Method of determining driving parameters and vehicle control device

Abstract

A method for determining driving parameters includes: collecting driving data from a driver; extracting driving characteristics from the driving data; wherein the driving characteristics include a first operating frequency of a first component of the vehicle and a second operating frequency of a second component of the vehicle; determining a first probability that the driver belongs to a first driving style type and a second probability that the driver belongs to a second driving style type based on the driving characteristics; and determining driving parameters based on the first and second probabilities; wherein the driving parameters include at least longitudinal acceleration and longitudinal deceleration. This method can dynamically adjust autonomous driving parameters to match the driver's driving style, effectively improving the user experience.

Description

Technical Field

[0001] This invention relates to the field of autonomous driving technology, and more specifically, to a method for determining driving parameters and a vehicle control device. Background Technology

[0002] With the development of artificial intelligence technology, automakers are incorporating autonomous driving functions as an important module into their products. However, existing autonomous driving functions often rely on drivers to manually set the parameters, making it difficult to dynamically adjust these parameters based on the driver's driving style and the driving environment. Summary of the Invention

[0003] According to one aspect of the present invention, a method for determining driving parameters is provided, which can dynamically adjust the parameters of autonomous driving.

[0004] Therefore, the present invention provides the following technical solution:

[0005] A method for determining driving parameters includes: collecting driving data from a driver; extracting driving characteristics from the driver based on the driving data; wherein the driving characteristics include a first operating frequency of a first component of the vehicle and a second operating frequency of a second component of the vehicle; determining a first probability that the driver belongs to a first driving style type and a second probability that the driver belongs to a second driving style type based on the driving characteristics; and determining driving parameters based on the first probability and the second probability; wherein the driving parameters include at least longitudinal acceleration and longitudinal deceleration.

[0006] Optionally, the method also includes performing autonomous driving operations on the vehicle based on driving parameters.

[0007] Optionally, the driver's driving characteristics may also include a third frequency of steering wheel operation.

[0008] Optionally, the driver's driving characteristics may also include the expected time it will take for the vehicle to travel to a identifiable target ahead.

[0009] Optionally, determining the driver's style type based on driving characteristics includes calculating a first probability and a second probability using a logistic regression model, wherein the logistic regression model is trained using at least a subset of the driving characteristics.

[0010] Optionally, determining the driving parameters based on the comparison between the first probability and the second probability includes determining the driving style type corresponding to the larger probability between the first probability and the second probability; and determining that the longitudinal acceleration and longitudinal deceleration are linear functions of the exponent D corresponding to the determined driving style type.

[0011] Optionally, determining the driving parameters also includes determining the vehicle's constant speed as (k1*D+c1)*V during autonomous driving.max V max This represents the maximum speed limit for the current road segment, with k1 and c1 being adjustment coefficients.

[0012] Optionally, determining the driving parameters also includes determining the lane change duration during automatic lane change as (c2-c3*D) / V. t V t c1 represents the vehicle speed when changing lanes, and c2 and c3 are adjustment coefficients.

[0013] According to another aspect of the present invention, a vehicle control device is provided, comprising: a data acquisition module configured to acquire driving data of a driver; a feature extraction module configured to extract driving features of the driver based on the driving data; wherein the driving features include at least a first operating frequency of a first component of the vehicle and a second operating frequency of a second component of the vehicle; and a parameter setting module coupled to the feature extraction module, configured to determine a first probability that the driver belongs to a first driving style type and a second probability that the driver belongs to a second driving style type based on the driving features, and to determine driving parameters based on the first probability and the second probability; wherein the driving parameters include at least longitudinal acceleration and longitudinal deceleration.

[0014] Optionally, the first component is an accelerator pedal and the second component is a brake pedal.

[0015] Optionally, the feature extraction module is also configured to: collect the third operating frequency of the steering wheel; and determine the expected time for the vehicle to travel to a identifiable target ahead.

