Search

CN-122009163-A - Control method of vehicle and vehicle

CN122009163ACN 122009163 ACN122009163 ACN 122009163ACN-122009163-A

Abstract

The embodiment of the application provides a vehicle control method and a vehicle, wherein the method comprises the steps of responding to running of the vehicle in a first running mode according to a running direction, obtaining image data of at least one environment object in the running direction and point cloud data of the environment object in the environment where the vehicle is located, determining first confidence coefficient of the image data, second confidence coefficient of the point cloud data and third confidence coefficient, determining a second running mode to be entered into by the vehicle based on the first confidence coefficient, the second confidence coefficient and the third confidence coefficient, wherein the probability of collision between the vehicle and the environment object in the second running mode is smaller than that of collision between the vehicle and the environment object in the first running mode, controlling the vehicle to switch from the first running mode to the second running mode, and controlling the vehicle in the second running mode to run based on a risk coefficient of the vehicle in the second running mode. The application solves the technical problem of low control accuracy of the vehicle.

Inventors

  • LIU JIA

Assignees

  • 奇瑞汽车股份有限公司

Dates

Publication Date
20260512
Application Date
20260212

Claims (10)

  1. 1. A control method of a vehicle, characterized by comprising: responding to running of the vehicle in a first running mode according to a running direction, and acquiring image data of at least one environment object in the running direction and point cloud data of the environment object in the environment where the vehicle is located; Determining a first confidence coefficient of the image data, a second confidence coefficient of the point cloud data and a third confidence coefficient, wherein the first confidence coefficient is used for representing the completeness degree of identifying the environmental object from the image data, the second confidence coefficient is used for representing the credibility degree of identifying the number of the environmental objects from the point cloud data, and the third confidence coefficient is used for representing the similarity degree between attribute information of identifying the same environmental object from the image data and the point cloud data respectively; Determining a second operation mode to be entered by the vehicle based on the first confidence, the second confidence and the third confidence, wherein the probability of the vehicle colliding with the environmental object in the second operation mode is smaller than the probability of the vehicle colliding with the environmental object in the first operation mode; Controlling the vehicle to switch from the first operation mode to the second operation mode; Controlling the vehicle in the second operation mode to run based on a risk coefficient of the vehicle in the second operation mode, wherein the risk coefficient is used for representing the risk degree of the collision.
  2. 2. The method according to claim 1, wherein the method further comprises: Acquiring a mode type of the second operation mode, wherein different mode types correspond to different safety degrees of the vehicle; and determining the risk coefficient according to a determination strategy corresponding to the mode type.
  3. 3. The method of claim 2, wherein the pattern types include a first pattern type and a second pattern type, the security level corresponding to the second pattern type being less than the security level corresponding to the first pattern type, determining the risk factor according to a determination policy corresponding to the pattern type, comprising: Identifying a first candidate environment object set from the image data in response to the mode type being the first mode type or the second mode type, and performing cluster tracking on the point cloud data to obtain a second candidate environment object set; Determining at least one environmental object based on the first candidate environmental object set and the second candidate environmental object set, wherein the environmental object is the first candidate environmental object set or the second candidate environmental object set, and affects the candidate environmental object of the vehicle running according to the running direction; And inputting at least one environmental object into a risk quantification model of the vehicle, and carrying out risk prediction on at least one environmental object by utilizing the risk quantification model to obtain the risk coefficient under the first mode type or the second mode type, wherein the risk quantification model is obtained by training a neural network model.
  4. 4. A method according to claim 3, wherein performing risk prediction on at least one of the environmental objects using the risk quantification model to obtain the risk coefficient comprises: determining a collision time of the vehicle and at least one environmental object by using the risk quantification model; Determining a target type weight of at least one environmental object and a traffic channel relation corresponding to at least one environmental object by using the risk quantification model, wherein the target type weight is used for representing the dangerous degree of the collision event between the vehicle in the second operation mode and the environmental object of different types, and the traffic channel relation is used for representing the position relation between the traffic channel where the environmental object is located and the traffic channel where the vehicle is located; and carrying out comprehensive risk prediction on at least one environmental object based on the collision time, the target type weight and the traffic channel relation by using the risk quantification model to obtain the risk coefficient under the first mode type or the second mode type.
