quantized model based on scene and communication environment requirements is constructed, scene urgency score and communication constraint degree are calculated, and a road side unit calculates scene urgency score e_score in real time, which is: , Wherein S critical is scene criticality, a score is preset according to the position of the vehicle, D traffic is traffic density, normalization is needed, and the method comprises the following steps: , wherein W impact is weather or illumination influencing factor, A priority represents application priority; Is the corresponding weight coefficient; the road side unit evaluates the current communication link quality and calculates the communication constraint degree C_constraint: , wherein, B available is the current available bandwidth, B required is the reference bandwidth requirement, η loss and η delay are the packet loss rate and the delay impact factor respectively, which are: , 。
  5. 5. The method for matching data skeleton based on intelligent car networking environment according to claim 1, wherein the method for constructing a quantization model based on space vision level requirements, calculating vision requirement factors and determining space vision level requirements according to vision division threshold values is as follows: Defining a view requirement factor v_need as: , Wherein, the For the current vehicle speed, The speed limit is the road speed limit; the minimum curvature radius of the road in front is set, and the straight road is infinite; is the local traffic density; And Is a weight coefficient; Setting a threshold value for dividing the field of view When (when) If the micro-view field is in the mesoscopic view field requirement, otherwise, the micro-view field requirement is met; if the road side unit is a microscopic view, the road side unit should transmit back high-precision (namely, space positioning error <10cm, end-to-end transmission delay <50ms, object attribute dimension including speed, acceleration, contour and micro motion state) small skeleton information, focus on the object in the direct sensing range of the vehicle, and provide detailed information of centimeter-level positioning and millisecond-level delay; If the mesoscopic view is calculated, the road side unit should provide information in a far range from the vehicle (namely, in the space dimension, the effective detection distance exceeds the effective detection distance of the sensor of the vehicle, the specific value is positively correlated with the current vehicle speed, or in the time dimension, the road side unit corresponds to the planned reachable space range of the vehicle running at the current speed for 5 to 15 seconds in the future).
  6. 6. The data skeleton matching method based on the intelligent internet of vehicles environment according to claim 5, wherein the method is characterized by carrying out global fusion on the automatic driving level of the vehicle, the specific data demand intensity of the vehicle, the scene emergency score, the communication constraint degree and the space view level demand based on the fuzzy inference model, and outputting the precision level and the scale of the final data skeleton, and comprises the following specific steps: defining system input variables, constructing an input matrix, wherein the input X is expressed as: , Wherein, the Represents the vehicle class (L0-L5), Representing the data demand intensity ), Represents scene urgency (E score), Represents the communication constraint (C _ constraint), Indicating the field of view level (microscopic, mesoscopic); For continuous variable , , Fuzzification is carried out by adopting a triangle membership function, and each variable corresponds to three fuzzy sets, which are specifically defined as follows: data demand intensity R fuzzification based on vehicle perceptibility, fuzzy set is defined as { low (L), medium (M), high (H) }, membership function is defined as ; Blurring of scene urgency E, fuzzy set { loose (L), medium (M), urgency (H) }, membership function defined as ; Fuzzification of communication constraint Cc, fuzzy set is { good (G), limited (L), bad (B) }, membership function is defined as ; Defining P as the precision level (low, medium, high) of the final data skeleton, S as the final skeleton size (small, medium, large), wherein the triangle membership function of P is defined as low (0,0,0.3), medium (0.2,0.5,0.8), high (0.7,1,1), and S as small (0,0,0.3), medium (0.2,0.5,0.8), large (0.7,1,1); The output of fuzzy reasoning is defined as the matrix Y: , Wherein, the Representing the precision grade of the finally output data skeleton; representing the size of the final output data skeleton; For each variable element in the input matrix X, a rule is activated: rule R1: IF (C is low grade vehicle) ) THEN (P is low, S is large); rule R2: IF (R is high) AND (eis medium) AND (Cc is limited) AND (Lv is mesoscopic), THEN (P is, S is large); Rule R3: IF (rsi high) AND (eis medium) AND (Cc is general) AND (Lv is microscopic), THEN (p×is high, s×is); Rule R4: IF (in ris) AND (eis medium) AND (Cc is limited) AND (Lv is microscopic), THEN (P is, S is small); Rule R5: IF (in ris) AND (eis urgent) AND (Cc is limited) AND (Lv is mesoscopic), THEN (p×is high, s×is); For the ith rule, determining the activation intensity of each rule by adopting a minimum value method, and