Vehicle with identification system

Abstract

Vehicle that includes the following: an identification system located inside the vehicle and configured to capture information from a token in the vehicle's environment and to classify a driver of the vehicle based on that information; a sensor array located inside the vehicle and configured to measure several parameters representing the vehicle's current handling condition and the vehicle's limit handling condition; and at least one control unit in communication with the identification system and the sensor array, configured to determine a range between the current handling condition of the vehicle and the limit handling condition, to record a history of the range if the driver classification is of a special type, and to characterize the long-term behavior of the driver, combining several driver characterization possibilities and calculating fuzzy probabilities that the driver classified as being of the special type falls into one of several categories.

Description

BACKGROUND

[0001] Driver error is cited as the cause of 45% to 75% of road collisions and as a contributing factor in a majority of all collisions.

[0002] Lane Departure Warning (LDW) uses a vision sensor to detect a vehicle's position relative to its lane and warn the driver of an unintentional lane departure. Certain forward collision warning (FCW) systems use environmental sensors to detect potential safety hazards ahead of a vehicle and warn the driver in advance. However, these existing driver warnings operate during steady-state or near-steady-state driving conditions.

[0003] German patent DE 10 2009 039 774 A1 discloses a motor vehicle with a control unit for this purpose, in particular for switching on and off or adjusting driving functions or performance characteristics, wherein at least one driver-specific evaluation value describing the driver's driving skills is determined, wherein, depending on the evaluation value, performance characteristics or driving functions of a system of the motor vehicle available to the driver are activated or deactivated, or threshold values ​​for automatic driving intervention, in particular by a safety system, are adjusted. A point value can be used as the evaluation value, wherein points for specific driving situations are added to or subtracted from the evaluation value. The driver can be identified before the start of a journey, for example via a personalized key or a chip card. The evaluation value of the identified driver can be stored.

[0004] DE 10 2005 044 771 A1 discloses a motor vehicle with a recording unit for capturing vehicle driving data, an evaluation unit for processing the vehicle driving data and an output unit for displaying information to a driver based on the captured vehicle driving data, wherein a traffic hazard potential is continuously determined from current vehicle driving data including movement data of operating elements by the driver, about which the driver is informed.

[0005] DE 10 2009 049 592 B4 discloses a driver assistance system, in particular an accident avoidance system, wherein at least one limit value of at least one driving condition variable describing the current physical limit of the vehicle, as well as a warning value derived therefrom and lying within the physical limit, are determined from technical data of the motor vehicle and other data, including data from an electronic stability control system, wherein the current value of the at least one driving condition variable is determined and a warning is issued to a driver if the driving condition variable lies between the warning value and the limit value. Data from a driver assistance system or from acceleration sensors can be taken into account to determine the limit value or the current value of the driving condition variable or the current warning value.The driver-specific warning value can be determined by taking into account an input describing the driver's driving skills. The driver can be identified via a personalized key. SUMMARY

[0006] The invention specified in the independent claims makes it easier to always provide drivers with performance characteristics or driving functions or thresholds for automatic driving intervention that correspond to their driving skills.

[0007] Advantageous embodiments of the invention are specified in the dependent claims. BRIEF DESCRIPTION OF THE DRAWINGS Fig. Figure 1 is a block diagram of an embodiment of a vehicle control system. Fig. Figure 2 is a diagram of example profiles of vehicle speed, vehicle traction, and vehicle braking. Fig.3A to 3C are diagrams of example vehicle motion states of yaw rate and drift angle. Fig. 4A to 4C are diagrams of example yaw, longitudinal and drift handling limits. Fig. Figure 5 is a diagram of example profiles of vehicle speed, vehicle traction, and vehicle braking. Fig. Figures 6A to 6C are diagrams of example vehicle motion states of yaw rate and drift angle. Fig. Figures 7A to 7C are diagrams of example yaw, longitudinal and drift handling limits. Fig. Figure 8 is a diagram of example membership functions that characterize four driver categories based on a handling risk factor. Fig. 9A, Fig. 10A and Fig. 11A are diagrams of example end-handling limits and risk. Fig. 9B, Fig. 10B and Fig.11B are diagrams of example probabilities of a driving style. Fig. Figure 12 is a diagram of example determinants of smooth and abrupt driving behaviors. Fig. 13A and Fig. 13B are diagrams of mean example gap times for aggressive and cautious driving. Fig. 14A and Fig. Figure 14B shows diagrams of example standard deviations of accelerator pedal rate for aggressive and cautious driving, respectively. Fig. 15A and Fig. Figure 15B shows diagrams of example standard deviations of the brake pedal rate for aggressive and cautious driving. Fig. 16A and Fig. Section 16B contains diagrams of example driver indices for aggressive and cautious driving. Fig.Figure 17 is a diagram of a relative example distance, example distance error and example longitudinal acceleration between a leading and a following vehicle for aggressive driving. Fig. 18 is a diagram of selected example parameters that illustrate aggressive driving of Fig. 17 characterize. Fig. Figure 19 is a diagram of a relative example distance, example distance error and example longitudinal acceleration between a leading and a following vehicle for cautious driving. Fig. 20 is a diagram of selected example parameters that illustrate the careful driving of Fig. 19 characterize. Fig. Figures 21 to 23 are block diagrams of embodiments of driver guidance systems. DETAILED DESCRIPTION I. Introduction

[0008] One goal of existing electronic vehicle control systems is to facilitate the driving task by identifying the driver's intention and assisting the driver by steering the vehicle to achieve that intention safely, smoothly, and efficiently. The effectiveness of electronic control systems can be significantly enhanced when the driver and the electronic control system work together toward the same accident avoidance goal, maximizing the vehicle's accident avoidance capability with the driver in the feedback loop as a system. One method to achieve this is to provide the driver with clear and transparent guidance in a timely manner, enabling a responsible driver to react accordingly.Such advisory information can be acquired or calculated by sensors typically found in a vehicle, implementing a two-way control loop between the driver and the electronic control system. The electronic control system follows the driver's intentions, and the driver reacts to the advisory information from the electronic control system to modify their driving inputs (such as reducing throttle, loosening steering inputs, etc.). In this way, seamless coordination between the driver and the electronic control system is possible and likely minimizes the impact of potential safety risks due to driver error.