[0016] The method for determining driving parameters provided by this invention can dynamically adjust autonomous driving parameters based on real-time collected driver driving data using a logistic regression algorithm to match the driver's driving style and real-time needs. By introducing machine learning algorithms, this method can meet the need for autonomous driving technology to quickly assess the driver's driving style and adjust driving parameters in real time. It features simple implementation and convenient upgrades and maintenance, thereby effectively improving the user experience. The vehicle control device provided by this invention can not only automatically adjust driving parameters to match the driver's driving style, but also has good real-time performance and versatility, and is convenient to deploy and has low implementation costs. Attached Figure Description

[0017] Figure 1 This is a flowchart illustrating a method for determining driving parameters applied to a vehicle according to an embodiment of the present invention.

[0018] Figure 2 This is a schematic diagram illustrating the module structure of a vehicle control device according to another embodiment of the present invention. Detailed Implementation

[0019] Specific details are set forth in the following description in order to provide a thorough understanding of the invention. However, those skilled in the art will readily recognize that embodiments of the invention may be practiced even without these specific details. Specific numerical references may be used in this invention, such as "first element," "second device," etc. However, these specific numerical references should not be construed as necessarily adhering to their literal order, but rather as indicating that "first element" and "second element" are distinct.

[0020] The specific details presented in this invention are merely exemplary, and variations may be made while still falling within the spirit and scope of the invention. The term "coupled" is defined as either a direct connection to a component or an indirect connection to a component via another component. Furthermore, the terms "approximately" or "generally" used herein for any numerical value or range indicate appropriate tolerances without affecting the effectiveness of the invention.

[0021] Preferred embodiments of methods, systems, and apparatus suitable for implementing the present invention are described below with reference to the accompanying drawings. Although the embodiments are described with respect to a single combination of elements, it should be understood that the present invention includes all possible combinations of the disclosed elements. Therefore, if one embodiment includes elements A, B, and C, and a second embodiment includes elements B and D, the present invention should also be considered to include other remaining combinations of A, B, C, or D, even if not explicitly disclosed.

[0022] According to one embodiment of the present invention, a method for determining driving parameters applied to a vehicle is provided, such as... Figure 1 As shown, it includes steps S10, S12, S14, S16 and optional step S18.

[0023] Step S10: Collect the driver's driving data.

[0024] In this step, the driver's biometrics are first identified to determine their identity, and driving data for each driver is stored independently. This allows subsequent steps to extract and analyze the current driver's driving characteristics, thereby determining driving parameters that match that driver.

[0025] Driving data when the driver is in manual driving mode and traveling normally (e.g., at a speed of 30 km / h or higher) is highly valuable for analyzing driver characteristics. This data can be collected using onboard sensors. Onboard sensors include those that help the vehicle perceive its external and internal environment and objects. Common onboard sensors include, but are not limited to: cameras, millimeter-wave radar, ultrasonic radar, lidar, inertial measurement units, position sensors, pressure sensors, and cornering sensors. To achieve rapid response and dynamic adjustment of autonomous driving parameters, shorter sampling intervals can be used to improve sampling accuracy, such as real-time sampling at 50 ms intervals.

[0026] In some embodiments, the types of driving data to be collected can be selected according to the configuration requirements of the autonomous driving system. Driving data includes, but is not limited to, vehicle speed, accelerator pedal position data, brake pedal pressure data, steering wheel angle data, and distance to other vehicles, where distance to other vehicles includes the distance between the vehicle and an identifiable target ahead (such as a vehicle, person, or object ahead). In other embodiments, driving data can be categorized according to road conditions and weather conditions to extract driving characteristics of the driver in specific situations, and thus to determine driving parameters for those situations.

[0027] Step S12: Extract the driver's driving characteristics based on driving data.