determining the activation intensity of each rule For all its preconditions minimum value of membership: , Wherein, the Under the expression rule i, if the precondition comprises the data demand intensity R, the membership degree calculated value is equal to the data demand intensity R; under the expression rule i, if the precondition comprises a scene emergency degree E, the membership degree calculation value is equal to the scene emergency degree E; respectively representing that under the rule i, if the corresponding precondition exists, the corresponding membership value is equal to the corresponding membership value; Representing the minimum value of membership of all preconditions under rule i; For each output variable, let P's argument be [0,1], and discretize P's argument into m points For the following The point: Clipping the output of each rule: , Wherein, the Is P in point in the rule conclusion of the ith A value obtained by the corresponding output membership function; the activation intensity for rule i; Representing the output of the rule i after clipping; aggregating all rule outputs: , Wherein, the Representing the output of rule i after clipping; Is indicated at the point Taking the maximum value of all rule outputs; deblurring by barycentric method: , for final data accuracy, mapping the sharpness values to specific gear: , , wherein, T 1 and T 2 are preset thresholds; And The precision and skeleton size gear of the final data are obtained; Final roadside unit transmission The data skeleton information is given to the requesting vehicle.
  7. 7. The method for matching data skeleton based on intelligent Internet of vehicles environment according to claim 6, wherein the data demand intensity R based on the vehicle perception capability is fuzzified, the fuzzy set is defined as { low (L), medium (M), high (H) }, and the membership function is defined as The method specifically comprises the following steps: Fuzzy set is a low corresponding membership function, where r 1 is the starting point of low demand intensity and r 2 is the maximum point of low demand intensity: , Fuzzy sets are membership functions corresponding to the medium, wherein r 1 and r 2 are transition points for starting and ending the medium demand intensity, and r3 and r4 are starting and ending points for transitioning to high precision: , fuzzy sets are highly corresponding membership functions, where r 3 and r 4 are the starting and ending points of high demand intensity: 。
  8. 8. The method of claim 6, wherein the scene urgency E is blurred, the fuzzy set is { loose (L), medium (M), urgent (H) }, and the membership function is defined as The method specifically comprises the following steps: Fuzzy sets are loosely corresponding membership functions, where e 1 and e 2 are the starting and maximum points of a relaxed scene, respectively: , the fuzzy set is a moderate corresponding membership function, where e 1 to e 4 represent emergency transitions for a moderate scene: , Fuzzy sets are membership functions for emergency correspondence, where e 3 and e 4 represent the starting and maximum points of the emergency scene: 。
  9. 9. The method for matching data skeleton based on intelligent Internet of vehicles environment according to claim 6, wherein the communication constraint degree Cc is fuzzified, the fuzzy set is { good (G), limited (L), bad (B) }, and the membership function is defined as The method specifically comprises the following steps: The fuzzy set is a membership function with good correspondence, wherein c 1 and c 2 are a starting point and a maximum point of good communication constraint, respectively: , Fuzzy sets are membership functions corresponding to a constraint, where c 1 to c 4 are transition intervals of a constrained scene: , The fuzzy set is a membership function corresponding to the severe, wherein c 3 and c 4 are a starting point and a maximum point of the severe communication degree respectively: 。
  10. 10. A data skeleton matching system based on the method of any one of claims 1-9, deployed at a roadside unit, comprising: the vehicle classification module is used for classifying the vehicles according to the automatic driving grades of the vehicles; The demand analysis module is used for constructing a demand intensity quantization model based on the perception capability of the vehicle so as to calculate the specific data demand intensity of the vehicle; the constraint evaluation module is used for calculating scene emergency degree scores and communication constraint degrees; the visual field level demand module is used for constructing a quantitative model based on the spatial visual field level demand, calculating a visual field demand factor and determining the spatial visual field level demand according to a visual field division threshold; the framework decision module is used for carrying out global fusion on the automatic driving grade of the vehicle, the specific data demand intensity of the vehicle, the scene emergency score, the communication constraint degree and the space vision grade demand based on the fuzzy reasoning model, and outputting the precision grade and the scale of the final data framework; And the data packaging module is used for packaging the perception data according to the precision grade and the scale of the final data skeleton and sending the perception data to the vehicle.