[0009] We consider, among other things, warnings that occur near the handling limit, a driving or maneuvering condition under which vehicle stability controls typically intervene. In addition to problems encountered near the handling limit, the driver guidance system method discussed here can also be used to improve fuel economy; that is, a system that can use guidance and / or training to help the driver learn fuel-saving driving habits. We also discuss the use of vehicle stability control data to provide real-time warnings when the vehicle approaches the handling limit. This can be part of a grouping of warning functions that can be defined as an intelligent personal advisor (IPB) system.In general, the intelligence calculated for the IPB system can be sent to warn or advise a driver through various devices, including a haptic pedal, a windscreen display, an audio warning device, a voice system, etc.

[0010] Fig. Figure 1 represents the interaction of an embodiment of an IPB system 10 with other components / subsystems 12 of a vehicle 14. The other components / subsystems 12 may include vehicle sensors 16, 18 (e.g., a yaw rate sensor, a steering angle sensor, a lateral acceleration sensor, a longitudinal acceleration sensor, a wheel speed sensor, a brake pressure sensor, etc.), actuators 20, and one or more control units 22. The one or more control units 22 may include a stability control system 24, a decision logic unit 26, and other control units / systems 28 (e.g., an anti-lock braking system, a traction control system, etc.).

[0011] For any control system, the plant model can play a role in designing an effective control strategy. Likewise, a driver model is important for generating effective and appropriate driver guidance signals. Therefore, driving style characterization may be necessary. We discuss methods for identifying a driver's characteristics based on their vehicle handling ability. Although driver modeling and driver behavior characterization have been studied, we propose a method in which driving behavior / style and / or driver experience level can be derived, for example, based on the frequency and duration of driving near the handling limit (as well as other techniques). Such driver characterization information can be used in a variety of applications, some of which are discussed below. II. Brief discussion of vehicle stability controls

[0012] A vehicle's handling determines its ability to corner and maneuver. To maximize its handling, the vehicle must maintain contact with the road surface using its four tire contact patches. A tire exceeding its limit of grip will either spin, skid, or slide. A condition under which one or more tires exceed their limits of grip can be called a limit of handling, and the limit of grip can be called the handling limit. Once a tire reaches its handling limit, the average driver typically loses control. In what is known as understeer, the vehicle under-responds to the driver's steering input, its front tires exceed their handling limit, and the vehicle continues straight ahead regardless of the driver's steering input.In a so-called oversteer situation, the vehicle excessively responds to the driver's steering inputs, its rear tires exceed their handling limits, and the vehicle continues to skid. For safety reasons, most vehicles are designed to understeer at their handling limits.

[0013] To compensate for vehicle steering issues if a driver is unable to control the vehicle at or beyond its handling limits, electronic stability control (ESC) systems are designed to redistribute tire forces to generate a torque that can effectively turn the vehicle in accordance with the driver's steering input. This is done to control the vehicle and avoid understeer and oversteer.

[0014] Since its inception in 1995, ESC systems have been implemented in various platforms. With gradual adoption throughout the 2010 model year and achieving full installation by the 2012 model year, Federal Motor Vehicle Safety Standard 126 requires ESC systems in any vehicle with a gross vehicle weight under 10,000 lb. ESC systems can be implemented as an extension of anti-lock braking systems (ABS) and traction control systems (TCS) for all speeds. They can provide yaw and lateral stability assistance for vehicle dynamics centered around the driver's intentions. It can also proportion the brake pressure (above or below the pressure applied by the driver) to an individual wheel(s) to generate an active torque to counteract unexpected yaw and lateral movements of the vehicle.This results in improved steering control at the handling limits for any traction surface during braking, acceleration, or coasting. Specifically, current ESC systems compare the driver's intended path with the actual vehicle response derived from onboard sensors. If the vehicle's response differs from the intended path (either understeer or oversteer), the ESC control unit applies braking to a selected wheel(s) and reduces engine torque, if necessary, to keep the vehicle on the intended path and minimize loss of control.

[0015] A limit handling condition can be detected using data already present in ESC systems, thus eliminating the need for new sensors. For example, consider a vehicle equipped with an ESC system utilizing a yaw rate sensor, a steering wheel sensor, a lateral accelerometer, wheel speed sensors, a master cylinder brake pressure sensor, a longitudinal accelerometer, and so on. The vehicle motion variables are defined in the coordinate systems specified in ISO 8855, where, from a frame fixed to the vehicle body, the vertical axis points upwards, the longitudinal axis runs along the longitudinal direction of the vehicle body, and a transverse axis points from the passenger side to the driver's side.

[0016] In general, vehicle-level feedback controls can be calculated from individual motion variables such as yaw rate, drift angle, or their combination, together with decisions between other control commands such as driver braking, engine torque demand, ABS, and TCS. Vehicle-level control commands are discussed below.

[0017] The well-known bicycle model captures the vehicle dynamics, its yaw rate ω z along the vertical axis of the vehicle body and its drift angle β r , which is defined at its rear axle, and obeys the following equations Izω˙z=−bfcf(βr+bωztvx−1−δ)+brcrβr+Mz M(v˙xβr+vxβ˙r+brω˙z+ωz+vx)=−cf(βr+bωzvx−1−δ)−crβr where v x the vehicle's speed is, M and I z the total mass and the moment of inertia of the vehicle are, c f and c rthe lateral stiffness of the front and rear tires are, b f and b r The distances from the vehicle's center of gravity to the front and rear axles are b = b f + b r , M z The active moment is applied to the vehicle, and δ is the front wheel steering angle.

[0018] A target yaw rate ω zt and a target drift angle β rt , which are used to reflect the driver's steering intention, can be derived from (1) using the measured steering wheel angle δ and the estimated vehicle speed v x The inputs are used for calculations. In such a calculation, we assume that the vehicle is on a road with normal surface conditions (e.g., high friction level with nominal lateral stiffness c). f and c rThe signal processing, filtering, and non-linear corrections for steady-state cornering can also be performed to fine-tune the target yaw rate and target drift angle. These calculated target values ​​characterize the driver's intended path on a normal road surface.

[0019] The yaw rate feedback control unit is essentially a feedback control unit calculated from the yaw error (the difference between the measured yaw rate and the target yaw rate). When the vehicle turns left and ω z ≥ ω zt + ω zdbos (where ω zdbos a time-varying deadband), or the vehicle turns right and ω z ≤ ω zt - ω zdbosThis causes the vehicle to oversteer and activates the oversteer control function in the ESC. The active torque request (applied to the vehicle to reduce the oversteer tendency) could be calculated, for example, as follows. during a left curve: Mz=min(0,kos(ωz−ωzt−ωzdbos)) during a right turn: Mz=max(0,−kos(ωz−ωzt+ωzdbos)) where k os a velocity-dependent gain, which could be defined as follows kos=k0+(vx−vxdbl)kdbu−kdblvxdbu−vxdbl where the parameters k0, k dbl , k dbu , V xdbl , vxdbu are tunable.