[0028] Based on real-time collected driver data, the current driver's driving characteristics can be determined. These characteristics include the driver's different operating frequencies on different vehicle components. In some embodiments, the driver's rapid acceleration frequency and rapid deceleration frequency can be determined as driving characteristics based on the degree (position of the pedal or the pressure applied) and / or number of times the driver depresses the accelerator or brake pedal. In other embodiments, the operating frequency of the transmission (automatic shift frequency) and / or the driver's clutch operation frequency (first operating frequency), and the intervention frequency of the active safety system and / or electronic stability system (second operating frequency) can be determined as some of the driver's driving characteristics. It should be noted that although the operating frequencies of the automatic transmission, active safety system, and electronic stability system are the frequencies of vehicle components or external modules, their operation (or activation) is caused by the driver's driving behavior and therefore can also reflect the driver's driving characteristics.

[0029] In some embodiments of the invention, the driver's driving characteristics also include the frequency of the driver's steering wheel operations; for example, the frequency of the driver's sudden steering wheel turns can be determined based on the collected angle and speed at which the driver turns the steering wheel. In other embodiments, the driver's driving characteristics also include the expected time for the vehicle to reach a identifiable target ahead, which reflects the driver's field of vision, reaction speed, and sensitivity of active safety systems (if any), and this expected time can be largely determined by vehicle speed and distance.

[0030] The process of extracting the current driver's driving characteristics described above may include the statistics and calculations of various driving data. As an example, the following driving characteristics are extracted: frequency of rapid acceleration, frequency of rapid deceleration, frequency of sharp steering wheel turns, and expected time for the vehicle to reach a identifiable target ahead. The statistical and calculation process is as follows:

[0031] Acceleration frequency: The number of accelerations during manual driving is counted. The method is to increment the acceleration count by 1 when the driver depresses the accelerator pedal to more than 80% and the duration exceeds 200ms. The acceleration frequency f is calculated by dividing the number of accelerations by the corresponding cumulative manual driving time. a ;

[0032] Rapid deceleration frequency: The number of rapid decelerations during manual driving is counted. The count is incremented by 1 when the driver depresses the brake pedal with a pressure exceeding 40 Bar and a duration exceeding 200 ms. The rapid deceleration frequency f is calculated by dividing the number of rapid decelerations by the corresponding cumulative manual driving time. b ;

[0033] Sharp steering wheel frequency: The number of sharp steering wheel turns during manual driving is counted. The method is to increment the count by 1 when the steering wheel angle changes by more than 10 degrees and lasts for more than 2 frames while the vehicle is not turning or making a U-turn. The criteria for not turning or making a U-turn are a vehicle speed greater than 0 and a cumulative steering wheel angle change in the same direction greater than or equal to 180 degrees within 3 seconds. The sharp steering wheel frequency f can be calculated by dividing the number of sharp steering wheel turns by the corresponding cumulative manual driving time. s ;

[0034] Expected time for the vehicle to reach a identifiable target ahead: When the driver encounters a target ahead during manual driving, the distance d between the vehicle and the target, the vehicle speed v, and the expected time ttc = d / v are collected. During the manual driving period, the expected time ttc for each encounter with a target ahead is recorded, and the average is calculated.

[0035] Step S14: Determine the first probability that the driver belongs to the first driving style type and the second probability that the driver belongs to the second driving style type based on driving characteristics.

[0036] Before performing this step, you can first define the first driving style type and the second driving style type.

