CN103231710B - Driver workload based system and method for scheduling driver interface tasks - Google Patents

Abstract

The invention discloses a driver workload based system and method for scheduling driver interface tasks. One or more measurements of the driver workload can be determined through checking of power control states of vehicles, input from drivers to vehicles and the like, and then execution of driver interface tasks can be delayed and/or stopped based on the driver workload, and therefore, the driver workload is not increased. For example, optionally, execution of driver interface tasks can be scheduled based on the driver workload, so that driver interface tasks are executed according to scheduling, and affects of execution of driver interface tasks on workloads of drivers are minimized.

Description

Based on the system and method that driver's work load dispatches driver interface task

The application is the applying date to be on July 29th, 2010, and Application No. 201080068204.9 is entitled " to be used for The divisional application of the application for a patent for invention of the system and method based on driver's work load scheduling driver interface task ".

Technical field

Embodiments of the invention relate in general to a kind of for dispatching driver interface task based on driver's work load System and method.

Background technology

Particular vehicle can provide information entertainment information, navigation information etc. to strengthen driving experience.With driver and these Interaction increase between vehicle, promotes such interaction to be beneficial in the case where driver's work load is not increased.

The content of the invention

The measurement of driver's work load can be determined from vehicle, driver and/or environmental information.Can based on a determination that driving Member's work load optionally postpones or prevents the execution of specific driver interface task.Selectively, can based on a determination that drive The person's of sailing work load dispatches the execution of driver interface task, and driver interface task is performed according to the scheduling.

Description of the drawings

Fig. 1 is the example block diagram that hybrid working bears estimating system.

Fig. 2 is the exemplary graph of car speed, traction and braking curve.

Fig. 3 A to Fig. 3 C are the exemplary graphs of the state of motion of vehicle represented with yaw rate and yaw angle.

Fig. 4 A to Fig. 4 C are the exemplary graphs of driftage, longitudinal direction and manipulation limit surplus of breakking away.

Fig. 5 is the exemplary graph of car speed, traction and braking curve.

Fig. 6 A to Fig. 6 C are the exemplary graphs of the state of motion of vehicle represented with yaw rate and yaw angle.

Fig. 7 A to Fig. 7 C are the exemplary graphs of driftage, longitudinal direction and manipulation limit surplus of breakking away.

Fig. 8 and Fig. 9 are the exemplary graphs for finally manipulating limit surplus and risk.

Figure 10 and Figure 11 are respectively the exemplary graphs for high request environment and the low accelerator pedal position for requiring environment.

Figure 12 and Figure 13 are respectively the histograms of the standard deviation of the accelerator pedal position of Figure 10 and Figure 11.

Figure 14 is the curve map of the curve being consistent with the histogram of Figure 12 and Figure 13.

Figure 15 A to Figure 15 D are respectively accelerator pedal position, steering wheel angle, driver's control action（DCA）Index and car The exemplary graph of speed.

Figure 16 A to Figure 16 C are respectively running direction indicator activation, airconditioning control activation, instrument board（IP）The example of index is bent Line chart.

Figure 17 is the schematic diagram that vehicle follows another vehicle.

Figure 18, Figure 19 and Figure 20 are respectively the exemplary graphs of car speed, closing speed and distance travelled.

Figure 21 and Figure 22 are respectively to be spaced and be spaced（HW）The exemplary graph of index.

Figure 23 A to Figure 23 E are respectively rule-based index, IP indexes, DCA indexes, synthetic work burden estimation（WLE） The exemplary graph of index and car speed.

Figure 24 is the exemplary graph for characterizing driver requested membership function based on WLE indexes.

Specific embodiment

I. introduction

Driver's work load/requirement may indicate that such vision, body and perceive require, such as information amusement, phone, Driver is placed on primary flight activation and is surmounted the primary flight activation by the secondary activation of perspective suggestion etc.（So that Driver also carries out the secondary activation of suggestion in addition to carrying out primary flight activation）.

Driver there may come a time when improperly to think they can primary flight discussed above activation and secondary activation it Between take sb's mind off sth.Therefore, if vehicle system of the driver requested operation for modulation communication and with driver will be estimated System interaction, then estimate that driver requested operation can have significantly value.However, the driving environment of complexity may need novelty Forecasting Methodology is estimating driver's work load.The development of the intelligence system of driver's work load identification can be carried out to be beneficial to Man-machine interface of the customization output to driver（HMI）.

In order to continuously estimate work load, it may be necessary to which design is pre- in the case of different driving environments and/or driver Survey the estimator of work load.Communication service in adaptive compartment can be based on the environment for predicting driving demand wherein, and The value of service is sent to driver.Additionally, characterizing the driver's work load in a period of time（For example, long-term characterization）Can be with It is beneficial.For such estimation of driver's work load can allow not only high workload burden the time period during suppress or Postpone the communication technology in compartment, and cause the communication technology in compartment to be adapted to long drives requirement.

Specific embodiment described herein can provide for work load estimation（WLE）Method and system.WLE can be from Driver's work load is performed for the observable vehicle of the real-time HMI task managements of self adaptation, driver and environmental data State estimation/classification.In some cases, WLE can be estimated using single real-time technique and/or using real-time mixed method Meter work load.For example, can be based on driver, vehicle and environmental interaction to rule-based（rule-based）Algorithm supplement it is right In the additional continuous prediction of driver's work load.WLE algorithms can be combined to calculate with special study and Computational intelligence technology And the WLE indexes that prediction collects（For example, the continuous signal of the work load load estimation for driver is represented）.In some feelings Under condition, the driving demand of driver can be inferred from observable information of vehicles, the information of vehicles includes speed, acceleration, system The change of dynamic, steering, interval, instrument board and/or console interaction etc..

As an example, WLE indexes can be used to arranging/avoid/limit/customized voice order and/or present to driver's Other tasks/informations are improving function.During needing vehicle performance to operate, in dangerous driving environment, by instrument Plate carry out the time period of high activity during etc., can limit/customize/prevent the customizing messages for driver.

Intelligent Hybrid algorithmic method is contemplated that long-term and short-term driver actions.WLE mixed methods can catch driver's thing Part, situation and behavior, to adjust the communication of vehicle and driver.These and other technology described here can aid forecasting driving The sensed condition state of member's increase/reduction simultaneously can use existing vehicle sensors.