[0020] If ω z ≤ ω z - ω zdbus (where ω zdbus a time-varying deadband is), when the vehicle turns left, or ω z ≥ ω z + ω zdbusWhen the vehicle turns right, the understeer control function in the ESC is activated. The active torque request can be calculated as follows. during a left turn: Mz=max(0,−kus(ωz−ωzt+ωzdbus)) during a right turn: Mz=min(0,−kus(ωz−ωzt+ωzdbus)) where k us a tunable parameter.

[0021] The drift angle control unit is an additional feedback control unit to the aforementioned override yaw feedback control unit. It compares the drift angle estimate β. r with the target drift angle β rt If the difference has a threshold value β rdb If the drift angle feedback control is activated, the active torque demand is calculated, for example, as follows during a left turn. βr≥0:Mz=min(0,kss(βr−Brt−Brdb)−ksscmpβ˙rcmp) during a right turn βr<0:Mz=max(0,kss(β−Brt−Brdb)−ksscmpβ˙rcmp) where k ss and k sscmp tunable parameters are and β̇ rcmp a compensated time derivative of the drift angle.

[0022] Other feedback control terms based on variables such as yaw acceleration and drift gradient can also be generated. If the dominant vehicle motion variable is either the yaw rate or the drift angle, the aforementioned active torque can be used directly to determine the required steering wheel (or wheels) and the amount of brake pressure to send to the corresponding steering wheel (or wheels). If the vehicle dynamics are dominated by multiple motion variables, a control decision and prioritization are performed. The final decided active torque is then used to determine the final steering wheel (or wheels) and the corresponding brake pressure (or brake pressures).During an oversteer event, for example, the front outside wheel is selected as the steering wheel, whereas during an understeer event, the rear inside wheel is selected as the steering wheel. During a large drift, the front outside wheel is always selected as the steering wheel. If both drifting and oversteer occur simultaneously, the amount of brake pressure can be calculated by integrating both the yaw error and the drift angle control commands.

[0023] In addition to the above cases where the handling limit is exceeded due to the driver's steering maneuvers, a vehicle can reach its handling limit in its longitudinal direction of travel. For example, braking on a snowy and icy road can lead to wheel lock-up, which increases the vehicle's braking distance. Open throttle on a similar road can cause the drive wheels to spin without moving the vehicle forward. For this reason, the handling limit can also be used for these non-steering driving conditions. That is, the conditions under which longitudinal tire braking or drive forces reach their peak values ​​can also be included in a definition of the handling limit.

[0024] The ABS function monitors the rotational movement of the individual wheels in relation to the vehicle's speed, which is determined by the longitudinal slip ratios λ. ican be characterized, where i = 1, 2, 3, 4 for the left front wheel, the right front wheel, the left rear wheel and the right rear wheel, calculated as follows λ1=κ1ω1max((vx+ωztf)cos(δ)+(vγ+ωzbf)sin(δ),vmin)−1 λ2=κ2ω2max((vx+ωztf)cos(δ)+(vγ+ωzbf)sin(δ),vmin)−1 λ3=κ3ω3max(vx−ωztr,vmin)−1,λ4=κ4ω4max(vx+ωztr,vmin)−1 where t f and t r the half track widths for the front and rear axles are, ω i The speed sensor output of the i-th wheel is K i the speed scaling factor of the i-th wheel is, v y the lateral velocity of the vehicle at its center of gravity, and v min a predetermined parameter that reflects the permissible minimum longitudinal speed. Note that (6) is only valid if the vehicle is not in reverse mode. If the braking initiated by the driver causes excessive slip at one wheel (e.g., - λ)i ≥ λ bp If the wheel speed is less than 20%, the ABS module releases the brake pressure at that wheel. Similarly, during a large accelerator pedal application that causes significant wheel slip at the i-th driven wheel, the TCS module requests a reduction in engine torque and / or brake pressure applied to the opposite wheel on the same axle. Consequently, ABS or TCS activations can be detected by monitoring how close the wheel speed is to the i-th driven wheel. i s to λ bp and λ tp are, can be predicted. III. Handling Limit Advisor

[0025] Although the aforementioned ESC (including ABS and TCS) is effective in achieving its safety objective, further improvement is always possible. For example, enhancements to ESC systems may be desirable for roll stability control. However, the appropriate correction that the ESC attempts to make can be counteracted by the driver or environmental conditions. An accelerating vehicle whose tire forces far exceed the traction capacity of the road and tires might not be able to avoid an understeer accident even with ESC intervention.

[0026] We introduce an integration of the driver and the ESC system so that they can work together with the driver in the feedback loop to improve the system's control performance. In certain embodiments, the proposed Handling Limit Advisor (HLM) determines how close the current driving condition is to the handling limit.

[0027] In general, accurately determining handling limits would involve direct measurements of road and tire properties, or very intensive information from many related variables if direct measurements are unavailable. Currently, neither of these methods is mature enough for real-time implementation.

[0028] Due to their feedback feature, ESC systems can be configured to determine potential handling limits by monitoring a vehicle's motion variables, such as those described in the previous section. If these motion variables deviate from their reference values ​​by a certain amount (e.g., beyond certain deadbands), the ESC systems can begin calculating differential brake control command(s) and determining the steering wheel(s). The corresponding brake pressure(s) is then sent to the steering wheel(s) to stabilize the vehicle. The starting point of ESC activation can be considered the beginning of the handling limits.

[0029] In particular, we can define a relative handling limit h. x as defined below hx={x - xx if 0 ≤ x ≤ x x - x_x if x ≤ x < 00 otherwise where x is the deviation of a motion variable from its reference value and [x,x] defines the deadband interval into which x falls without initiating ESC, ABS, or TCS. x can be any of the control variables defined in the last section (or any other suitable control variable).

[0030] The benefit of h defined in (8) x The key is that the driving conditions can be quantitatively characterized into different categories. For example, if h x If the gradient is ≤ 10%, the driving condition can be classified as a red zone condition, requiring the driver to exercise special attention or take certain specific actions (e.g., slowing down the vehicle); if 10% < h x If the gradient is less than 40%, the driving condition can be classified as a yellow zone condition, requiring a certain level of special attention from the driver; if 40% < h xIf the level is ≤ 100%, the driving condition can be characterized as normal. Under normal conditions, the driver only needs to maintain their normal driving attention. Other parameters can, of course, also be used.