[0037] Here is a specific example: When a driver manually drives the vehicle at a certain speed for a cumulative period of time, for example, accumulating 2 hours of driving at a speed exceeding 30 km / h, the probability of the current driver belonging to different driving style types can be determined based on the driving characteristics calculated in step S12. Multiple thresholds for the frequency of rapid acceleration, rapid deceleration, and sharp steering wheel turns can be predetermined, thereby classifying driving style types into three categories: conservative, average, and aggressive. Machine learning methods can then be used to determine the probability of the driver belonging to each of these three different driving style types. For example, a logistic regression model can be used to calculate the probability of the driver belonging to driving style type i, as shown in the following formula:

[0038]

[0039] Where y i (i = 1, 2, 3) represent the probabilities of having a conservative, average, and aggressive driving style, respectively; w 1i w 2i w 3i w 4i (i = 1, 2, 3) are the weight coefficients of the model, w 0i (i = 1, 2, 3) represents the intercept term of the model. The logistic regression model can be pre-trained using at least a subset of the driving features determined in S12, and the coefficients w in the above formula... 1i w 2i w 3i w 4i w 0i (i = 1, 2, 3) can all be obtained through training using the extracted driving feature data. For example, during model training, the driving feature data used for training can be labeled into three types—conservative, normal, and aggressive—according to predetermined standards, and then imported into the model for training.

[0040] The logistic regression model constructs a functional relationship between the probability that a driver belongs to driving style type i and the driver's various driving characteristics. The maximum likelihood method is used to train the weights, which is not only suitable for adjusting the weight coefficients, but also for using different weight coefficients under different road conditions and weather conditions. This makes the autonomous driving parameters suitable for matching the current driver's driving style, specific needs and the current driving environment.

[0041] Step S16: Determine driving parameters based on the comparison between the first probability and the second probability.

[0042] In this step, the driver's driving style type can be determined by comparing the probabilities of belonging to different driving style types as determined in step S14, and then the corresponding driving parameters can be determined. In some embodiments of the present invention, the highest probability among the different probabilities calculated in step S14 can be used as the basis for determining the driver's driving style type. For example, if the calculated probability of the driver belonging to the first driving style type is 0.7 and the second probability of belonging to the second driving style type is 0.3, then the driver is determined to belong to the first driving style type. The determined driving style type can be displayed on the in-vehicle display screen for the driver's reference, or it can be uploaded to the cloud for backend data analysis.

[0043] In some embodiments of the present invention, driving parameters include at least longitudinal acceleration and longitudinal deceleration, and may further include vehicle cruise control and lane change time. Driving parameters can be used to plan the vehicle's autonomous driving and to assist the driver's driving operations. To quantify each driving parameter, an index D corresponding to different driving style types can be defined. For example, when a driver's driving style is aggressive, normal, or conservative, the corresponding driving style indices D are 3, 2, and 1, respectively. This driving style index D can serve as a specific parameter indicator for subsequent adjustment of autonomous driving parameters, including longitudinal acceleration, longitudinal deceleration, cruise control, and lane change time. Longitudinal acceleration is an important parameter used by the autonomous driving system to control vehicle acceleration; longitudinal deceleration is an important parameter used by the autonomous driving system to control vehicle deceleration in non-emergency braking situations; cruise control can be used to determine the vehicle's cruise speed during autonomous driving; and lane change time refers to the time consumed when the vehicle changes lanes during autonomous driving.

[0044] In some specific embodiments of the present invention, the longitudinal acceleration and longitudinal deceleration can be linear functions of the driving style index D. For example, the longitudinal acceleration can be set as (0.03 + 0.015 * D)g, and the longitudinal deceleration can be set as (-0.025 - 0.015 * D)g, where g is the acceleration due to gravity. The cruise control speed during autonomous driving can also be determined as (k1 * D + c1) * V based on the driving style index D. max V max This represents the maximum speed limit for the current road segment, which can be obtained using high-precision maps from in-vehicle or mobile applications. k1 and c1 are adjustment coefficients. Additionally, the speed V during lane changes is also considered. t The lane change duration can be determined. The autonomous driving system controls the vehicle's lane change actions based on the lane change duration, which is also related to driving style. As an example, the lane change duration can be determined as (c2-c3*D) / V tWhere c2 and c3 are adjustment coefficients.