WLE indexes may also allow for that the level of communication is presented to into driver based on driving demand/work load.Message prioritization Level（For example, it is low, high）The burden that prediction can be based on determines whether to post messages to driver during special time.May be used also Long drives based on driver require for specific HMI information to present to driver.Selectively, mixing WLE frameworks can combine GPS and digital map database are considering road scene situation and condition.With regard to the physiological status of driver（Including heart rate, Sight line and breathing）Information can as input extraly be attached to WLE frameworks, for anormal detection.In other examples, The WLE indexes of prediction may pass to driver to remind driver to avoid carrying out secondary task under high workload burden. Other schemes are also feasible.

Fig. 1 is the block diagram of the embodiment of the WLE systems 10 for vehicle 11.System 10 may include that rule-based work is born Load index subsystem 12, vehicle, driver and/or environment tracking and evaluation work bear index subsystem 13, rely on environment Work load index collects subsystem 14 and entirety collects/WLE long-term characterizations subsystem 16.Subsystem 12,13,14,16（It is single Solely or in combination）One or more controller/processing meanss etc. can be implemented as.

Subsystem 12（Explain such as in following part VII）Can be by driver information and/or environmental information（For example, may be used From the controller local area network of vehicle（CAN）Obtain）As input information of vehicles, and export represent driver's work load based on The index of rule.Subsystem 13（Explain such as in following part III to VI）Can be by driver information and/or environmental information （For example, can obtain from the CAN of vehicle）As input information of vehicles, and export one or more for representing driver's work load Chain index（For example, the limit is manipulated（HL）Index, driver's control action（DCA）Index, instrument board（IP）Index, interval （HW）Index）.Subsystem 14（Explain such as in following part VIII）The index produced by subsystem 13 can be used as input, And output tracking（T）Index.Subsystem 16（Explain such as in following part VIII）Can be by rule-based index and T indexes As input, and export WLE indexes（Explain such as in following part Ⅸ）And/or the long-term characterization of WLE indexes.

In other embodiments, system 10 can lack subsystem 12,14,16.That is, specific embodiment can be constructed Only to produce one or more work load indexes.As an example, system 10 may be structured to be based only upon particular vehicle information（ Hereinafter describe）Produce IP indexes.In the case of these for the single measurement for only existing driver's work load, it is not necessary to collect. Therefore, in this example, WLE indexes are IP indexes.In these and other embodiment, scheduler 18 may be structured to produce The long-term characterization of WLE indexes.Other arrangements are also feasible.

WLE indexes may be sent to that scheduler 18, and scheduler 18 can be implemented as one or more controllers/process dress Put/etc..Scheduler 18（Explain such as in following part Ⅹ）Can be used as wave filter-based on WLE indexes prevent/delay by The information for being transferred to driver reaches driver.For example, if WLE indexes are more than 0.8, can prevent all to be intended for driving The information of the person of sailing.If WLE indexes close 0.5, the information of types of entertainment can be only prevented, etc..Scheduler 18 can also be based on WLE Exponent pair is scheduled the transmission for being transferred to the information of driver.For example, can prolong during the time period of high workload burden The information of vehicle maintenance late, Text To Speech reading, Inbound Calls etc..Additionally, scheduler 18 can be characterized based on long-term WLE indexes making Obtain vehicle to be output as customized in driver, as discussed in more detail below.For example, including cruise control, adaptive cruise control The output of the particular vehicle system of system, music suggestion, configurable HMI etc. can be required based on long drives.

The work load state of driver can be inferred from observable information of vehicles, the information of vehicles includes speed, adds The change of speed, braking, steering, interval, instrument board interaction etc..Table 1 lists the example related to driver's work load load Feature/standard (metric).

Table 1

Exemplary characteristics/the standard related to driver's work load

Standard	It is intended to the behavior effect for quantifying
		Average speed	Big speed increases/reduces
Maximal rate	Big speed increases
		Average time interval（Interval time）	The interval of reduction
Minimum interval	The minimum interval of reduction
		Brake reaction time（BRT）	The BRT of reduction
Braking abrupt change	The frequency of increase
		Steering wheel Overturn ratio	The frequency of the increase of little reversion
With interacting for IP（For example, IP buttons are pressed）	The frequency of increase
		Traffic density	The density of increase
Steering position	New driving environment
		Average speed	Big speed increases/reduces
Maximal rate	Big speed increases

Table 2a and table 2b list the example information that can be obtained via CAN known in the art/access.Following information can by with Make the input of any algorithm described here.

Table 2a

The example information that can be obtained via CAN

Accelerator pedal position	Microphone input
		Steering wheel angle	Glass stand sensor
Seat sensor	Speed
		Turn signal	Yaw rate
Defrosting signal	Transverse acceleration
		Temperature control	Longitudinal acceleration
Headlight state	Wheel speed
		High beam state	Throttle position
Fog lamp state	Master cylinder pressure
		Radio tuner order	Driver's requested torque

Rain brush state	Total axle torque
		Gear	Total moment of torsion distribution
Rain sensor	Roll velocity
		Configurable HMI	Yaw angle
Touch HMI	Related angle of heel

Table 2b

The example system information that can be obtained via CAN

Pull-in control system
	Anti-lock braking system
Electronic stability is controlled
	Adaptive learning algorithms
Relaxed by the collision braked
	Blind spot is monitored
Automatic stopping is helped

II. the simple discussion of vehicle stabilization control

The manipulation of vehicle determines turn inside diameter and the ability for manipulating.Vehicle is needed by its adjacent road in four tire contact planes Face, so as to maximize its handling.Understand spinning more than the tire of its limit of adhesion, skid or trackslip.One or more wheels Tire exceedes the condition of its limit of adhesion and is referred to alternatively as limit manipulation condition, and limit of adhesion is referred to alternatively as manipulating the limit.Once tire Its manipulation limit is reached, common driver generally can no longer control vehicle.In the case of so-called understeer, vehicle does not have The front tyre for having the steering input for fully performing driver, vehicle exceedes its manipulation limit, and vehicle ignores turning for driver Continue straight trip to request.In the case of so-called ovdersteering, vehicle excessively performs the steering input of driver, vehicle Rear tyre exceedes its manipulation limit, and vehicle continues spinning.For safety purposes, most vehicles are manufactured to be grasped at it Understeer during the control limit.