[0031] Various audible and / or visual warnings can be activated to warn a driver of the handling limit. For example, if h x ≤ 10%, a warning light / haptic device can be activated to inform the driver that they need to slow down. Alternatively, a voice-activated display system can instruct the driver to take a specific action. If 10% < h x < 40%, an audible tone or a display can inform the driver that he is approaching unstable driving conditions, etc.

[0032] In particular, we want to use the control variables calculated in the last section to calculate h xto discuss. The yaw handling limit of the vehicle during oversteer situations h os (where ω z > ω zt , when the vehicle turns left, and ω z > ω zt , when the vehicle turns right) can be calculated from (8) by x = ω z - ω zt and x = ω zdbos is set to -x, where ω zdbos The overload yaw rate deadband is as defined in (2).

[0033] Similarly, the greed handling limit can be h US the vehicle for understeer situations from (8) can be calculated by x = ω z -ω zt and x = ω zdbus is set to -x, where ω zdbusThe understeer yaw rate deadband is as defined in (4). It should be noted that the aforementioned deadbands could be functions of the vehicle speed, the amplitude of the target yaw rate, the amplitude of the measured yaw rate, etc. The deadbands for the understeer situation (x < 0) and the oversteer situation (x > 0) are different and are tunable parameters.

[0034] The drift handling limit h SSRA of the vehicle can be calculated from (8) by x = β r - β rt and x = β rdb = -x is set.

[0035] The vehicle's longitudinal handling limits encompass the conditions under which either the driving or braking force of the tires approaches the handling limit. The traction control handling limit range for the i-th driven wheel h TCS , can be calculated from (8) by x = λ i , x = 0 and x = λ tbis set. The ABS handling limit for the i-th wheel h ABS , can also be calculated from (8) by x = λ i , x = λ tb = and x = 0. The final traction and braking handling limits can be defined as hABS=mini∈{1,2,3,4}hABSi,hTCS=mini∈{1,2,3,4}hTCSi

[0036] It should be noted that further selection conditions may be used when calculating the aforementioned handling limits. For example, one or a combination of the following conditions could be used to set the handling limit to 0: the amplitude of the target yaw rate exceeds a certain threshold; the amplitude of the measured yaw rate is greater than a certain threshold; the steering input of a driver exceeds a certain threshold; or extreme conditions such as the vehicle's cornering acceleration exceeding 0.5 g, the vehicle's deceleration exceeding 0.7 g, the vehicle being driven at a speed exceeding a threshold (e.g., 100 mph), etc.

[0037] To test the aforementioned handling limit span calculations and to verify their effectiveness in relation to known driving conditions, a vehicle equipped with a research ESC system developed at the Ford Motor Company was used to carry out the vehicle test.

[0038] For the effects of vehicle speed, throttling and braking, which are in Fig. Figure 2 shows the profiled driving conditions, which are the measured and calculated vehicle motion variables in Fig. 3A to 3C are shown. The corresponding individual handling limits h US , h OS , h TCS , h ABS and h SSRA are in Fig. 4A to 4C are shown. This test was performed as a freeform slalom on a snow cushion with all ESC calculations running. Brake pressure application was deactivated to bring the vehicle closer to the true limit handling condition.

[0039] For a further test, the vehicle was driven on a road surface with a high level of friction. The vehicle speed, vehicle traction, and vehicle braking profiles for this test are in Fig. 5 is shown. The vehicle motion states are in Fig. 6A to 6C shown. The corresponding individual handling limits h US , h OS , h TCS , h ABS and h SSRA are in Fig. 7A and Fig. 7B shown.

[0040] An envelope curve variable of all individual handling limit spans is defined as henv=min{hOS,hUS,hTCS,hABS,hSSRA}

[0041] Given that sudden changes in the envelope handling limit span could be due to signal noise, a low-pass filter F(z) is used to h env to smooth out in order to obtain the final handling limit. h=F(z)henv

[0042] For the in Fig. 2 and Fig. The vehicle test data shown in sections 3A to 3C represents the final handling limit range in Fig. 9A is shown, while for the in Fig. 5 and Fig. The vehicle test data shown in sections 6A to 6C represent the final handling limit range. Fig. 10A is shown. IV. Handling limits - Driving style characterization

[0043] In this section, we use the final handling limit calculated in (11) to characterize driving conditions and driving style related to vehicle handling. We introduce the concept of a handling risk factor (HRF) as a measure of how a driving condition relates to the handling limit. The handling risk factor r is defined as the complement of the final handling limit h, i.e., r=lh

[0044] The handling risk factor is minimal (r = 0) when the final handling limit h is maximal (h = 1), and vice versa. The HRF can also be used to develop a probabilistic model that describes different categories of driving styles reflected by the current driving conditions in relation to the handling limit.

[0045] Generally, a cautious driver typically drives without frequent aggressiveness, meaning rapid changes in steering, speed, and acceleration. Therefore, it is reasonable to characterize a cautious driver as one who consistently avoids using extreme driving inputs and approaching the maximum handling risk. An average driver is likely to have a higher level of HRF than a cautious driver. A skilled driver might be more adept at controlling the vehicle, meaning they can maintain a relatively high level of HRF for extended periods without the vehicle exceeding its maximum handling limits. A reckless driver exhibits careless handling behavior that is unpredictable and could induce rapid changes.It is expected that the reckless driver will operate with a handling risk factor that may occasionally approach the maximum (r = 1) very briefly, consequently causing frequent activation of the associated safety systems (e.g. ABS, TCS, ESC).

[0046] It is important to note that the difference between an experienced driver and a reckless driver is that the former can maintain a driving condition at a relatively high HRF level for an extended period, while the latter can only maintain a similar level for a short time before the vehicle exceeds its maximum handling limits due to the driver's poor control skills. Since the handling risk factor ranges that define, for example, cautious, average, experienced, and reckless driving behavior (with respect to the limit handling conditions) may not be well-defined, we use fuzzy subsets to quantify the four categories of drivers. We further evaluate these categories probabilistically based on a specific driving style.The fuzzy subsets that correspond to the categories of cautious, average, experienced, and reckless drivers can be described by the following membership functions. μc(r),μe(r),μa(r),μr(r) which are defined over the HRF totality [0, 1]. Fig. Figure 8 shows the relationship between the degrees of membership for each of these categories and the HRF.