[0045] The aforementioned method for determining driving parameters can be used to determine the driver's driving style and then adjust the various driving parameters in real time. When the vehicle is in autonomous driving mode, using these driving parameters can make the autonomous driving more in line with the driver's intentions and driving habits, thereby effectively improving the user experience.

[0046] Another embodiment of the present invention provides a vehicle control device 20, such as... Figure 2 As shown, it includes: a data acquisition module 202, a feature extraction module 204, a parameter setting module 206, and an optional model training unit 2060. The modules can be connected in a specific manner. Figure 2 The parameters are coupled to each other. In some embodiments, the parameter setting module 206 can provide feedback information to the data acquisition module 202 and the feature extraction module 204.

[0047] The data acquisition module 202 is configured to collect the driver's driving data and includes multiple onboard sensors to help the vehicle perceive its external and internal environment and objects and collect data. The data acquisition module 202 can use shorter sampling intervals to improve sampling accuracy.

[0048] The feature extraction module 204 is configured to extract the driver's driving characteristics based on driving data, including the driver's different operation frequencies on different parts of the vehicle. In some embodiments of the present invention, the feature extraction module 204 can calculate the driver's rapid acceleration frequency and rapid deceleration frequency based on the degree and number of times the driver depresses the accelerator pedal and brake pedal. In other embodiments, the feature extraction module 204 also calculates the driver's steering wheel operation frequency and the expected time for the vehicle to reach a identifiable target ahead; for example, the expected time for the vehicle to reach a identifiable target ahead can be determined based on the collected distance and speed.

[0049] The parameter setting module 206 is configured to determine different probabilities of a driver belonging to different driving style types based on the extracted driving features. In some embodiments of the present invention, the parameter setting module 206 uses a logistic regression model to calculate the probability of a driver belonging to different driving style types. Before calculating the above probabilities, the parameter setting module 206 uses at least a portion of the driving features extracted by the feature extraction module 204 to train (optimize) the logistic regression model, and can also use another portion of the driving features to test (verify) the logistic regression model. During model training, the feature extraction module 204 can import the combination of input driving features into the model for training according to certain predetermined labels and obtain the corresponding weight coefficients.

[0050] The parameter setting module 206 is also configured to compare the different probabilities of the determined driver belonging to different driving style types to determine the driving style type to which the driver belongs, and then determine (quantify) autonomous driving parameters including longitudinal acceleration, longitudinal deceleration, constant speed, lane change time, etc.

[0051] In some embodiments of the present invention, the vehicle control device may be implemented using a set of distributed computing devices connected by a communication network, or based on the "cloud". For example, a data acquisition module is located on the vehicle side, a parameter setting module is located on the cloud side, the collected driving data is uploaded from the vehicle side to the cloud side, and the determined driving parameters are transmitted from the cloud side to the vehicle side. See also... Figure 2 The model training unit 2060, as a subunit of the parameter setting module 206, is located in the cloud. The collected driving data is uploaded from the vehicle to the cloud, the model training unit 2060 trains the model, and the determined model weight parameters are transmitted back to the vehicle via the cloud. In this device, multiple distributed computing devices operate collaboratively to provide services by utilizing their shared resources. This cloud-based implementation offers one or more advantages, including: openness, flexibility and scalability, centralized management, reliability, and scalability.

[0052] According to some embodiments of the present invention, a non-transitory readable storage medium is also provided, on which a batch of machine-executable instructions are stored, which, when executed by a processor, can implement the various steps of the method described above for determining driving parameters applied to a vehicle.

[0053] Those skilled in the art will understand that the various illustrative logic blocks, modules, circuits, and algorithmic steps described in connection with the aspects disclosed herein can be implemented as electronic hardware, computer software, or a combination of both. To demonstrate the interchangeability between hardware and software, the various illustrative components, blocks, modules, circuits, and steps have been generally described above in terms of their functionality. Whether such functionality is implemented as hardware or software will depend on the specific application and the design constraints imposed on the overall system. Those skilled in the art can implement the described functionality in varying ways for a specific application; however, such implementation decisions should not be construed as causing a departure from the scope of the invention.