  5. 5. The method of claim 2, wherein the security level corresponding to the third pattern type is less than the security level corresponding to the first pattern type, and the security level corresponding to the third pattern type is greater than the security level corresponding to the second pattern type, and determining the risk factor according to the determination policy corresponding to the pattern type comprises: Identifying a first set of candidate environmental objects from the image data in response to the pattern type being the third pattern type; performing cluster tracking on the point cloud data to obtain a second candidate environment object set; Determining at least one environmental object based on the first candidate environmental object set and the second candidate environmental object set, wherein the environmental object is the first candidate environmental object set or the second candidate environmental object set, and affects the candidate environmental object of the vehicle running according to the running direction; inputting at least one environmental object into a risk quantification model and a standby risk model of the vehicle, performing risk prediction on at least one environmental object by using the risk quantification model to obtain a first risk coefficient, and performing risk prediction on at least one environmental object by using the standby risk model to obtain a second risk coefficient, wherein the operand of the standby risk model is smaller than that of the risk quantification model; and determining the maximum value of the first risk coefficient and the second risk coefficient as the risk coefficient in the third mode type.
  6. 6. The method of claim 5, wherein performing risk prediction on at least one of the environmental objects using the alternate risk model to obtain a second risk factor comprises: Determining a collision time of the vehicle and at least one environmental object by using the standby risk model; and determining the second risk coefficient based on the collision time by using the standby risk model.
  7. 7. The method of claim 1, wherein the mode type of the second operating mode is a first mode type, a second mode type, and a third mode type, the degree of security corresponding to the third mode type being less than the degree of security corresponding to the first mode type, and the degree of security corresponding to the third mode type being greater than the degree of security corresponding to the second mode type, determining a second operating mode to be entered by the vehicle based on the first confidence, the second confidence, and the third confidence, comprising: Determining that the mode type of the second mode of operation is the first mode type in response to the first confidence being greater than a confidence threshold and the second confidence being greater than the confidence threshold and the third confidence being greater than the confidence threshold; Determining that the mode type of the second operation mode is the second mode type in response to one of the first confidence level, the second confidence level, and the third confidence level being less than or equal to the confidence level threshold; And determining that the mode type of the second operation mode is the third mode type in response to at least two of the first confidence level, the second confidence level and the third confidence level being less than or equal to the confidence level threshold.
  8. 8. The method of any of claims 1 to 7, wherein determining a first confidence level of the image data comprises: Determining an image information entropy of the image data; determining that the first confidence level is less than or equal to a confidence threshold value in response to the image information entropy representing the integrity of the environmental object in the image data being less than or equal to an integrity threshold value; determining that the first confidence level is greater than the confidence threshold in response to the image information entropy indicating that the integrity level is greater than the integrity level threshold; And/or the number of the groups of groups, Determining a second confidence of the point cloud data, comprising: acquiring the number of the environmental objects respectively contained in the plurality of point cloud data in a target time period; determining that the second confidence level is less than or equal to the confidence threshold in response to the rate of change of the number being greater than or equal to a rate of change threshold within the target time period; Determining that the second confidence level is greater than the confidence level threshold in response to the rate of change being less than the rate of change threshold for the target period of time; And/or the number of the groups of groups, Determining a third confidence level, comprising: Acquiring first attribute information of the environmental object from the image data, and acquiring second attribute information of the environmental object from the point cloud data, wherein the first attribute information is used for representing the appearance state of the environmental object identified from the image data, and the second attribute information is used for representing the appearance state of the environmental object identified from the point cloud data; Determining that the third confidence level is less than or equal to the confidence level threshold in response to a degree of difference between the first attribute information and the second attribute information of the same environmental object being greater than or equal to a difference threshold; In response to the variance being less than the variance threshold, determining that the third confidence is greater than the confidence threshold.
  9. 9. The method according to any one of claims 1 to 7, characterized in that controlling the vehicle running in the second operation mode based on a risk factor of the vehicle in the second operation mode comprises: Controlling the vehicle in the second operation mode to send prompt information in response to the risk coefficient being larger than a first risk coefficient threshold value, and controlling the vehicle in the second operation mode to run according to the prompt information, wherein the prompt information is used for prompting the approaching degree between the vehicle and the environment object to a driving object in the vehicle; Controlling the vehicle in the second operation mode to run according to a first braking strategy in response to the risk coefficient being greater than a second risk coefficient threshold, wherein the second risk coefficient threshold is greater than the first risk coefficient threshold; And controlling the vehicle in the second running mode to run according to a second braking strategy in response to the risk coefficient being greater than a third risk coefficient threshold, wherein the third risk coefficient threshold is greater than the second risk coefficient threshold, and the degree of urgency of the second braking strategy is greater than the degree of urgency of the first braking strategy.
  10. 10. A vehicle, characterized by comprising: A memory storing an executable program; a processor for executing the program, wherein the program when run performs the method of any of claims 1 to 9.