[0047] The affiliation functions in Fig. 8 can be assigned to any event that is defined by a specific HRF with a value r k using a four-dimensional vector Dk=[μc(rk)μe(rk)μa(rk)μr(rk)]T His degree of belonging to each of the four example categories is represented: cautious, average, experienced, and reckless. An HRF value r k = 0.4 (corresponding to the handling limit value h) k= 0.6) is, for example, translated into the degrees of belonging to the cautious, average, experienced and reckless categories. μc(0,4)=0.46, μe(0,4)=0.85 μa(0,4)=0.09, μr(0,4)=0.22

[0048] The membership degrees encode the possibilities that the event characterized by an HRF with a value r = 0.4 (or the handling limit h = 0.6) could be assigned to any of the four example subdivisions. The vector of membership values ​​d kThis performs the mapping between a single driving event and the possible driver characterization with respect to the HRF of that event. To characterize the long-term behavior of the driver, we need a probabilistic interpretation of the possibilities generated by multiple events. By adding the membership values ​​for each event, we essentially combine the total possibilities that a specific driver can be classified as cautious, average, experienced, and reckless; that is, the vector d*=∑k=1N[μc(rk)μe(rk)μr(rk)]T where N is the number of samples. The combined possibilities can be viewed as frequencies (sometimes called fuzzy frequencies) because they reveal how often and to what extent the HRFs for the multiple events can be cascaded into the four example categories. The alternative to combining the possibilities, i.e., adding the membership functions, is to add 1 when the specific membership class µ i (r k If i ∈ {c, a, e, r} is greater than a prescribed threshold, e.g., 0.8, or otherwise 0, this leads to the calculation of the conventional frequencies of the four example categories. From the combined possibilities, we can calculate the probabilities of the cautious, average, experienced, and reckless driving styles. pi=di*(∑j∈{c,a,e,r}dj*)−1 where i ∈ {c, a, e, r}. The probabilities are calculated from the combined possibilities (fuzzy frequencies) and can be considered fuzzy probabilities. The reason for the fuzzyness here is the lack of certainty in characterizing the relationship between the four example categories and the HRF. For the special case of coarsely defined categories (represented by intervals as fuzzy subsets), the probabilities transform into Boolean values, their combined values ​​become frequencies, and consequently, the fuzzy probabilities are converted into conventional probabilities.

[0049] The most probable driver category i* is the one characterized with the highest probability, i.e. i.=argi∈{c,a,e,r}max(pi)

[0050] The calculation of probabilities based on frequencies can be expressed in terms of mean frequencies. pi=di* / N(∑j∈{c,a,e,r}dj* / N)−1

[0051] Alternatively, it can be expressed by exponentially weighted mean frequencies, with the higher weights assigned to the probabilities associated with the most recent events. Numerically, the process of generating a weighted mean, with higher weights corresponding to the more recent observation, can be accomplished by applying a low-pass filter that implements the exponential smoothing algorithm in the time domain. dnew*=(1−α)dold*+αdk=dold*+α(dk−dold*) where the constant forgetting factor 0 < α ≤ 1 controls the rate of updating the mean d* by assigning a set of exponentially decreasing weights to the older observations. For a constant forgetting factor α, expression (17) recursively generates a vector with positive weights. W=⌊(1−α)kα(1−α)k−1(1−α)k−2...α⌋ with a unit sum. The vector W designs a weighted mean combination operator with exponentially decreasing weights parameterized by the forgetting factor α. The parameter α defines the memory depth (the length of the moving window) of the weighted mean combination operator. Therefore, the filtered value d* of the membership class vector in (17) represents the weighted means of the individual possibilities over the weights W. Since all combined possibilities over the same moving window of length K aSince the values ​​of 1 / α can be calculated, we can view them as representations of the frequencies of the assignments to each of the four concepts. The weighted mean (17) is calculated over events with indices that correspond to a soft interval. s∈{k−Kα+1,k] belong, where the symbol { indicates a soft lower bound, the values ​​with lower indices than (k - K a ) with a relatively small contribution. Consequently, the combined possibilities that form the vector d* can be converted into probabilities according to expression (14).

[0052] In certain embodiments, α can be selected such that a characterization is performed for a desired duration. For example, a user can provide an input regarding α such that a characterization is performed every half hour. Other scenarios are also possible.

[0053] For the Fig.The vehicle tests shown in sections 2, 3A to 3C and 4A to 4C are the individual p i 's in Fig. 9B showed that for most of the driving the driver exhibited reckless driving behavior, which was associated with the large value of the drift angle in Fig. 3C is consistent (the peak amplitude of the drift angle exceeds 10 degrees). For the by Fig. The vehicle inspection shown in sections 5 to 7C are the individual p i 's in Fig. 10B shows that the driver initially exhibited average driving behavior and then transitioned to reckless driving behavior.

[0054] The calculated probabilities define the most likely HRF-based characterization of a driver for the time window determined by the forgetting factor α. By modifying the moving window, we can learn and combine the long- and short-term characterizations for a specific driver based on the HRF.

[0055] To predict the impact of HRF changes on driver characterization, we introduce transition probabilities. The Markov model P probabilistically describes the set of transitions between the current and predicted driver category values: pj(k+1)→pi(k)p11p12p13p14p21p22p23p24p31p32p33p34p41p42p43p44 where p ij the probability of switching from category i at time k to category j at time k+1, and p ii = max(p i) is the probability assigned to the dominant category i at time k, i, j ∈ {c,a,e,r}. The transition probabilities p ij are derived from the combined transition probabilities, which are only updated when i = arg max(p l ) at time k and j = arg max(p l ), l ∈ {c,a,e,r}. dif,new*={(1−α)dif,old*+αdi,k if j=argl∈{c,a,e,r}max(pl)(1−α)dif,old* otherwise}

[0056] The transition probabilities are then calculated by converting the combined transition probabilities into individual probabilities. The maximum transition probability p ij determines the transition from category i to category j as the most probable transition.

[0057] Fig. 11A and Fig.11B uses a driving vehicle test to verify long-term driving behavior characterization. The driver generally exhibits a cautious driving style (who could be a novice, average, or experienced driver). At approximately 190 seconds, the vehicle was turned with some degree of aggression, as seen from the peak in the HRF graph, and the driving style transitioned into the average category. Since no further significant HRF events were identified, this category was used for the remainder of the driving cycle in conjunction with the long-term characterization concept.

