Probabilistic driving behavior modeling system for vehicles

Abstract

The invention relates to a weather-conditioned hybrid probabilistic driving behavior modeling system for a vehicle, comprising: one or more controllers for executing instructions, to determine a probabilistic longitudinal velocity of the vehicle relative to a current weather condition based on a vehicle following model, a semantic rule system, and a velocity and visibility model through a longitudinal driving model stored by the one or more controllers. The current weather condition represents an adverse weather condition that affects a driving condition of the vehicle. The one or more controllers determine one or more lane options of the vehicle with respect to a current weather condition based on a route plan of the vehicle and perceived data representing an environment surrounding the vehicle through a probabilistic lateral driving model of the one or more controllers, and select a final lane option from the one or more lane options.

Description

Technical Field

[0001] The present disclosure relates to a hybrid probabilistic driving behavior modeling system for a vehicle that takes adverse weather conditions into account. Background Art

[0002] Autonomous vehicles are becoming more common, but still share the roads with manually driven vehicles. Driver behavior models can be used to predict driving maneuvers, driver intentions, and vehicle states. One type of driver behavior model is a microscopic driver behavior model, which includes several sub-models that describe the driver's behavior. Two examples of sub-models are vehicle following models and lane change models. The vehicle following model simulates how the driver of a constrained vehicle reacts to changes in the relative position and speed of the leading vehicle in an uninterrupted flow of traffic. The lane change model refers to when a vehicle changes from the current lane of travel to an adjacent lane of travel. It should be recognized that the vehicle following model and the lane change model are considered separately without determining the impact of one model on the other. Considering the models separately may lead to unrealistic behavior because the speed profile and lane decision are selected separately from each other.

[0003] Adverse weather conditions, such as rain, fog, or snow, may reduce the visibility of objects in the vehicle's surroundings. Adverse weather conditions may also reduce road friction and the vehicle's maneuverability and stability. Therefore, a driver's behavior may be different when operating a vehicle in adverse weather conditions compared to when the driver is operating the vehicle in clear or non-adverse weather conditions. However, current methods tend to ignore the impact of adverse weather conditions on driver behavior and typically only consider non-adverse weather conditions.

[0004] Therefore, while driver behavior models for autonomous vehicles achieve their intended purpose, there is a need in the art for improved driver behavior models that account for adverse weather conditions. Summary of the invention

[0005] In several aspects, the present invention discloses a hybrid probabilistic driving behavior modeling system for a vehicle conditioned on weather. The hybrid probabilistic driving behavior modeling system includes: one or more controllers, executing instructions to determine a probabilistic longitudinal speed of the vehicle relative to current weather conditions based on a vehicle following model, a semantic rule system, and a speed and visibility model through a longitudinal driving model stored by the one or more controllers, wherein the current weather conditions represent adverse weather conditions that affect the driving conditions of the vehicle. The probabilistic lateral driving model of the one or more controllers determines one or more lane options for the vehicle relative to the current weather conditions based on the vehicle's route planning and perception data representing the vehicle's surroundings. The controller selects a final lane option from the one or more lane options, wherein the final lane option includes a maximum probability of being selected at the probabilistic longitudinal speed compared to the remaining lane options that are part of the one or more lane options. The controller determines a next state of the vehicle based on the final lane option and the probabilistic longitudinal speed.

[0006] In another aspect, the probabilistic longitudinal velocity is determined by the vehicle following model by determining the velocity of the vehicle by calculating a first maximum velocity and a second maximum velocity and selecting a minimum value between the first maximum velocity and the second maximum velocity.

[0007] In yet another aspect, the first maximum speed indicates when to cause the driven vehicle to reach a target speed, and the second maximum speed indicates when to cause the driven vehicle to maintain a recommended distance from a leading vehicle.

[0008] In one aspect, the probabilistic longitudinal speed is determined by determining a constraint speed by a semantic rule system that limits the vehicle speed determined by a vehicle following model based on a plurality of high-level rules.

[0009] In another aspect, a plurality of high-level rules limits a vehicle speed determined by a vehicle following model based on one or more driving conditions resulting from specific adverse weather conditions represented by current weather conditions.

[0010] In one aspect, the probabilistic longitudinal speed is determined by a speed and visibility model by determining a probabilistic longitudinal speed of the vehicle relative to current weather conditions based on a visibility range limit constraint speed.

[0011] In another aspect, the speed and visibility model represents a relationship between a restricted speed and a visibility distance of a vehicle driver at a particular location, wherein the restricted speed is negatively correlated to an inverse of the visibility distance.

[0012] In yet another aspect, the probabilistic lateral driving model classifies a current road segment as part of a road plan as including a required lane change or a discretionary lane change.

[0013] In one aspect, in response to determining that a lane change is required on the current road segment to reach a next road segment as part of the route plan, the current road segment is classified as including a necessary lane change.