In order to compensate wagon control, electronics in the case of the vehicle that can not control to meet or exceed the manipulation limit in driver Stability contorting（ESC）System is designed to redistribute tire force can effectively make the vehicle please with the steering of driver to produce Seek the torque for turning to being consistent.That is, control vehicle avoids the situation of understeer or ovdersteering.

Since coming out from nineteen ninety-five, ESC system has been implemented in various platforms.Gradually carry out in vehicle year 2010, To installation comprehensively is realized vehicle year 2012, FMVSS 126 requires that ESC system is installed on gross weight grade and exists On less than 10,000 pounds of any vehicle.ESC system can be implemented as anti-lock braking system（ABS）With full speed traction control system System（TCS）Extension.ESC system can provide driftage and lateral stability side centered on the intention of driver to dynamics of vehicle Help.It can also make brake pressure（Higher or lower than the pressure that driver applies）It is proportional to single wheel, so as to produce effectively Torque is transported with resisting the unexpected driftage of vehicle and laterally sliding.This causes to be led for any during braking, accelerating or sliding Drawing surface course changing control when the limit is manipulated strengthens.More specifically, current ESC system is by the intention path of driver and from car The actual vehicle response that set sensor is inferred is compared.If the response of vehicle is different from path is intended to（Understeer turns To excessive）, and if requiring that vehicle is maintained on intention path and minimizes the out of control of vehicle by ESC controllers, then ESC controls Device processed applies to brake and reduce engine torque to the wheel for selecting.

Detectable limit manipulation situation can be carried out using the data present in ESC system, so not needing new sensing Device.As an example, it is considered to be equipped with the vehicle of ESC system, ESC system uses Yaw rate sensor, steering wheel sensor, laterally Accelerometer, vehicle-wheel speed sensor, master cylinder brake-pressure sensor, longitudinal accelerometer etc..As defined in ISO-8855 As, vehicle movement variable is defined in a coordinate system, wherein, on car body fixed car frame have vertical axis upwards, The lateral shaft of driver side is pointed to along the axle of the longitudinal direction of car body and from passenger side.

In general, can be from single kinematic variables（Such as yaw rate, yaw angle or combinations thereof）And other Control command（The such as request of operator brake, engine torque, ABS and TCS）In judgement come calculate vehicle feedback rank control. Wagon control level commands are discussed below.

Known auto model obtains dynamics of vehicle, vehicle along the yaw rate ω of the vertical axis of car body_zAnd at it The defective steering stabilizer angle beta of back axle definition_r, and meet below equation：

(1)

Wherein, v_xIt is the travel speed of vehicle, M and I_zIt is the gross mass of vehicle and the yaw rotation inertia of vehicle, c_fAnd c_r It is the cornering stiffness of front tyre and rear tyre, b_fAnd b_rIt is from the center of gravity of vehicle to front axle and the distance of back axle, b=b_f+b_r, M_zThe effective torque of vehicle is applied to, δ is front vehicle wheel steering angle.

The steering wheel angle δ that measurement can be used and travel speed v estimated_xCome from formula as input（1）Calculating is used for Target yaw rate ω that the steering of reflection driver is intended to_ztWith target side slip angle β_rt.In such calculating, it will be assumed that Prevailing roadway condition（For example, high friction level and small cornering stiffness c_fAnd c_r）Lower driving vehicle.Can also carry out for steady Signal Regulation, filtering and the gamma correction of the turning of the state limit are determined to finely tune target yaw rate and target side slip angle.These meters The desired value of calculation represents intention path of the driver on prevailing roadway.

Yaw rate feedback controller is mainly from yaw error（Difference between the yaw rate and target yaw rate of measurement）Meter The yawer of calculation.If vehicle is turned left and ω_z≥ω_zt+ω_zdbos（Wherein, ω_zdbosIt is time dependent dead Area）, or vehicle turns right and ω_z≤ω_zt-ω_zdbos, then vehicle oversteering and the ovdersteering control in ESC is activated Function processed.For example,（It is applied to vehicle to reduce ovdersteering trend）Effective torque request can be calculated as below：

During left steering：M_z=min(0,-k_os(ω_z-ω_zt-ω_zdbos)) (2)

During right turn：M_z=max(0,-k_os(ω_z-ω_zt+ω_zdbos))

Wherein, k_osIt is the gain of velocity dependent, it can be defined as below：

Wherein, parameter k_o、k_dbl、k_dbu、v_xdbl、v_xdbuIt is adjustable.

If ω when the vehicle turns to the left_z≤ω_zt-ω_zdbos（Wherein, ω_zdbosIt is time dependent dead band）, or The ω when vehicle is turned right_z≥ω_zt+ω_zdbos, then the understeer control function in ESC be activated.Effective torque request can be such as Lower calculating：

During left steering：M_z=max(0,-k_us(ω_z-ω_zt+ωzdbus)) (4)

During right turn：M_z=min(0,-k_us(ω_z-ω_zt-ω_zdbus))

Wherein, k_usIt is adjustable parameter.

Sideslip angle controller is the supplementary feedback controller of above-mentioned ovdersteering driftage feedback controller.Sideslip angle controller Sideslip angle beta will be estimated_rWith target side slip angle β_rtIt is compared.If the difference between the two exceedes threshold value beta_rdb, then yaw angle feedback control It is activated.For example, effective torque request is calculated as below：

During left steering： (5)

During right turn：

Wherein, k_ssAnd k_sscmpIt is adjustable parameter,(β_rcmp) be yaw angle compensation time-derivative.

Can similarly produce based on other feedback control terms of variable, such as yaw acceleration and sideslip gradient.When When main vehicle movement variable is yaw rate or yaw angle, above-mentioned effective torque can be directly used for determining the wheel that must be controlled and incite somebody to action It is sent to the amount of the brake pressure of corresponding control wheel.If dynamics of vehicle depends on multiple kinematic variableses, will carry out Control judges and arranges order of priority.Subsequently the final effective torque for judging is used for determining finally to control wheel and corresponding system Dynamic pressure.For example, during the event of ovdersteering, outer front vehicle wheel is chosen as controlling wheel, and in the event phase of understeer Between, interior rear wheel is chosen as controlling wheel.In the case of big sideslip, outer front vehicle wheel is chosen as controlling wheel all the time.When simultaneously When generation is breakked away and ovdersteering is gone off course, brake pressure can be calculated by overall consideration yaw error and yaw angle control command Amount.