Description

Data skeleton matching method and system based on intelligent Internet of vehicles environment Technical Field The invention belongs to the technical field of intelligent transportation, and relates to a data skeleton matching method and system based on an intelligent Internet of vehicles environment. Background With the acceleration of the urban process and the continuous increase of traffic demand, the problems of traffic safety, congestion, energy consumption and the like are increasingly outstanding, and the rapid development of an Intelligent Transportation System (ITS) is promoted. The core goal of intelligent traffic is to improve road use efficiency and driving safety through information sharing and collaborative decision-making, and under the background, the limitation of single main body (vehicle or road side) perception and decision-making is difficult to meet the safety requirement in complex environments. Based on the method, the vehicle-road cooperative system realizes information interconnection by integrating dynamic traffic information of people, vehicles, roads and clouds and utilizing a vehicle network (Vehicle to Everything, V2X) communication technology, so that traffic information such as vehicle conditions, road conditions and the like can be shared, and vehicles and other traffic participants can be effectively assisted to make safer and more efficient decisions. However, with the rapid development of intelligent internet-connected vehicles (ICV) and road collaboration technologies, conventional internet-of-vehicles data interaction mechanisms increasingly expose a bottleneck and a deficiency to some extent. Firstly, a road side unit in the current internet of vehicles system mostly adopts a unified data packet response mode, when the road side unit responds to surrounding environment information of a request vehicle, the perception capability, the automatic driving level or the actual demand of the vehicle are not distinguished, so that the information redundancy is serious, the bandwidth resource waste is caused, and the real-time performance is difficult to guarantee. Secondly, along with the improvement of the perception capability and the intelligent level of the vehicle, the requirements of different vehicles on the precision-granularity of the environmental information are obviously different, but the existing system lacks an effective packaging and matching method for the data to distinguish what kind of fine degree data the different vehicles should receive. In order to meet the different demands of different vehicles on data, a matching method capable of dynamically responding to vehicle requests and organizing the vehicle requests into corresponding data frameworks according to data precision and scale is needed. The method needs to realize flexible adjustment of the precision and scale of the environmental information required by the vehicle on the basis of ensuring the bandwidth efficiency, the communication instantaneity and the data credibility. Disclosure of Invention The invention aims to solve the problems in the prior art and provides a data skeleton matching method and system based on an intelligent Internet of vehicles environment. In order to achieve the purpose, the basic scheme of the invention is that the data skeleton matching method based on the intelligent Internet of vehicles environment comprises the following steps: according to the automatic driving grade of the vehicle, classifying the types of the vehicle; Aiming at an auxiliary driving vehicle and an intelligent driving vehicle in the vehicle types, constructing a demand intensity quantization model based on the perception capability of the vehicle, and calculating the specific data demand intensity of the vehicle by combining the behavior data and the equipment conditions carried in the request vehicle; constructing a quantitative model based on scene and communication environment requirements, and calculating scene urgency scores and communication constraint degrees; Constructing a quantitative model based on the space visual field level requirement, calculating a visual field requirement factor, and determining the space visual field level requirement according to a visual field division threshold; Based on a fuzzy reasoning model, carrying out global fusion on the automatic driving level of the vehicle, the specific data demand intensity of the vehicle, the scene emergency score, the communication constraint degree and the space vision level demand, and outputting the precision level and the scale of a final data skeleton; and packaging the perceived data by the road side unit according to the precision grade and the scale of the final data skeleton and sending the perceived data to a requesting vehicle. The basic scheme has the working principle and beneficial effects that the technical scheme analyzes real-time requests sent by vehicles, the data skeleton matching process takes the automatic d