[0058] As mentioned above, various audible and / or visual warnings can be activated to alert a driver to the handling limit. The range thresholds, which define whether a warning (and / or which type) should be issued, can be modified (or removed) based on driver characterization. For example, if the driver is determined to be an experienced driver, the warning threshold can be set to [value missing in original text]. x ≤ 10% per hour x The warnings regarding the handling limit range can be reduced by ≤ 2% or removed (an experienced driver may not need the warnings). IV. Unsupervised driving style characterization

[0059] During normal driving maneuvers, a driver's long-term longitudinal vehicle control can be used to determine driving behavior regardless of the vehicle's dynamic response. For example, a driver may exhibit a specific longitudinal control pattern while driving on a highway for an extended period. Their accelerator pedal activation pattern may be smooth or abrupt, even in the absence of emergency conditions. The variability of the pedal and its rate of change can be used to differentiate between smooth and abrupt application. Such smooth or abrupt application shows a strong correlation with fuel economy and acceleration performance when driving conditions are unrestricted. Identifying such driving behaviors can be used, for example, to implement a fuel economy advisor.

[0060] Anomaly detection can be used to estimate significant changes in the overall variability of driver actions, indicating alterations in corresponding behaviors. Anomaly detection is a technique that relies heavily on continuous monitoring, machine learning, and unsupervised classification to identify trends deviating from normal behavior and predict potentially significant changes. The determinant of the covariance matrix of a driver's actions can be used as a measure of the generalized variance (dispersion) of the population and therefore as an indicator of a change in driver behavior.

[0061] The feature space of the driving torque requirement τ d and its derivative is given by the vector y = [τ d τ̇ d] spans. The determinant D of the population covariance matrix can be calculated recursively as Dk+1=(1−α)k−1Dk(1−α+(γk−vk)Qk(yk−vk)T) with vk+1=(1−α)vk+αyk Qk+1=(I−Gk(yk+−vk))Qk(1−α)−1 Gk+1=Qk(yk−vk)Tα(1−α+α(yk−vk)Qk(yk−vk)T where v k a filtered version of y k is, Q k the estimated inverse covariance matrix and α is a constant that reflects the forgetting factor in terms of the filter memory depth.

[0062] The D calculated in (21) k It presents initial means and standard deviations for abrupt and smooth behaviors. The instantaneous behavior is classified as abrupt if its value exceeds a control threshold l. abrupt , and is classified as soft if its value is lower than a tax threshold u smooth · l abrupt and u smooth are as l abrupt = µabrupt - 3σ abrupt , u smooth , = µ smooth + 3σ smooth defined, where µ abrupt and σ abrupt The mean and standard deviation of the class of abrupt behavior are µ. smooth and σ smooth are also defined for the class of smooth behavior. If the current behavior is classified as either abrupt or smooth, the corresponding mean and standard deviation of the corresponding behavior are recursively updated. wk+1=(1−β)wk+βDk+1 Hk+1=(1−β)Hk+(β−β2)(Dk+1−wk)T(Dk+1−wk) σk+1=(Hk+1)1 / 2 where w and H are the estimated mean and variance, and □ is another forgetting factor.

[0063] Fig.Figure 12 shows the determinant of the covariance matrix of the accelerator pedal position vector and its rate of change for eight runs of vehicle tests. The four runs with solid lines representing the determinant were for abrupt accelerator pedal applications. These determinants show a large value, greater than 7, for example. The four runs with dashed lines representing the determinant were for smooth accelerator pedal applications. These determinants show a small value, less than 4, for example. Therefore, the magnitude of the determinant reveals unique information patterns that can be used to distinguish smooth from abrupt driving behavior.

[0064] Because the interactions between the driver and the driving environment involve frequent vehicle stops of varying durations, interrupting the continuous update may be necessary to prevent numerical problems during recursive computation. The following interruption conditions can be used: (i) when the vehicle speed is less than 1 mph, recursive computations related to vehicle speed and acceleration are interrupted. (ii) when the accelerator pedal position is less than 1%, recursive computations related to the pedal are interrupted.

[0065] Although the above deviation focuses on the accelerator pedal, it can easily be applied to braking. Since sudden, aggressive braking can occur during emergency situations (which do not necessarily reflect the driver's general behavior), quasi-steady-state driving, where braking is not extreme, can be used for computational selection.

[0066] During temporary acceleration and deceleration, certain wheels of the vehicle may experience significant longitudinal slip, and the longitudinal tire forces on these wheels may reach their peak values. Such conditions can be identified by monitoring the rotational motion of individual wheels in relation to the vehicle's speed, and consequently, driver behavior during transitional maneuvers can be determined, as discussed above. V. Semi-supervised driving style characterization

[0067] Not all driver inputs may be accessible through electronic control systems. However, certain variables can construct an input / output pair that can be used to derive the driver control structure. For example, during a vehicle following maneuver, the relative distance between the leading and following vehicles and the driver's braking and throttling requests are usually well coordinated. Here, we consider using a Tagaki-Sugeno model (TS model) to relate the variance of the driver's braking and throttling commands to the relative distance and speed between the leading and following vehicles.

[0068] A fuzzy system can use the signal-processed mean following distance (gap time) relative to the other vehicle, as well as the standard deviation of accelerator and brake pedal rate changes, to determine whether the driver is aggressive or cautious. The driver index score from the fuzzy calculation and rule evaluation can determine the driver's aggressiveness based on vehicle following, vehicle speed, and the driver's control actions on acceleration and deceleration.

[0069] For a real-time vehicle implementation, the recursive estimation of the mean and variance of a variable of interest is applied. The signal-conditioned average gap time at sampling time k can be calculated as gk=gk−1+α(Δsk / vfk−gk−1) where Δs× is the relative distance between the vehicle in front and the vehicle behind, and v fkThe speed of the following vehicle is α. α is a filter coefficient similar to the one used in (22). Fig. 13A and Fig. Figure 13B shows the mean gap times calculated from two runs of a vehicle test: one for aggressive driving and the other for cautious driving.

[0070] The average accelerator pedal rate can be calculated as ρ¯k=ρ¯k−1+α((ρk−ρk−1) / ΔT−p¯k−1) where ρ is the average accelerator pedal input and ΔT is the sampling time. The corresponding variance can be calculated as υkαυk−1+(1−α)(ρk−ρ¯k)2 and the standard deviation is obtained from the square root of the variance. Fig. 14A and Fig. Figure 14B shows the standard deviations of two runs of test data for aggressive and cautious driving.