More than due to the steering operation of driver cause exceed manipulation the limit situation in addition to, vehicle can be at it Lengthwise movement direction reaches its limit manipulation condition.For example, being braked on accumulated snow and eisbahn may cause to pin Wheel, which increases the stop distance of vehicle.On similar road open up the engine may cause driving wheel skid without Vehicle is set to advance.For this purpose, the manipulation limit may be alternatively used for these non-steering driving situations.That is, tire longitudinally braking or Driving force reaches the situation of its peak value and also is included within the definition for manipulating the limit.

ABS functions monitor rotary motion of each wheel relative to Vehicle Speed, and this can be by longitudinal sliding motion rate λ_iTable Show, wherein, i=1,2,3,4 are respectively directed in the near front wheel, off-front wheel, left rear wheel and off hind wheel, λ_iIt is calculated as below：

Wherein, t_fAnd t_rIt is the half of the wheelspan of front axle and back axle, ω_iIt is i-th vehicle-wheel speed sensor output, κ_iIt is i-th wheel velocity scale factor, v_yIt is lateral velocity of the vehicle in its c.g. position, v_minIt is the admissible minimum of reaction The parameter preset of longitudinal velocity.Note formula（6）Only vehicle not under reversing mode when it is just effective.When the system that driver starts Move and produce at wheel too big slip（For example ,-λ_i≥λ_bp=20%）, then ABS module will be released in the brake pressure at wheel.Class As, during the big throttle of applying causes to produce big slip on i-th driven wheel, TCS modules subtract request Little engine torque and/or request brake pressure are applied on the relative wheel on identical axletree.As a result, can be by monitoring λ_iWith λ_bp And λ_tpHas be close to predict ABS or TCS activation more.

III. manipulation limit index

Although above-mentioned ESC（Including ABS and TCS）Security purpose is effectively realized, but further enhancing is also feasible 's.For example, control to may expect to increase ESC system for roll stability.However, the appropriate correction that ESC attempts carrying out can Offset by driver or surrounding environment.Accelerate vehicle tire force considerably beyond the driving power of road/tire, can cause Even if the vehicle can not avoid the event of understeer under the intervention of ESC.

In general, manipulating the accurate determination of limiting case would generally be related to the direct measurement of road and tire characteristics, or Person obtains the information for including in the case of direct measurement is infeasible from many correlated variables.At present, both approaches are all inadequate Maturation is with real-time implementation.

Due to the feedback characteristic of ESC system so that ESC system can be configured to monitor vehicle movement variable（Vehicle Operating parameter）（The variable of such as last point description）To determine potential limit manipulation situation.When kinematic variables and its reference value Difference specified quantitative（For example, more than specific dead band）When, ESC system can start to calculate the order of difference control for brake and determine control Wheel.Subsequently corresponding brake pressure is sent to control wheel to stablize vehicle.The starting point of ESC activation can be considered manipulation The beginning of the limit.

More specifically, we can be to relative manipulation limit surplus h_xIt is defined as below：

Wherein, x is the deviation of kinematic variables and its reference value,It is defined in the case of not starting ESC, ABS or TCS The dead band interval that x falls into.X can be any control variables defined in last point（Or any other suitable control becomes Amount）.

In formula（7）Defined in h_xAdvantage to be that driving situation is expressed as in which can be quantized different classes of.For example, h is worked as_x When≤10%, driving situation can be classified as red area situation, in the case of red area, driver pay particular attention to or Take some special actions（For example, vehicle deceleration is made）；When 10%<h_x<When 40%, driving situation can be classified as yellow area feelings Condition, in the case of yellow area, driver needs paying special attention to for certain grade；When 40%<h_xWhen≤100%, driving situation can It is classified as general case.In usual cases, driver only needs to keep his normal driving notice.Certainly, can also make Use other scopes.

More specifically, let us discusses h using the control variables in last point of calculating_xCalculating.Can be by setting Put x=ω_z-ω_ztAndCome from formula（7）Calculate in ovdersteering situation（Now, when the vehicle turns to the left ω_z>ω_zt, the ω when vehicle is turned right_z>ω_zt）Driftage manipulation limit surplus h of period vehicle_OS, wherein, ω_zdbosIt is in formula Son（2）Defined in ovdersteering yaw rate dead band（OSDB）.

Similarly, can be by arranging x=ω_z-ω_ztAndCome from formula（7）Calculate in understeer feelings Driftage manipulation limit surplus h of vehicle under condition_US, wherein, ω_zdbusIt is in formula（4）Defined in understeer yaw rate dead band （USDB）.Note, above-mentioned dead band is probably the function of speed, the amount of target yaw rate, the amount of yaw rate of measurement etc..Turn to not Sufficient situation（x<0）Under dead band and ovdersteering situation（x>0）Under dead band be different, and they are adjustable parameters.

Can be by arranging x=β_r-β_rtAndCome from formula（7）Calculate the sideslip manipulation limit surplus of vehicle h_SSRA。

The longitudinal direction manipulation limit of vehicle is related to the situation that the driving force or brake force of tire reaches the manipulation limit.Can be by setting Put x=λ_i、x=0 andCome from formula（7）Calculate the traction control manipulation limit surplus for i-th driven wheelCan also be by arranging x=λ_i、x=λ_bpAndCome from formula（7）Calculate more than the ABS manipulation limit for i-th wheel AmountFinal traction manipulation limit surplus and braking manipulation limit surplus can be defined as：

Note, when computationally stating manipulation limit surplus further screening conditions can be used.For example, following bar can be used The combination of some conditions in a condition or following condition in part is set to 0 by limit surplus is manipulated：Target yaw rate Amount exceedes specific threshold；The amount of the yaw rate of measurement is more than specific threshold；The steering input of driver exceedes specific threshold；Or Person, the turning acceleration of limiting case, such as vehicle is more than 0.7g, vehicle with more than threshold value more than the deceleration of 0.5g, vehicle （For example, 100mph）Speed traveling etc..