[0014] In another aspect, the current road segment is classified as including a free lane change in response to determining that the vehicle is performing one or more of the following operations: changing lanes to increase the distance to a leading vehicle or a following vehicle and to achieve a higher speed.

[0015] In yet another aspect, one or more controllers store a plurality of joint probability distribution models representing the probability of a particular lane being selected under particular adverse weather conditions and a particular longitudinal speed of the vehicle.

[0016] In one aspect, a joint probability distribution model is determined by defining a set of random variables and a set of weather conditions, wherein a first random variable represents longitudinal behavior and a second random variable represents lateral behavior.

[0017] In another aspect, the adverse weather condition is one of the following weather conditions: rain, fog, snow, and low light.

[0018] In yet another aspect, a method for determining a next state of a vehicle by a hybrid probabilistic driving behavior modeling system for a vehicle is disclosed herein. The method includes: determining a probabilistic longitudinal speed of the vehicle relative to current weather conditions based on a vehicle following model, a semantic rule system, and a speed and visibility model by a longitudinal driving model stored by one or more controllers, wherein the current weather conditions represent adverse weather conditions that affect vehicle driving conditions. The method includes: determining one or more lane options of the vehicle relative to current weather conditions based on a route plan of the vehicle and perception data representing the vehicle's surroundings by a probabilistic lateral driving model stored by one or more controllers. The method includes: selecting a final lane option from one or more lane options by one or more controllers, wherein the final lane option includes a maximum probability of being selected at the probabilistic longitudinal speed compared to the remaining lane options that are part of the one or more lane options. The method includes: determining the next state of the vehicle based on the final lane option and the probabilistic longitudinal speed.

[0019] In another aspect, the method includes determining the probabilistic longitudinal speed by determining, by a vehicle following model, a speed of the vehicle by calculating a first maximum speed and a second maximum speed and selecting a minimum value between the first maximum speed and the second maximum speed.

[0020] In yet another aspect, the method includes determining, by a semantic rule system, a speed constraint that limits a vehicle speed determined by a vehicle following model based on a plurality of high-level rules.

[0021] In one aspect, the method includes determining, from a speed and visibility model, a probabilistic longitudinal speed of the vehicle relative to current weather conditions by constraining the speed based on a visibility distance limit.

[0022] In another aspect, the method includes classifying a current road segment as part of a road plan as including a necessary lane change or a discretionary lane change.

[0023] In yet another aspect, a hybrid probabilistic driving behavior modeling system for a vehicle includes: one or more controllers executing instructions to determine a probabilistic longitudinal speed of the vehicle relative to current weather conditions based on a vehicle following model, a semantic rule system, and a speed and visibility model by a longitudinal driving model stored by the one or more controllers, wherein the current weather conditions represent adverse weather conditions that affect vehicle driving conditions; and wherein the speed of the vehicle is determined by the vehicle following model by calculating a first maximum speed and a second maximum speed, selecting a minimum between the first maximum speed and the second maximum speed, determining a constraint speed by the semantic rule system to limit the speed of the vehicle determined by the vehicle following model based on a plurality of high-level rules, and determining the probabilistic longitudinal speed of the vehicle relative to the current weather conditions by the speed and visibility model of the one or more controllers by limiting the constraint speed based on a visible distance, thereby determining the probabilistic longitudinal speed. The probabilistic lateral driving model stored by the one or more controllers determines one or more lane options for the vehicle relative to the current weather conditions based on a route plan of the vehicle and perception data representing the vehicle's surroundings. The one or more controllers select a final lane option from the one or more lane options, wherein the final lane option includes a maximum probability of being selected at the probabilistic longitudinal speed compared to remaining lane options that are part of the one or more lane options. One or more controllers determine a next state of the vehicle based on the final lane options and the probabilistic longitudinal speeds.

[0024] In another aspect, the probabilistic lateral driving model classifies a current road segment as part of a road plan as including a necessary lane change or a free lane change.

[0025] Further areas of applicability will become apparent from the description provided herein.It should be understood that the description and specific examples are intended for purposes of illustration only and are not intended to limit the scope of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] The drawings described herein are for illustration purposes only and are not intended to limit the scope of the present disclosure in any way.

[0027] Figure 1 is a schematic diagram of a vehicle including a disclosed hybrid probabilistic driving behavior modeling system according to an exemplary embodiment, the hybrid probabilistic driving behavior modeling system including one or more controllers;

[0028] Figure 2 According to an exemplary embodiment, Figure 1 the software architecture of one or more controllers shown;

[0029] Figure 3 A method for determining a probabilistic longitudinal speed of a vehicle under current weather conditions using a longitudinal driving model block according to an exemplary embodiment is shown. Figure 2 a portion of one or more controllers as shown;

[0030] Figure 4 A method for determining one or more lane options at longitudinal speed using a lateral driving model block according to an exemplary embodiment is shown. Figure 2 a portion of one or more controllers as shown; and

[0031] Figure 5 is a diagram showing a method for Figure 1 A flow chart of a method for modeling the behavior of a hybrid probabilistic driving behavior modeling system is shown. DETAILED DESCRIPTION

[0032] The following description is merely exemplary in nature and is not intended to limit the present invention, application, or uses.