[0071] Similar to (25) and (26), the mean and variance of the brake pedal rate change can be calculated. Fig. 15A and Fig.Figure 15B shows the standard deviations of two runs of test data for aggressive and cautious driving. The variables are first normalized before being passed to the fuzzy derivative system. The fuzzy theorems and membership functions were determined for the features to transform the crude inputs into fuzzy terms. The fuzzy theorem G s The mean gap time is defined by Gs={(g,μ(g))|g∈G} where G is given by the limited collection of gap times g in the vehicle path. The gap time membership function µ is chosen as a Gaussian function.

[0072] A zero-order TS model was used to calculate the driver index level. A normalized output scale from 0 to 1.0 represented the levels of cautious to low-aggressive to aggressive driving behavior. The driver index is obtained from the fuzzy calculation and rule evaluation. Table 1 shows the rules used. It should be noted that a longer gap time indicates a relatively more safety-conscious driving style compared to a shorter gap time. Table 1 Rules for characterizing driving behavior When the gap time is If the accelerator pedal rate standard is If the brake pedal rate standard is Then the driver index Low Low Low Low-aggressive High Low Low Cautious Low High Low Aggressive Low Low High Aggressive Low High High Aggressive High High High Low-aggressive High Low High Cautious High High Low Not very aggressive

[0073] Fig. 16A and Fig. Figure 16B shows the driver index calculated from two runs of vehicle test data: one for aggressive driving with a driver index greater than 0.8 and the other for cautious driving with a driver index less than 0.2. VI. Monitored Driving Style Characterization

[0074] The vehicle following task requires the driver to maintain one of the following with the vehicle in front: (i) a speed difference of zero; (ii) a constant relative distance; and (iii) a constant relative gap time, defined by dividing the relative distance by the relative speed.

[0075] A human driver can be modeled as a PD feedback control unit. The closed-loop control system during a vehicle tracking maneuver can be expressed as (x¨l−x¨f−x¯¨g)=−cv(x˙l−x˙f−x¯˙g)−cs(xl−xf−x¯g) where x l and x f The distance between the vehicle in front and the vehicle behind is x gThe gap offset reference is used. Due to the implementation of radar, which is used in vehicles equipped with an adaptive cruise control function, the relative distance and relative speed are measured and defined as Δs=xl−xf,Δv=x˙l−x˙f

[0076] A vehicle equipped with stability control features a longitudinal accelerometer with the output α x up, the ẍ f measures. (28) can further be expressed as ax=cv(Δv−x¯˙g)+cs(Δs−x¯g)+(x¨l−x¯¨g)

[0077] The unknown parameters c v and c s The values ​​in (30) can be used to characterize a driver's control structure during vehicle tracking. Using the low-pass filtered Δs and Δv to replace the gap offset reference x g and their derivative x¯˙g And taking into account the time delays, we have the following equations axk+i=cv[Δsk−μk(Δs)]+cs[Δvk−μk(Δv)]+w[μk(Δs)μk(Δs)]=(1−α)[μk−1(Δs)μk−1(Δv)]+α[ΔskΔvk] where the subscript i is in α xk-i The time delay between the driver's brake / throttle application and the relative distance and measured relative velocity and acceleration is reflected, α is a low-pass filter coefficient similar to the one used in (22), and w is an uncertain high-frequency signal that can be treated as white noise. Using a conditional least-squares identification algorithm, c v and c s can be identified in real time from (31). The response time t p and the damping ratio ζ of the system with the driver in the loop can be determined using c v and c s to be related as tp=2πcs / 4cs−cc2, ς=cv / 2cs This can be used to infer the driver's driving behavior: (i) for a normal driver it is desirable that the system's temporary response to the driver in the loop is fast (sufficiently small t) p , e.g. less than 0.5 s) and damped (sufficiently large ζ); (ii) for an elderly driver or a driver with a physical disability t p be large; (iii) for an aggressive driver, ζ is likely to show a small value, such as less than 0.5, and the system response is likely to exhibit excessive overshoot; (iv) for a cautious driver, ζ is likely to show a reasonably large value, such as greater than 0.7.

[0078] A least squares parameter identification was used to calculate c v and c sImplemented. Two runs of a vehicle test were conducted. In the first run, the driver in the following vehicle attempted to use aggressive throttle and braking to achieve a constant relative gap time between his vehicle and a preceding vehicle, resulting in a larger distance error Δs. k - µ k (Δs) led. See Fig. 17. The identified c v is approximately 0.2 and the identified c s is approximately 0.05. See Fig. 18. The damping ratio calculated from (32) showed a value of less than 0.5, which indicates a slightly damping system with the driver in the loop, and therefore corresponds to aggressive driving behavior.

[0079] In the second run, the driver used cautious throttle and brake application to achieve vehicle tracking; the relative distance error Δs k - µ k (Δs) in Fig.19 had a lower amplitude compared to the one in Fig. 17 shown. The identified c v and c s are in Fig. Figure 20 shows the damping ratio to be greater than 0.8, except during the first 150 seconds. See [reference]. Fig. 20. This is a specification of a highly damping system with the driver in the loop, which therefore corresponds to a cautious driving behavior. VII Applications

[0080] Handling thresholds and / or driving style characterization, with appropriate monitoring and reinforcement, can be used for driver guidance, driver training, driver monitoring, and safety enhancement. Another potential application is the opportunity for vehicle personalization by adjusting control parameters to match the specific driver's style. For example, the ESC or brake control system can utilize such driver style characterization to adjust the activation threshold to suit individual driving behavior. As an example, an experienced driver might require less frequent ESC activations compared to a less experienced driver facing the same driving conditions. (However, there may be a minimum requirement to adjust the thresholds such that an error by an experienced driver can still be mitigated by the ESC function.)

[0081] As another example, steering sensitivity (the degree of vehicle steering response to a given steering input) and accelerator pedal sensitivity (the degree of vehicle acceleration response to a given accelerator pedal input) can be tuned based on driver characterization. The steering wheel and / or accelerator pedal can be made more sensitive if the driver is characterized as experienced (resulting in greater vehicle responsiveness). The steering wheel and / or accelerator pedal can be made less sensitive if the driver is characterized as cautious (potentially leading to improved fuel economy). Other applications are also possible.