For known riving condition, their validity is calculated and verified in order to test above-mentioned manipulation limit surplus, assemble The vehicle for having the research ESC system developed by Ford Motor Company is used to carry out vehicle testing.

For the riving condition drawn by car speed, throttle and braking described in fig. 2, measurement and calculating Vehicle movement variable is as shown in Fig. 3 A to Fig. 3 C.Corresponding individually manipulation limit surplus h_US、h_OS、h_TCS、h_ABS、h_SSRAIn Fig. 4 A extremely Illustrate in Fig. 4 C.The test execution is the free form obstacle skiing in the case where all ESC calculate operation, on snow pad Traveling.Close brake pressure to apply, so as to vehicle manipulates condition be close to the real limit.

As another test, on the road surface with high friction level vehicle is driven.In Figure 5 describe speed, traction and Braking curve.State of motion of vehicle is illustrated in Fig. 6 A to Fig. 6 C.Corresponding individually manipulation limit surplus h_US、h_OS、h_TCS、h_ABS、 h_SSRAIllustrate in Fig. 7 A and Fig. 7 B.

The envelope variable of all of independent manipulation limit surplus is defined as

h_env=min(h_OS,h_US,h_TCS,h_ABS,h_SSRA} (9)

In view of due to signal noise envelope may be caused to manipulate the suddenly change of limit surplus, using low pass filter F Z () is making h_envIt is smooth, so as to obtain the final manipulation limit（HL）Index or surplus：

h=F(z)h_env(100,000

For the vehicle testing data illustrated in Fig. 2 and Fig. 3 A to Fig. 3 C, final manipulation limit surplus is retouched in fig. 8 Paint, and be directed to the vehicle testing data illustrated in Fig. 5 and Fig. 6 A to Fig. 6 C, final manipulation limit surplus is described in fig .9.

HL indexes can provide the continuous variable between 0 and 1, and indicate that driver has with the manipulation limit of vehicle and be close to more （Wherein, value 1 represents the manipulation limit of the driver in vehicle）.The HL indexes that model should be based on can be in for example low μ driving paths The driving demand information being even more important is provided during condition.

Assume, with close its manipulation limit of vehicle, to need more visions, body and perception attention to maintain vehicle control System, can infer driver's work load information from HL indexes.As the work load of driver increases, HL indexes increase.With The work load for driver is reduced, and HL indexes are reduced.

IV. driver's control action index

Driver's control action（DCA）Index can provide the continuous variable between 0 and 1, and indicate driver for for example Total rate of change of the control action accelerate, braked and turn to.Increasing rate of change from the operating level of driver can reflect increase Driving demand, vice versa.Therefore, DCA indexes can provide the different drivers from the wagon control action for carrying out different aspects Related rate of change（Driving demand）Measurement.

For example, it is contemplated that impact of the accelerator pedal rate of change to driving demand.With reference to Figure 10 and Figure 11, for example, respectively low Real-time accelerator pedal position is temporally drawn in the case of requirement situation and high request.Situation is required compared to low, in high request feelings The rate of change of the accelerator pedal under condition is clearly relatively larger.

The standard deviation of the accelerator pedal position of Figure 10 and Figure 11 is illustrated respectively in Figure 12 and Figure 13.

With reference to Figure 14, probability corresponding with the distribution of Figure 12 and Figure 13 is produced using the gamma function of canonical form：

Wherein, a is scale factor, and b is form factor.Dotted line represents low driving demand distribution of standard deviation, and solid line is represented High driving demand distribution of standard deviation.These probability distribution of accelerator pedal rate of change illustrate driving demand classification and current class The grade of the difference between opportunity.For example, 2% standard deviation can represent low driving demand with more maximum probability, and 10% standard Deviation can represent high driving demand etc. with more maximum probability.The technology can be applied similarly to brake pedal position, steering wheel angle And/or other driver's control action parameters.Therefore, DCA indexes can be based on for accelerator pedal, brake pedal, steering wheel etc. Driver actions rate of change it is driver requested to estimate.

The average of the standard deviation rate of change illustrated in Figure 14 can change for different drivers.DCA indexes are calculated can Consider these averages for changing and calculate relative change rate.The derivative of driver's input can also be combined into the expected action of acquisition. Can be from each factor of analysis（For example, accelerator pedal position/speed, brake pedal position/speed, steering wheel angle position/speed Deng）Covariance determinant come obtain variance calculate.

In a particular embodiment, by the covariance of the impact driving demand based on following formula recursive calculation each factor Determinant is calculating DCA indexes：

Wherein, x_kIt is（At moment k）The two-dimensional vector of each driver's control action and its derivative,It is average （It can be constantly updated during each drive cycle, and be reset after each drive cycle）, α is calibration forgetting factor, G_kIt is the covariance inverse matrix estimated, I is unit matrix, P_kIt is the covariance matrix estimated,It is from formula（12）Δ x_kTransposed matrix.

The recursive calculation determinant det of covariance matrix, is provided below：

Wherein, n is vector x_kSize.It utilizes these parameters to provide the driving of the average relative to specific driver The measurement of the estimation rate of change of member's acceleration, braking and/or steering behaviour.It also provides the one-dimensional measurement of population variance, wherein, can be with One-dimensional measurement described in track is catching the significant changes of the rate of change for collecting of driver's control action.

Final DCA indexes（index）Can be continuous signal between 0 and 1 by ratio, and can be given by：

DCA index=max（Accelerator pedal variance, brake pedal variance, steering variance） (17)

The use above technical Analysis such as accelerator pedal position drawn in Figure 15 A and such as the side drawn in Figure 15 B To disk angle.Figure 15 C are shown and exported based on the example for DCA indexes of the input of Figure 15 A and Figure 15 B.In this example, The determinant of covariance matrix（16）The measurement of the estimation rate of change of driver's acceleration and steering behaviour is provided.By using each The maximum of rate of change is uniformed and collected to produce the DCA indexes drawn in figure 15 c to each rate of change.Speed Draw in Figure 15 D, as reference.The rate of change of increase is captured as into the value closer to 1 in DCA indexes（Indicate higher Driving demand）, and the value being captured as the rate of change of reduction in DCA indexes between such as 0 and 0.2（Indicate that low driving will Ask）.