[0033] See also Figure 1 , an exemplary hybrid probabilistic driving behavior modeling system 10 for a vehicle 12 is shown. It should be appreciated that the vehicle 12 may be any type of vehicle, such as, but not limited to, a car, a truck, a sport utility vehicle, a van, or an RV. The vehicle 12 includes an automated driving system (ADS) or an advanced driver assistance system (ADAS) for assisting a driver with steering, braking, and / or acceleration. The hybrid probabilistic driving behavior modeling system 10 includes one or more controllers 20 in electronic communication with a plurality of perception sensors 22, which are configured to collect perception data representing the environment surrounding the vehicle 12. In Figure 1 In the non-limiting embodiment shown, the plurality of perception sensors 22 include one or more cameras 30, an inertial measurement unit (IMU) 32, a global positioning system (GPS) 34, radar 36, and lidar 38. However, it should be appreciated that additional sensors may be employed.

[0034] like Figure 1As shown, one or more controllers 20 of the hybrid probabilistic driving behavior modeling system 10 include a longitudinal driving model block 40, a lateral driving model block 42, a semantic driving model block 44, a model association module 46, a weather calibration module 48, and a driver behavior module 50. The one or more controllers 20 receive an input state X of the vehicle 12, wherein the input state X represents the current state of the vehicle 12 observed by the plurality of perception sensors 22 in conjunction with other sensors (not shown) as part of the vehicle 12 (such as, for example, wheel speed sensors). The input state X of the vehicle 12 includes the current weather condition W, the current motion state of the vehicle 12, and the road condition R.

[0035] It should be appreciated that the current weather condition W represents an adverse weather condition that affects the driving conditions of the vehicle 12. Some examples of adverse weather conditions include, but are not limited to, rain, fog, snow, and low light that occurs at dusk or at night. In a specific embodiment, rain and / or fog can be further classified into different intensities, such as, for example, very light rain, light rain, moderate rain, and heavy rain. Adverse weather conditions affect driving conditions by limiting the driver's visual distance or by affecting the road surface. For example, the road surface may be affected by rain, causing the road surface to be slippery. The road condition R represents the friction coefficient μ of the road surface. coefficient The friction coefficient of the road is μ coefficient Indicates when the road is covered with ice or snow, when the road is wet due to precipitation, or when it is dry. The friction coefficient of the road μ coefficient is associated with the uncertainty of the vehicle position due to adverse weather conditions and is used to determine the uncertainty of the vehicle position due to adverse weather conditions. Specifically, the deviation σ of the normal distribution of the data samples modeling the vehicle behavior is related to the friction coefficient μ coefficient Inversely proportional.

[0036] The current motion state of the vehicle 12 represents variables such as, but not limited to, the current position x, y, the current longitudinal velocity vx, the current lateral velocity vy, the current acceleration a, and the current heading angle θ of the vehicle 12. The one or more controllers 20 determine a next state X' of the vehicle 12 based on the input state X, wherein the next state X' of the vehicle 12 represents the next motion state of the vehicle 12 and includes variables such as, but not limited to, the next position x', y', the next longitudinal velocity vx', the next lateral velocity vy', the next acceleration a', and the next heading angle θ' of the vehicle 12.

[0037] Figure 2The software architecture of one or more controllers 20 is shown. The longitudinal driving model block 40 executes a longitudinal driving model that determines a probabilistic longitudinal speed of the vehicle 12 relative to the current weather condition W and is represented as (ls)|W, where the driver behavior module 50 receives the probabilistic longitudinal speed (ls)|W as input. The longitudinal driving model includes a vehicle following model 60, a semantic rule system 62, and a speed and visibility model 64. The lateral driving model block 42 determines one or more lane options (lc)|W of the vehicle 12 relative to the current weather condition W, where the driver behavior module 50 receives the lane change maneuver (lc)|W as input. The one or more lane options (lc)|W of the vehicle 12 include turning right, turning left, or going straight.

[0038] As described below, the model association module 46 stores a plurality of joint probability distribution models employed by the driver behavior module 50. The joint probability distribution model represents the probability of a particular lane option being selected under a particular adverse weather condition and a particular longitudinal speed of the vehicle 12, and is determined based on training data. The driver behavior module 50 receives the probabilistic longitudinal speed (ls)|W and one or more lane options (lc)|W as inputs, and selects a final lane option from the one or more lane options (lc)|W, wherein the final lane option includes the maximum probability of being selected at the longitudinal speed (lc)|W compared to the remaining lane options that are part of the one or more lane options (lc)|W. The driver behavior module 50 then outputs the final lane option lc and the probabilistic longitudinal speed ls, wherein the one or more controllers 20 convert the final lane option lc and the probabilistic longitudinal speed ls into the next state X' of the vehicle 12.