[0054] The above description is only for preferred embodiments of the present invention and is not intended to limit the scope of protection of the present invention. Those skilled in the art may make various modifications without departing from the spirit of the present invention and the appended claims.

Claims

1. A method for determining driving parameters, comprising: a) Collect driver's driving data; b) Extract the driver's driving characteristics based on the driving data; wherein, the driving characteristics include a first operating frequency of a first component of the vehicle and a second operating frequency of a second component of the vehicle; c) Based on the driving characteristics, determine a first probability that the driver belongs to a first driving style type and a second probability that the driver belongs to a second driving style type; and d) Determine the driving parameters based on the first probability and the second probability; wherein the driving parameters include at least longitudinal acceleration and longitudinal deceleration, and step d) includes: Determine the driving style type corresponding to the larger probability between the first probability and the second probability; The longitudinal acceleration and the longitudinal deceleration are determined to be linear functions of an exponent D corresponding to the determined driving style type.

2. The method as described in claim 1, characterized in that, The method further includes: The vehicle performs autonomous driving operations based on the driving parameters.

3. The method as described in claim 1, characterized in that, The first component is an accelerator pedal, and the second component is a brake pedal.

4. The method as described in claim 1, characterized in that, The driving features also include a third frequency of steering wheel operation.

5. The method as described in claim 1, characterized in that, The driving characteristics also include the expected time for the vehicle to travel to a identifiable target ahead.

6. The method as described in claim 1, characterized in that, Step c) includes: The first probability and the second probability are calculated using a logistic regression model, wherein the logistic regression model is trained using at least a portion of the driving features.

7. The method as described in claim 1, characterized in that, Step d) also includes: The constant speed of the vehicle during autonomous driving is determined to be (k1*D+c1)*V. max V max This represents the maximum speed limit for the current road segment, with k1 and c1 being adjustment coefficients.

8. The method as described in claim 7, characterized in that, Step d) also includes: The lane change time during automatic lane change of the vehicle is determined to be (c2-c3*D) / V t V t c1 represents the vehicle speed when changing lanes, and c2 and c3 are adjustment coefficients.

9. A vehicle control device, comprising: The data acquisition module is configured to collect the driver's driving data; The feature extraction module is configured to extract the driver's driving features based on the driving data; wherein the driving features include at least a first operating frequency of a first component of the vehicle and a second operating frequency of a second component of the vehicle; A parameter setting module, coupled to the feature extraction module, is configured to determine a first probability that the driver belongs to a first driving style type and a second probability that the driver belongs to a second driving style type based on the driving features, and to determine driving parameters based on the first probability and the second probability; wherein, the driving parameters include at least longitudinal acceleration and longitudinal deceleration; the parameter setting module is further configured to: Determine the driving style type corresponding to the larger probability between the first probability and the second probability; The longitudinal acceleration and the longitudinal deceleration are determined to be linear functions of an exponent D corresponding to the determined driving style type.

10. The apparatus as claimed in claim 9, characterized in that, The first component is an accelerator pedal, and the second component is a brake pedal.

11. The apparatus as claimed in claim 9, characterized in that, The feature extraction module is also configured to: Collect the third operating frequency of the steering wheel; Determine the expected time for the vehicle to travel to a identifiable target ahead.

12. The apparatus as claimed in claim 9, characterized in that, The parameter setting module is configured as follows: The logistic regression model is trained using at least some of the driving features described. The first probability and the second probability are calculated using the logistic regression model.

13. The apparatus according to any one of claims 9-12, characterized in that, The device is deployed based on a cloud computing system.

14. A non-transitory readable storage medium having stored thereon a batch of machine-executable instructions, wherein the machine-executable instructions, when executed by a processor, implement the method as described in any one of claims 1-8.

Images