V. instrument partitioned index

Driver can provide driver activation's to instrument board and/or interacting for other interfaces related with touch/voice Indicate.The increase of such driver activation's rank can increase the perception requirement to driver.As shown in table 1, driver presses The increase of button pressing activation can increase driver's work load.Can be aggregated as composite index with the frequency of interaction of cabin control, Wherein, control including rain brush control, climate controlling, volume control, running direction indicator, console with interacting for cabin control Platform, vehicle window control, automatic seat control, voice command interface etc..Therefore, instrument board（IP）Index is provided and represents driver and instrument The continuous output of dash board, electronics and/or any other HMI interaction（Between zero and one）.

For example, when k pressing/trigger buttons/interface arrangement at any time, output is given by：

BP_i(k)=α·BP_i(k-1)+(1-α)·1 (18)

When without pressing/trigger button/interface arrangement, output is given by：

BP_i(k)=α·BP_i(k-1)+(1-α)·0 (19)

Wherein, BP_iIt is to press/trigger pursuit gain for the button/interface of each tracked button/interface, α is calibration Forgetting factor.

Then, IP indexes output can be given by：

IP index=max(BP₁,BP₂,BP₃,BP₄.......BP_n) (20)

Wherein, n is the quantity of tracked button/interface.It is also possible to use and described herein any collect technology to determine IP indexes.As an example, can use with referring to formula（28）With（29）Similar technology of the technology of description etc..

Example running direction indicator and air-conditioning activation input are drawn respectively in Figure 16 A and Figure 16 B.Thus obtained IP indexes According to formula（18）、（19）With（20）It is determined that and drawing in Figure 16 C.In this example, rise time and steady-state value are based on sharp Duration living.

VI. interval index

Interval index provides the continuous variable between 0 and 1, and indicates the vehicle that just driven and front（Or side）Vehicle （Or other objects）Degree of closeness.As shown in Table 1, can be from the minimum of the average time interval and/or reduction for reducing Infer increased work load load in interval.

Can obtain relying on the interval of present speed from following formula：

Wherein, r_pK () is the position of k front vehicles at any time, r_fK () is the position of following vehicle, v_f(k) be with With the speed of vehicle.Equispaced HW_mK () can be obtained by following formula：

HW_M(k)=HW_M(k-1)+α(HW_curr-HW_M(k-1)) (22)

Wherein, α is the time constant for Exponential Filtration, can be selected as needed.Then, HW indexes can be obtained from following formula ：

Wherein, γ is HW index sensitive gains, HW_MAXIt is calibration value.Can be according to needed for meeting maximal index 1 interval when Between select/adjust gain.

In other embodiments, such as driver's type selecting/regulation sensitive gain can be based on.If it is known that driver's class Type（" youth ", " old age ", " teenager ", " new hand ", " veteran " etc.）, then sensitive gain can correspondingly be adjusted.Can be based on Driver is identified as " youth ", " old age ", " teenager " etc. by the certificate carried by driver known in the art.Can be by car Detection certificate, and certificate be used for recognize driver type.Selectively, vehicle can be provided and allow driver to recognize them The select button of the type of oneself.However, can be classified to driver using any suitable/known technology.For " teenager " " new hand " driver can increase sensitive gain, and being directed to " veteran " driver etc. can reduce sensitive gain.In other embodiments In, it is greater value that can select sensitive gain for " teenager " and " new hand " driver, and be directed to " veteran " driver etc. can It is smaller value that sensitive gain is selected.Therefore, in the case of same intervals, HW indexes can be for " teenager " driver be for more Big value and be smaller value for " veteran " driver waits.

Selectively（Or extraly）, can be selected based on ambient conditions/adjust sensitive gain.By suitable/known skill Art（Such as by detecting wheel skid）It is determined that moist or icy roads situation sensitive gain can be caused to increase.Dry roads Situation can cause sensitive gain to reduce.Can be selected using any suitable ambient conditions including traffic density, geographical position etc. Select/change sensitive gain.

Can with formula（21）、（22）With（23）In calculate similarly calculate with infrastructure interval apart, Wherein, infrastructure includes crossroad, highway, high request highway geometry etc..In this case, HW indexes can be by under Formula is obtained：

HW index=max(HW₁,HW₂,......HW_n) (24)

Wherein, n is the quantity of the project of the separation of just tracked high driving demand.It is also possible to use for formula （24）Weighting function.

In other embodiments, the interval of the increase for returning from the volume of traffic of the increase of adjacent lane is used as HW indexes Deviation（bias）Input.（The traffic density of increase can increase driving demand, such as illustrate in table 1.）

In other embodiments, collision time can be tracked in the scheme less than 1000ms.Will collide potential In the case of, the output of HW indexes can be defaulted as maximum 1.

With reference to Figure 17, collision time t_cCan be calculated by following formula：

Or

Wherein, V_xIt is closing speed, A_xIt is relative acceleration, X is the distance between vehicle.Can be from any suitable/known Radar system, vision system, laser radar system, vehicle-to-vehicle communication system etc. obtain range and range rate information.

Consider the calculating of the HW indexes in example vehicle follows scene, Figure 18 to Figure 20 illustrates the main car during the scene Closing speed and distance travelled between speed, vehicle.Figure 21 and Figure 22 respectively illustrate interval（By formula（22）Calculate） With HW indexes（By formula（23）Calculate）.

VII. rule-based subsystem

Referring again to Fig. 1, the fact that rule-based subsystem 12 may include for determining event binary system output identification And knowledge base.Subsystem 12 can provide specific specialists engineering science and vehicle driver's environmental interaction rule using as system 10 its The supplement of its component.Knowledge may be expressed as one group it is regular.The specific activation of Vehicular system can be included.

Suggestion of the specified output services burden of each rule, and if with（Condition）And then（Action）Structure.Work as satisfaction Action part is performed during the condition part of rule.Each rule may specify the suggestion that output services are born（0 or 1）.Can with appoint What suitable/known mode monitors/obtains multiple vehicle parameters, the vehicle by subsystem 12 from the CAN of such as vehicle Parameter includes that longitudinal acceleration, transverse acceleration, deceleration, steering wheel angle, button are used（For example, 2a and table are shown in Table 2b）.The fact that related to these parameters and combinations thereof, can be used to arrange conditional plan.