[0039] The longitudinal driving model block 40 of the one or more controllers 20 will now be described. The longitudinal driving model block 40 implements a longitudinal behavioral driving model that determines a probabilistic longitudinal speed (ls)|W of the vehicle 12 under current weather conditions W based on a vehicle following model 60, a semantic rule system 62, and a speed and visibility model 64. Figure 3 is a flow chart illustrating an exemplary method 300 for determining a probabilistic longitudinal speed (ls) | W of the vehicle 12 relative to current weather conditions W. Figure 2 and Figure 3 , the method 300 may start at step 302. In step 302, the longitudinal driving model block 40 receives an input state X of the vehicle 12, which represents the current weather conditions W, the current motion state of the vehicle 12, and the road conditions R. The longitudinal driving model block 40 also receives the input state X of the vehicle 12 from the plurality of perception sensors 22 (such as Figure 1 Then, the method 300 may proceed to the decision step 304.

[0040] In decision step 304, the longitudinal driving model block 40 determines whether a leading vehicle exists in front of the vehicle 12. In response to determining that there is no leading vehicle, the method 300 proceeds to step 306, otherwise the method 300 proceeds to step 308. In step 306, the longitudinal driving model block 40 establishes a virtual leading vehicle model that simulates the leading vehicle located in front of the vehicle 12. The virtual leading vehicle model includes a lane position standard deviation (SDLP) value that represents the impact of adverse weather conditions on the lane keeping ability of the simulated leading vehicle. The virtual leading vehicle model also represents the speed of the leading vehicle. The SDLP value is determined according to the following equation 1:

[0041] SPLD=0.306(Weather coeff )-0.168(speed_limit)*(Weather coeff ) Equation 1

[0042] Among them, speed_limit is the speed limit, Weather coeff Represents a weather coefficient, which represents the impact of adverse weather conditions on lane position deviation, where different adverse weather conditions (e.g., rain, snow, or fog) affect lane position in different ways. The longitudinal driving model block 40 determines the speed of the leading vehicle based on a speed selection model that takes adverse weather conditions into account. The speed selection model reduces the speed of the leading vehicle based on a specific type of adverse weather condition. For example, a speed selection model reduces the speed of the leading vehicle by 5% to 40% (snowy days), 3% to 13% (light rain), and 3% to 16% (heavy rain) based on data from the Federal Highway Administration (FHWA). However, it should be appreciated that other data sources may also be used. Then, the method 300 may proceed to step 308.

[0043] In step 308 , the longitudinal driving model block 40 predicts the speed V of the vehicle 12 under the current weather conditions W based on the vehicle following model 60 . nf (t+T), where the current weather condition W is an unfavorable weather condition. The following describes the predicted speed V of the vehicle 12 under the current weather condition W. nf (t+T). Then, the method 300 may proceed to step 310 .

[0044] In step 310, the semantic rule system 62 of the longitudinal driving model block 40 determines the constrained speed V based on a number of high-level rules described below. nf_constrained , which constrains the speed V of the vehicle 12 determined by the vehicle following model 60 nf(t+T). A plurality of high-level rules limit the speed V of the vehicle 12 determined by the vehicle following model 60 based on one or more driving conditions resulting from adverse weather conditions represented by the current weather condition W. nf (t+T). Then, the method 300 may proceed to step 312 .

[0045] In step 312, the speed and visibility model 64 of the longitudinal driving model block 40 constrains the speed V by limiting the visibility distance based on the nf_constrained The probability longitudinal speed (ls)|W of the vehicle 12 relative to the current weather conditions W is determined. Specifically, the speed and visibility model 64 represents the constraint speed V for a particular location. nf_constrained and the visual distance of the driver of the vehicle 12, where the constraint speed V nf_constrained The speed and visibility model 64 will be described in more detail below. Once the probabilistic longitudinal speed (ls)|W of the vehicle 12 relative to the current weather conditions W is determined, the method 300 may terminate.

[0046] The vehicle following model 60 will now be described. The vehicle following model 60 is determined based on training data that models adverse weather conditions. It should be appreciated that the training data may include real world data or simulated data. In an exemplary embodiment, the vehicle following model is the Gipps vehicle following model and the training data is the Naturalistic Driving Study (NDS) data set from the Wyoming Strategic Highway Research Program (SHRP2) II. However, it should be appreciated that other models and data sets may also be employed.