Description

Control method of vehicle and vehicle Technical Field The embodiment of the application relates to the technical field of vehicle safety control, in particular to a vehicle control method and a vehicle. Background At present, a collision early warning system for a vehicle generally lacks self-monitoring capability, and when a camera in the vehicle is stained, hard light or shielded or a radar is interfered to cause perception failure, the collision early warning system cannot sense failure, and still judges based on abnormal data output risk, so that 'silence failure/SILENTLY FAIL' is caused, false braking or missed braking is caused, and the functional safety requirement of the vehicle is violated. Therefore, there is still a technical problem that the control accuracy of the vehicle is low. There is currently no good solution to the above problems. Disclosure of Invention The embodiment of the application provides a vehicle control method and a vehicle, which are used for at least solving the technical problem of low control accuracy of the vehicle. According to one aspect of the embodiment of the application, a control method of a vehicle is provided, wherein the method can comprise the steps of responding to the vehicle to run in a running direction in a first running mode, obtaining image data of at least one environmental object in the running direction and point cloud data of the environmental object in the environment where the vehicle is located, determining first confidence coefficient of the image data, second confidence coefficient of the point cloud data and third confidence coefficient, wherein the first confidence coefficient is used for representing the completeness degree of the environmental object identified in the image data, the second confidence coefficient is used for representing the credibility degree of the quantity of the environmental object identified in the point cloud data, the third confidence coefficient is used for representing the similarity degree between attribute information of the same environmental object identified in the image data and the point cloud data respectively, determining a second running mode to be entered by the vehicle based on the first confidence coefficient, the second confidence coefficient and the third confidence coefficient, wherein the probability of the vehicle colliding with the environmental object in the second running mode is smaller than the probability of the vehicle colliding with the environmental object in the first running mode, and controlling the vehicle to be switched from the first running mode to the second running mode, and the risk coefficient is used for controlling the vehicle to run in the running mode based on the second risk coefficient. Further, the method further comprises the steps of obtaining the mode type of the second operation mode, wherein different mode types correspond to different safety degrees of the vehicle, and determining risk coefficients according to a determination strategy corresponding to the mode type. The method comprises the steps of determining a risk coefficient according to a determination strategy corresponding to a mode type, wherein the mode type comprises a first mode type and a second mode type, the security degree corresponding to the second mode type is smaller than that corresponding to the first mode type, the method comprises the steps of responding to the mode type being the first mode type or the second mode type, identifying a first candidate environment object set from image data, clustering point cloud data to obtain a second candidate environment object set, determining at least one environment object based on the first candidate environment object set and the second candidate environment object set, wherein the environment object is the first candidate environment object set or the second candidate environment object set and affects candidate environment objects of a vehicle driving according to a driving direction, inputting the at least one environment object into a risk quantization model of the vehicle, and performing risk prediction on the at least one environment object by using the risk quantization model to obtain the risk coefficient under the first mode type or the second mode type, and training the risk quantization model by using a neural network model. Further, the risk predicting is carried out on at least one environmental object by using a risk quantification model to obtain a risk coefficient, the risk predicting comprises the steps of determining collision time of a vehicle and the at least one environmental object by using the risk quantification model, determining target type weights of the at least one environmental object and a traffic channel relation corresponding to the at least one environmental object by using the risk quantification model, wherein the target type weights are used for representing the vehicle in a second running mode and the risk degre