The system convention realized by subsystem 12 can be according to following form：

If

Vehicle_parameter1>x_iAnd Vehicle_parameter2>y_i

Then (26)

Enable during event for the specific delays or restriction of cabin system or information entertainment from Expert Rules.Base In the output of rule can also be processed to advocate based on expert for condition driving demand and special characteristic using carrying Collect for correlation output.

Rule can be based on information, for example, list in the table 2a and 2b more than.For example, if steering wheel angle>105 degree, Then Event_Flag=1（Event flag=1）.Certainly, Else Rule can also be built.

VIII. collect

One or more in HW indexes, DCA indexes, IP indexes and HL indexes can be collected with using following by subsystem 14 Technology formed tracking（T）Index.However, in the embodiment for use/calculating/determining an index is only needed, can be not required to Collect.

In a particular embodiment, short-term collect can be used for dispatching/postpone/postpone the information of driver of being transferred to/ Task.In the case where the highest driving demand estimated is needed, T indexes can be given by：

T Index=max(DCA Index,IP Index,Hl Index,HW Index) (27)

In other embodiments, the average/maximum output combination to exponential quantity as described below adopts the remittance for relying on environment Always.For example, referring to Fig. 1, DCA indexes, IP indexes, HL indexes and HW indexes can be combined by subsystem 14 and be given by with being formed T indexes：

Wherein, w_iIt is the weight of the dependence environment according to the driving demand value being added in input.Launch formula（28）Obtain：

Wherein, WLE_DCA、WLE_IP、WLE_HL、WLE_HWIt is respectively the output of DCA indexes, IP indexes, HL indexes and HW indexes.Phase The weight answered is by w_DCA、w_IP、w_HL、w_HWBe given.

Table 3 and table 4 list the example rule for collecting.

Table 3

For the example rule for collecting based on environment

Table 4

For the more example rules for collecting based on environment

Subsystem 16 can use the technology above by reference to the description of subsystem 14 to collect rule-based index and T indexes For WLE indexes.As an example, WLE indexes can be given by：

WLE Index=max(T Index,Rule-Based Index) (30)

Exemplary rule-based index, IP indexes and DCA indexes are drawn respectively in Figure 23 A and Figure 23 C.For considering The situation of the highest driving demand situation estimated, summarizes these indexes and in figure using technology described here Draw in 23D.Speed is drawn in Figure 23 E, for referring to.

Ⅸ. long-term characterization

In other embodiments, can be by subsystem 16 and/or scheduler 18（According to configuration）With time representation WLE indexes To provide HMI suggestions.Long-term WLE is characterized can cause HMI to be customized in driver based on the driving demand with the time.For example, examine Consider r_kIt is reflection（K at any time）For the variable of the WLE exponential quantities of driver.Assume driving demand be classified as with a, B, c } 3 classes that represent, and with the fuzzy membership functions μ for such as defining in fig. 24_a、μ_b、μ_c.Then, driving behavior d_kCan be under Infer in the example calculations in face：

d_k=[μ_a(r_k),μ_b(r_k),μ_c(r_k)] (31)

For example, if r_kValue be 0.4, then d_kMay be expressed as [0.18,0.62,0]（According to Figure 24）.It is filtered（It is long Phase）Deformation driving behaviorCan be expressed from the next：

Wherein, α is calibration forgetting factor（So as to α specifies/determine assessment TERM DEFORMATION driving behaviorTime period）.Pin Long term probability (p to each classification_k)_iCan obtain from following formula：

According to formula（33）, for the filtered deformation driving behavior of each classificationDivided by for all categories Filtered deformation driving behavior sumFor example, if[0,0.16,0.38] is represented as, then (p_k)_aWill be equal to 0 divided by 0+0.016+0.38（(p_k)_aWill be equal to 0）, (p_k)_bWill be equal to 0.16 divided by 0+0.016+0.38（(p_k)_b Will be equal to 0.29）, (p_k)_cWill be equal to 0.38 divided by 0+0.016+0.38（(p_k)_cWill be equal to 0.71）.

Then, the final long-term WLE indexes of driving demand characterize i_kCan infer from following formula：

Using above example, (p_k)_iThe maximum of value is 0.71（(p_k)_c）.Therefore, can be from formula（34）Infer and drive row To be currently at " high request " classification.

Ⅹ. scheduler

Scheduler 18 can the WLE indexes of computation, the long-term characterization of WLE indexes or DCA indexes, IP indexes, HL indexes With any one in HW indexes（For only using/calculate/determine the embodiment of single index）Come to information entertainment systems and/ Or interacting between other conversational systems and driver is modulated.WLE indexes provide the work load load estimated, for setting The voice command and other tasks for being presented to driver is put/avoids/customize/limit/dispatch, to improve feature and safety Property.

Interacting with the example of driver may include：Text To Speech conversion is produced, incarnation communication is produced, is produced with regard to incoming call The prompting of phone, the perspective dynamical system order of generation, the perspective voice of generation advise, produce tactile via such as thoughts and feelings steering wheel Response produces other audio frequency, vision and/or tactile output etc..Each task in these example driver interface tasks can With relative priority.For example, generation can have high priority with regard to the prompting of Inbound Calls, and produce perspective language Sound suggestion can have low priority.

Any suitable/known technology can be used for for priority type distributing to given driver interface task.As showing Example, scheduler 18 is capable of achieving high/low priority protocol, wherein, all promptings distribution with regard to Inbound Calls being generated is high Priority, the suggestion distribution low priority that all vehicles for being sent to driver are initiated.However, other priority can be used Scheme.As an example, the numeral between 0 and 1.0 can represent the priority of task：Particular task can distribute 0.3 priority, and Other task dividables are with 0.8 priority etc..In other embodiments, can by produce the controller of task known in the art/ Processor/subsystem（It is not shown）To distribute the priority type related to driver interface task.