[0047] Measure the velocity v at time intervals (t+T) nf (t+T), where t represents the current time and T represents the apparent reaction time, which is constant among all vehicles in the environment (i.e., vehicle 12). The vehicle following model 60 calculates the first maximum speed and the second maximum speed and then select the first maximum speed and the second maximum speed The speed v of the vehicle 12 is determined by the minimum value between nf (t+T), which is expressed in Equation 2 as:

[0048]

[0049] The first maximum speed Indicates when the vehicle 12 reaches the target speed, and the second maximum speed Indicates when the driving vehicle 12 maintains the recommended distance from the leading vehicle. The first maximum speed is solved based on Equation 3 And solve the second maximum speed based on Equation 4

[0050]

[0051]

[0052] Where f is a subscript representing the following vehicle (ie, vehicle 12), l is a subscript representing the leading vehicle, and x n represents the position of the leading vehicle or the following vehicle at time t, v n represents the speed of the leading vehicle or the following vehicle at time t, represents the maximum expected acceleration of vehicle n, represents the maximum expected speed of vehicle n (where vehicle n represents the leading vehicle or the following vehicle), represents the maximum expected deceleration of vehicle n based on the maximum braking performed by the driver of the following vehicle, represents an estimate of the maximum deceleration expected of the leading vehicle n and an estimate of the maximum braking capacity of the leading vehicle, L represents the length of the leading vehicle, S represents the vehicle spacing when stopped or the minimum following distance when stopped, and s=(L l +S) represents the sum of the leading vehicle size and the minimum following distance when stopped. Equation 3 and Equation 4 (e.g., ) is stored by the weather calibration module 48 of one or more controllers 20. For example, the Gipps parameter values ​​of the vehicle spacing s are approximately 2.9 meters (non-bad weather conditions), 4.4 meters (fog), 3.1 meters (drizzle), and 3.5 meters (snow).

[0053] The semantic rule system 62 of the longitudinal driving model block 40 is now described. It should be appreciated that the semantic driving model block 44 provides a plurality of high-level rules for the semantic rule system 62 of the longitudinal driving model block 40. The plurality of high-level rules of the semantic rule system 62 constrain the speed v of the vehicle 12 determined by the vehicle following model 60 based on one or more driving conditions. nf (t+T), these driving conditions are caused by specific adverse weather conditions represented by the current weather conditions W. In one embodiment, the driving conditions include one or more of the following: road surface conditions, visibility distance, road geometry, and construction areas. Road surface conditions indicate when the friction coefficient changes due to precipitation accumulation on the road. It should be recognized that the vehicle 12 may be required to reduce speed in icy and snowy weather, so multiple high-level rules limit the speed v of the vehicle 12 accordingly. nf(t+T). For example, if the road surface is dry, the multiple high-level rules instruct the vehicle 12 to travel at the current speed with a probability value of 0.95. However, if the road surface is icy and the current speed of the vehicle 12 is less than the threshold, the vehicle 12 will reduce the speed by 80% with a probability value of 0.9. The road geometry indicates curves in the road and when the vehicle 12 approaches an on-ramp or an off-ramp. The multiple high-level rules instruct the vehicle 12 to reduce the vehicle speed by forty percent when the vehicle 12 enters a work zone and the visibility distance is normal, and to reduce the vehicle speed by seventy percent when visibility is low.

[0054] The speed and visibility model 64 of the longitudinal driving model block 40 of the one or more controllers 20 is now described. The speed and visibility model 64 constrains the speed v by limiting the visibility distance based on the nf_constrained to determine the probabilistic longitudinal speed (ls)|W of the vehicle 12 relative to the current weather conditions W. As described above, the speed and visibility model 64 represents the constraint speed v for a particular location. nf_constrained and the visual distance of the driver of the vehicle 12, where the constraint speed v nf_constrained is negatively correlated to the inverse of the visibility distance. In other words, the speed and visibility model 64 takes into account how the driver reacts and limits the speed of the vehicle 12 when the visibility distance decreases. The specific location refers to a specific portion of the road. For example, the specific location may refer to a section of the road between a first milepost and a second milepost. It should be appreciated that the specific location includes various site-specific variables that cannot be directly transferred to another location, however, the speed and visibility model 64 for a specific geographic area can be used to set speed constraints based on the visibility distance. An example of a speed and visibility model 64 representing a highway section measured between two mileposts is expressed in Equation 5 as:

[0055]

[0056] Wherein, S represents the average speed every 5 minutes, Vis represents the visible distance, DayNight represents a daytime or nighttime dummy variable (1 represents daytime and 0 represents nighttime), SB6 represents a variable (1 represents the first milepost on the road), and SB7 represents another variable (1 represents the second milepost on the road). If the probabilistic longitudinal speed (ls)|W of the vehicle 12 relative to the current weather condition W is less than the average speed S, the probabilistic longitudinal speed (ls)|W of the vehicle 12 and the average speed S are set to the probabilistic longitudinal speed (ls)|W (S=ls), otherwise the probabilistic longitudinal speed (ls)|W is limited to the average speed S.

[0057] The lateral driving model block 42 of the one or more controllers 20 will now be described. The lateral driving model block 42 is based on the path plan of the vehicle 12 and the information from the plurality of perception sensors 22 (e.g., Figure 1 The vehicle 12 may be configured to have a lane plan that includes a plurality of road segments, wherein one or more lane options (lc)|W for the vehicle 12 are determined for each road segment that is part of the route plan.