Therefore, specific embodiment can allow the driver interface task to be in by dispatching sequence based on work load and priority It is existing.For example, if WLE indexes（Or any one index for depending on the circumstances）With the value between 0.4 and 0.6, then scheduler 18 can only allow to perform high priority driver interface task.In the case where WLE indexes are less than 0.4 value, scheduler Relatively low priority tasks can be scheduling to more late execution by 18.For example, if WLE indexes have the value between 0.7 and 1.0, Scheduler 18 can prevent the execution of all of driver interface task.During these high workloads are born, reach in WLE indexes little In the case of 0.7 value, high-priority task can be scheduling to more late execution by scheduler 18, and in WLE indexes 0.4 is less than Value in the case of, relatively low priority tasks can be scheduling to more late execution by scheduler 18.

Similarly, if long drives behavior is characterized as " high request ", no matter its priority, specific/all task can It is postponed/delayed/scheduled, until long drives behavior is characterized as " middle requirement " or " low requirement ".Selectively, if driven for a long time Sail behavior and there is any probability for being in such as " high request " classification, then specific/all task can be postponed/delayed/scheduled, directly It is 0 to the probability in " high request ".Certainly, other schemes are also feasible.For example, priority type be not used for appoint In the embodiment of business classification, all tasks can be postponed/delayed/scheduled according to the work load inferred.

In the case of receiving Inbound Calls during high workload is born, Inbound Calls can be proceeded to voice postal by scheduler 18 Part system.Once WLE indexes reach appropriate value, scheduler 18 can produce the prompting for indicating to receive Inbound Calls.

Algorithm disclosed herein may pass to processing meanss, any/complete in such as system 12,13,14,16 and 18 Portion, the processing meanss may include any existing electronic control unit for taking various forms or special electronic control unit, The various ways is included but is not limited to：It is permanently stored in not writeable storage medium（Such as ROM device）In information and variable Be stored in writable storage media（Such as floppy disk, tape, CD, ram set and other magnetically and optically media）On information.It is described Algorithm is also capable of achieving as the executable object of software.Selectively, suitable nextport hardware component NextPort can be used（Such as special IC （ASIC）, field programmable gate array（FPGA）, state machine, controller）Or other nextport hardware component NextPorts or device or hardware, software Combination with fastener components is implementing whole or in part algorithm.

Although embodiments of the invention have been shown and described, it is not intended to make these enforcements exemplify and describe this Invention is possible to form.The word for using in the description is descriptive words rather than restricted word, it should be appreciated that Various changes can be carried out in the case of without departing from the spirit and scope of the present invention.

Claims (12)

1. a kind of vehicle, including：

At least one processor, is configured to：(I) monitors the state of activation of each interface of multiple driver-vehicle interfaces, (II) is directed to each interface of the plurality of driver-vehicle interface, and the previous value and state of activation based on parameter produces expression The parameter of the frequency of interaction between driver and driver-vehicle interface, (III) receives the multiple driver interfaces that will be performed Task, wherein, each the driver interface task in the plurality of driver interface task includes priority type, and (IV) is based on Maximum parameter and priority type in the parameter of the generation optionally postpones or prevents the plurality of driver interface At least some task in task is performed.

2. vehicle as claimed in claim 1, wherein, the plurality of driver interface task includes at least in following operation It is individual：Audio output is produced, visual output is produced and produces tactile output.

3. vehicle as claimed in claim 1, wherein, the plurality of driver interface task includes producing with regard to Inbound Calls Remind, wherein, the maximum parameter in the parameter based on the generation optionally postpones or prevents the plurality of driver from connecing The step of at least some task in mouth task is performed includes：Inbound Calls is proceeded to into voice-mail system.

4. vehicle as claimed in claim 1, wherein, the plurality of driver-vehicle interface includes rain brush control, weather control System, radio volume control, running direction indicator control, the control of console console, car door lock, automatic seat or voice command connect Mouthful.

5. a kind of vehicle, including：

At least one processor, is configured to：(I) determine each interface in driver and multiple driver-vehicle interfaces it Between frequency of interaction, (II) based on a determination that maximum frequency of interaction determine driver's work load, (III) is received and will be performed Multiple driver interface tasks, wherein, each task in the plurality of driver interface task includes priority type, (IV) If driver's work load falls into the scope of predetermined value, optionally postpone or prevent the plurality of driver interface to appoint At least some task in business is performed, wherein, the scope of predetermined value depends on the priority type.

6. vehicle as claimed in claim 5, wherein, the plurality of driver interface task includes producing with regard to Inbound Calls Remind, wherein, optionally postpone or prevent the plurality of driving if the scope that driver's work load falls into predetermined value The step of at least some task in the person's of sailing interface task is performed includes：Inbound Calls is proceeded to into voice-mail system.

7. vehicle as claimed in claim 5, wherein, the plurality of driver-vehicle interface includes rain brush control, weather control System, radio volume control, running direction indicator control, the control of console console, car door lock, automatic seat or voice command connect Mouthful.

8. vehicle as claimed in claim 5, wherein, the plurality of driver interface task includes at least in following operation It is individual：Audio output is produced, visual output is produced and produces tactile output.

9. a kind of method for managing driver interface task, including：

Monitor the state of activation of each interface in multiple driver-vehicle interfaces；

For each interface in the plurality of driver-vehicle interface, the previous value and state of activation based on parameter produces table Show the parameter of the frequency of interaction between driver and driver-vehicle interface；

Multiple driver interface tasks that reception will be performed, wherein, each task in the plurality of driver interface task Including priority type；

Maximum parameter in based on the parameter of the generation collects and priority type is come to described more for what is performed At least some task in individual driver interface task is scheduled；

It is performed the plurality of driver interface task.

10. method as claimed in claim 9, wherein, the plurality of driver interface task is included in following operation at least One：Audio output is produced, visual output is produced and produces tactile output.

11. methods as claimed in claim 9, wherein, the plurality of driver-vehicle interface includes rain brush control, weather Control, radio volume control, running direction indicator control, the control of console console, car door lock, automatic seat or voice command Interface.

12. methods as claimed in claim 9, wherein, the plurality of driver interface task includes producing with regard to Inbound Calls Prompting, wherein, maximum parameter in the parameter based on the generation or collect to the plurality of driver for performing The step of at least some task in interface task is scheduled includes：Inbound Calls is proceeded to into voice-mail system.