[0058] Figure 4 A method 400 for determining one or more lane options (lc) |W for a vehicle 12 is shown. Figure 2 and Figure 4 , the method 400 begins at decision step 402. In step 402, the lateral driving model block 42 of one or more controllers 20 classifies the current segment as part of the route plan as including a necessary lane change or a free lane change. In response to determining that a lane change is required in the current segment to reach the next segment as part of the route plan, the current segment is classified as including a necessary lane change. In response to determining that the vehicle 12 is performing one or more of the following operations, the current segment is classified as including a free lane change: changing lanes to increase the distance to the leading vehicle or the following vehicle and to achieve a higher speed. Necessary lane changes and free lane changes are described in more detail below. In response to classifying the current segment as including a free lane change, the method 400 can enter decision step 404.

[0059] In decision step 404, the lateral driving model block 42 determines whether the conditions for performing a free lane change are met. Specifically, the conditions for performing a free lane change include determining whether the absolute braking distance between the vehicle 12 and the following vehicle allows the vehicle 12 to stop completely, and whether there is a sufficient distance between the vehicle 12 and the following vehicle in the direction of travel. The sufficient distance is based on the speed of the vehicle 12 and the speed of the following vehicle. It should be appreciated that the absolute braking distance and the sufficient distance are calculated based on weather adjustment parameters stored by the weather calibration module 48 of one or more controllers 20. In response to determining that the conditions for performing a free lane change are not met, the method 400 can terminate and consider the next road segment. However, in response to determining that the conditions for performing a free lane change are met, the method 400 enters decision step 406.

[0060] In decision step 406, the lateral driving model block 42 of the one or more controllers 20 determines whether the values ​​corresponding to the maximum deceleration and the acceptable deceleration corresponding to the necessary lane change are satisfied. The maximum deceleration refers to the upper limit of the deceleration rate of the vehicle 12 and the following vehicles, while the acceptable deceleration refers to the lower limit of the deceleration rate of the vehicle 12 and the following vehicles. The values ​​of the maximum deceleration and the acceptable deceleration are adjusted to account for adverse weather conditions and are stored in the weather calibration module 48 (e.g., Figure 2 In response to determining that the maximum deceleration and the acceptable deceleration corresponding to the necessary lane change are not met, the method 400 terminates and the next road segment is considered. However, in response to determining that the maximum deceleration and the acceptable deceleration corresponding to the necessary lane change are not met, the method 400 proceeds to step 408.

[0061] In step 408, the lateral driving model block 42 of one or more controllers 20 is based on the route planning and the Figure 1 The perception data collected by the plurality of perception sensors 22 is shown to determine one or more potential lane options. The method 400 may then proceed to a decision step 410 .

[0062] In decision step 410, the lateral driving model block 42 of the one or more controllers 20 evaluates each potential lane option based on a first headway threshold measured between the vehicle 12 in the target lane and the following vehicle, a second headway threshold measured between the vehicle 12 in the target lane and the following vehicle, and a third headway threshold measured between the vehicle 12 in the current lane and the following vehicle. The first, second, and third headway thresholds are selected to ensure that the vehicle 12 has sufficient room to maneuver when switching lanes.

[0063] In response to determining that the specific lane option does not meet the first headway threshold, the second headway threshold, or the third headway threshold, the method 400 terminates and the specific lane option is no longer considered. However, in response to determining that the specific lane option meets each of the first headway threshold, the second headway threshold, or the third headway threshold, the specific lane option is added to the one or more lane options (lc)|W of the vehicle 12. Once all specific lane options are considered, the lateral driving model block 42 of the one or more controllers 20 transmits the one or more lane options (lc)|W of the vehicle 12 to the driver behavior module 50, and the method 400 terminates.

[0064] In a non-limiting embodiment, the lane change model is simulated by a traffic simulation program such as VISSIM. Necessary lane change is defined by three parameters (maximum deceleration, acceptable deceleration, and acceleration reduction rate). As mentioned above, maximum deceleration refers to the upper limit of the deceleration rate of the vehicle 12 and the trailing vehicle, and acceptable deceleration refers to the lower limit of the deceleration of the vehicle 12 and the trailing vehicle. As the distance from the emergency stop position increases, the lower acceleration rate reduces the maximum deceleration in a linear manner. Some examples of necessary lane changes include but are not limited to entrance ramps and exit ramps.

[0065] A free lane change is defined by three parameters (minimum headway, distance reduction factor, and maximum deceleration for cooperative braking). The minimum headway represents the minimum distance between the vehicle 12 and the following vehicle or the leading vehicle required to execute the lane change. During the lane change, the value of the minimum distance is reduced by the distance reduction factor. The maximum deceleration for cooperative braking refers to the cooperative deceleration of the following vehicle that allows the vehicle 12 to change lanes. Some examples of free lane changes include, but are not limited to, catching up or passing a slower vehicle in the current lane of travel, returning to the driver's preferred lane, and entering a new lane. It should be appreciated that the parameters for both the required lane change and the free lane change are adjusted to account for adverse weather conditions and are stored in the weather calibration module 48 (e.g., Figure 2 For example, the maximum deceleration corresponding to the vehicle 12 is -11.38 meters per second 2 (m / s2) (non-bad weather), -8.32 m / s2 (rainy day), -17.84 m / s2 (snowy day) and -17.43 m / s2 (foggy day).

[0066] The driver behavior module 50 of one or more controllers 20 will now be described. As described above, the model association module 46 stores a plurality of joint probability distribution models that represent the probability of a particular lane being selected under specific adverse weather conditions and a specific longitudinal speed of the vehicle 12. The joint probability distribution model is determined by defining a set of random variables {Lo, La} and a set of weather conditions as {W}, where a first random variable Lo represents longitudinal behavior and a second random variable La represents lateral behavior. The joint probability distribution on the set of random variables {Lo, La} conditioned on the set of weather conditions {W} is defined as P(Lo, La|W). During training, a longitudinal probability distribution set P(Lo|W) and a lateral probability distribution set P(La|Lo, W) are defined. The joint probability distribution P(S, Lo, La|W) also takes into account the semantic behavior S, where the joint probability distribution P(S, Lo, La|W) is represented in Equation 6 as:

[0067] P(S, Lo, La | W) = P(S|Lo, La W) * P(S,La|Lo, W) * P(Lo|W) Equation 6

[0068] The driver behavior module 50 selects a lateral behavior La that maximizes the longitudinal behavior random variable Lo and the lateral behavior random variable la based on the current weather condition W, which is expressed in Equation 7 as:

[0069]

[0070] Where p(La, Lo|W) represents the conditional probability of the lateral behavior random variable La given the longitudinal behavior random variable Lo and the current weather condition W, and p(Lo|W) represents the conditional probability of the longitudinal behavior random variable Lo given the current weather condition W. The driver behavior module 50 selects a final lane option lc from the one or more lane options (lc)|W received from the lateral behavior block 42, the final lane option lc producing the maximum probability compared to the remaining lane options as part of the one or more lane options (lc)|W, and is represented in Equation 8 as:

[0071] lc=argmax lc∈Lc logp (Lc| Ls, W)*p (Ls|W) Equation 8

[0072] The driver behavior module 50 receives the probabilistic longitudinal speed (ls)|W and one or more lane options (lc)|W as inputs, and selects a final lane option from the one or more lane options (lc)|W based on a joint probability distribution, wherein the final lane option includes a maximum probability compared to the remaining lane options that are part of the one or more lane options (lc)|W. The driver behavior module 50 then outputs the final lane option lc and the probabilistic longitudinal speed ls, wherein the one or more controllers 20 determine a next state X′ of the vehicle 12 based on the final lane option lc and the probabilistic longitudinal speed ls.

[0073] Figure 5 The following is a diagram showing the Figure 2 Flowchart of an exemplary method 500 for modeling the behavior of the hybrid probabilistic driving behavior modeling system 10 is shown. Figure 1 , Figure 2 and Figure 5 , the method 500 may begin at step 502. In step 502, one or more controllers 20 receive a road segment as part of a route plan. Then, the method 500 may proceed to decision step 504.

[0074] At decision step 504 , if the road segment represents an intersection, the method proceeds to decision step 506 . Otherwise, the method 500 proceeds to decision step 510 .

[0075] In decision step 506, the one or more controllers 20 continue to monitor the perception data collected by the plurality of perception sensors 22 to determine whether the vehicle 12 can maneuver at the intersection. For example, the one or more controllers 20 may wait for a traffic signal to indicate that the vehicle 12 can maneuver. Once the one or more controllers 20 determine that the vehicle 12 can turn, the method 500 proceeds to step 508.

[0076] At step 508 , the vehicle 12 performs a maneuver at the intersection, wherein the maneuver is selected from one of: a left turn, a right turn, or a straight ahead. The method 500 may then proceed to decision step 510 to evaluate the next road segment.

[0077] In decision step 510, the one or more controllers 20 determine whether the vehicle 12 is capable of driving. If the vehicle 12 is capable of driving, the method 500 proceeds to decision step 512. In decision step 512, the one or more controllers 20 determine whether the vehicle is currently driving. If the vehicle 12 is not currently driving, the one or more controllers 20 select a braking longitudinal model and the method terminates. The braking longitudinal model utilizes the same longitudinal driving model as the longitudinal driving model block 40 during normal operation of the vehicle 12 to determine the slowing target speed required when braking and decelerating the vehicle 12. However, if the vehicle is not driving, the method 500 returns to decision step 510. Referring to decision step 510, if the one or more controllers 20 determine that the vehicle 12 is capable of driving, the method 500 proceeds to decision step 512.

[0078] In decision step 512, the one or more controllers 20 determine the number of lanes in the road. In response to determining that there is only one lane in the road, the one or more controllers 20 select the lateral driving model block 42 to determine one or more lane options (lc)|W for the vehicle 12. The method 500 may then terminate. In response to determining that there is more than one lane in the road, the one or more controllers 20 select the longitudinal driving model block 40 to determine the probabilistic longitudinal speed (ls)|W. The method 500 may then terminate.

[0079] Referring generally to the accompanying drawings, the disclosed hybrid probabilistic driving behavior modeling system provides various technical effects and benefits. Specifically, the hybrid probabilistic driving behavior modeling system integrates a vehicle following model, a lane change model, a driver's semantic behavior, and a speed and visibility model to determine the next state of a vehicle under adverse weather conditions, which solves the situation where the models determine inconsistent behaviors when the models are considered independently of each other. The hybrid probabilistic driving behavior modeling system maximizes the number of lane options available to the vehicle at a specific section of route planning under current weather conditions and a specific longitudinal speed.

[0080] The controller may refer to an electronic circuit, a combinational logic circuit, a field programmable gate array (FPGA), a processor (shared processor, dedicated processor or processor group) that executes code, or a combination of some or all of them, or a part of them, such as in a system on a chip. In addition, the controller may be based on a microprocessor, such as a computer having at least one processor, a memory (RAM and / or ROM), and associated input and output buses. The processor may run under the control of an operating system residing in the memory. The operating system may manage computer resources so that a computer program code embodied as one or more computer software applications (such as applications residing in the memory) may have instructions executed by the processor. In an optional embodiment, the processor may directly execute the application, in which case the operating system may be omitted.

[0081] The description of the present invention is merely exemplary in nature, and changes that do not depart from the gist of the present invention are included within the scope of the present invention. These changes should not be regarded as departing from the spirit and scope of the present invention.

Claims

1. A hybrid probability driving behavior modeling system for a vehicle based on weather conditions, the hybrid probability driving behavior modeling system comprising: One or more controllers that execute instructions to: determining, using a longitudinal driving model stored by the one or more controllers, a probabilistic longitudinal speed of the vehicle relative to current weather conditions based on a vehicle following model, a semantic rule system, and a speed and visibility model, wherein the current weather conditions represent adverse weather conditions that affect driving conditions of the vehicle; determining, by a probabilistic lateral driving model of the one or more controllers, one or more lane options for the vehicle relative to the current weather conditions based on a path plan of the vehicle and sensory data representing an environment surrounding the vehicle; selecting a final lane option from the one or more lane options, wherein the final lane option comprises a maximum probability of being selected at the probabilistic longitudinal speed compared to remaining lane options that are part of the one or more lane options; as well as Based on the final lane option and the probabilistic longitudinal speed, a next state of the vehicle is determined.

2. The hybrid probabilistic driving behavior modeling system according to claim 1, wherein the probabilistic longitudinal speed is determined by: determining the speed of the vehicle by calculating a first maximum speed and a second maximum speed by the vehicle following model; and A minimum value between the first maximum speed and the second maximum speed is selected. 3 . The hybrid probabilistic driving behavior modeling system according to claim 2 , wherein the first maximum speed indicates when the driven vehicle reaches a target speed, and the second maximum speed indicates when the driven vehicle maintains a recommended distance from a leading vehicle.

4. The hybrid probabilistic driving behavior modeling system according to claim 2, wherein the probabilistic longitudinal speed is determined by: A speed constraint is determined by the semantic rule system, the speed constraint limiting a speed of the vehicle determined by the vehicle following model based on a plurality of high-level rules.

5. The hybrid probabilistic driving behavior modeling system of claim 4, wherein the plurality of high-level rules limit the speed of the vehicle determined by the vehicle following model based on one or more driving conditions resulting from specific adverse weather conditions represented by the current weather conditions.

6. The hybrid probabilistic driving behavior modeling system according to claim 4, wherein the probabilistic longitudinal speed is determined by: A probabilistic longitudinal speed of the vehicle relative to the current weather conditions is determined by the speed and visibility model by limiting the constrained speed based on visibility distance.

7. The hybrid probabilistic driving behavior modeling system according to claim 5, wherein the speed and visibility model represents the relationship between the constraint speed and the visibility distance of the vehicle driver at a specific location, wherein the constraint speed is negatively correlated with the inverse of the visibility distance. 8 . The hybrid probabilistic driving behavior modeling system of claim 1 , wherein the probabilistic lateral driving model classifies a current road segment as part of a road plan as including a necessary lane change or a free lane change. 9 . The hybrid probabilistic driving behavior modeling system of claim 8 , wherein in response to determining that a lane change is required in the current segment to reach a next segment as part of the route planning, the current segment is classified as including the necessary lane change.

10. The hybrid probabilistic driving behavior modeling system of claim 8, wherein the current road segment is classified as including a free lane change in response to determining that the vehicle is performing one or more of the following operations: changing lanes to increase the distance to a leading vehicle or a following vehicle, and achieving a higher speed.

Images