[0082] Fig.Figure 21 is a block diagram of an embodiment of an advisory system 30 for a vehicle 32. The advisory system 30 may include several vehicle condition sensors 34 (e.g., yaw rate sensor, steering angle sensor, lateral acceleration sensor, longitudinal acceleration sensor, wheel speed sensor, brake pressure sensor, etc.), one or more control units 36 configured to perform, for example, electronic stability control, anti-lock braking and traction control, and the handling limit and driver characterization described above, an audio and / or visual indicator system 38 (e.g., a display panel, a speaker system, an LED array, a USB port, etc.), and a vehicle input and / or vehicle control system 40 (e.g., an accelerator pedal, a powertrain control unit, a steering wheel, etc.).

[0083] The vehicle condition sensors 34 can detect the various parameters described above, such as vehicle speed, wheel slip, etc., which characterize the movement of the vehicle 32 (e.g., current handling condition and limit handling condition), as well as the driver inputs described above (e.g., accelerator and brake pedal position, etc.). The one or more control units 36 can use this information as inputs to the handling limit and / or driver characterization algorithms described above. Based on the output of these algorithms, the one or more control units 36 can, as described above, (i) activate the audio and / or visual indicator system 38 to, for example, warn or train the driver, or (ii) modify aspects of the vehicle input and / or vehicle control system 40 to specifically adapt the vehicle's response to the type of driver.

[0084] For example, based on information acquired by the sensors 34, the one or more control units 36, which execute the algorithms described above, can determine that a driver of the vehicle 32 is driving recklessly. Since the driver has been classified as reckless, the one or more control units 36 can, for example, begin to issue driving instructions via the audio and / or visual indicator system 38 to encourage the driver to change their behavior. The one or more control units 36 can also / alternatively activate haptic elements of the vehicle input and / or vehicle control system 40, such as a haptic accelerator pedal, to warn the driver of their reckless behavior.

[0085] A memory accessible through one or more control units 36 can contain a database of instructions (audio and / or visual) mapped to specific predefined rules. An example rule might be that if a driver is reckless and the rate of steering angle change falls within a certain defined range for a specified duration (that is, the driver continues to rapidly turn the steering wheel clockwise and counterclockwise), then the driver is instructed to reduce their steering inputs.

[0086] Fig. 22 is a block diagram of a further embodiment of an advisory system 130 for a vehicle 132, wherein the same symbols are used for descriptions similar to those of Fig.21. In this embodiment, however, the vehicle 132 includes a token recognition system 142 configured in a known manner to recognize a token 144. For example, the token 144 can be a key with an identification chip that identifies a specific driver or a specific class of drivers (e.g., juvenile drivers). In this example, the token recognition system 142 can include a chip reader arranged in a known manner within an ignition system of the vehicle 132 to read the identification chip and transmit this information to the one or more control units 136. As another example, the token 144 can be a key fob or a plastic card with an embedded RFID chip.In this example, the token recognition system 142 can include an RFID chip reader arranged in a known manner within the vehicle 132 to detect and read the RFID chip and transmit this information to the one or more control units 136. As yet another example, the token 144 can be a mobile phone. In this example, the token recognition system 142 can include known modules (such as Ford's SYNC technology) configured to detect the mobile phone and transmit this information to the one or more control units 136. Other arrangements and scenarios are also possible.

[0087] A variety of functions can be implemented based on rider identification and the handling limit and / or riding style characterization described above. For example, if the token recognition system 142 provides information to the one or more control units 136 that identify the rider as a juvenile rider, the one or more control units 136 can issue instructions to the rider via the audio and / or visual indicator system 138 to reduce the throttle or brake as the handling limit is approached. The one or more control units 136 can implement rule-based commands similar to those described above to effect such a riding instruction. (An example rule might be that if the rider is a juvenile rider and if the handling limit is less than 15%, then the rider is instructed to slow down.)

[0088] If the token recognition system 142 provides information to the one or more control units 136 that identify the driver as a juvenile driver, then, as a further example, the one or more control units 136 can record a history of handling limit span and / or driver style characterization calculations to generate reports describing driving behavior. These reports can, for example, detail the number of times the juvenile driver exceeded certain handling limit span levels during a given trip. These reports can also, for example, describe the juvenile driver as cautious, aggressive, reckless, etc., during any given trip. Such reports can be accessed, reported, or displayed in any suitable / known manner, such as through the audio and / or visual indicator system 138.

[0089] As a further example, driver training and / or driver instruction can be implemented based on the driver identification described above. The one or more control units 136 can, for instance, issue instructions to encourage a driver to provide driving inputs to the vehicle 132 that will result in the driver being classified as cautious. For example, if a driver begins to accelerate and brake frequently, the one or more control units 136 can instruct the driver, via the audio and / or visual indicator system 138, to increase the distance between the vehicle 132 and the vehicle in front of it in order to reduce the frequency of acceleration and braking. Rules similar to those described above, or any other technology with suitable intelligence, such as neural networks, etc., can be used to facilitate the instructions.In embodiments where driving behavior is recorded for a later report, the fact of whether the driver follows or ignores the instructions can also be recorded as an indication of the driver's behavior.

[0090] Fig. Figure 23 is a block diagram of yet another embodiment of an advisory system 230 for a vehicle 232, wherein the same symbols are used for descriptions similar to those of Fig. 21. This embodiment includes a radar and / or camera system 246 (however, any suitable forward detection system may be used) that can periodically / continuously detect the distance between the vehicle 232 and another vehicle in front of the vehicle 232 in a known manner. (Although not shown, the advisory system 230 may also include a token recognition system and associated capabilities similar to those described in reference to Fig. (22 discussed. Other configurations are also possible.)

[0091] Under certain circumstances, the distance information collected by the system 246 can be monitored by the one or more control units 236. If the distance is less than a certain predefined threshold (e.g., 20 feet), the one or more control units 236 can warn the driver via the audio and / or video indicator system 238 and / or active elements of the vehicle input and / or control systems 240, if they are of a haptic type (e.g., haptic accelerator pedal, haptic steering wheel, haptic seat, etc.).

[0092] Under other circumstances, the distance information X collected by System 246, together with the change in distance over time, V x , and the longitudinal acceleration of the vehicle 232, A x , from which one or more control units 236 are used to calculate a time until collision, t c, with the vehicle in front of vehicle 232 to be determined by the following equation tc=−Vx±(Vx)2+2(Ax)(X)(Ax) or tc=XVx

[0093] If the time until the collision is less than a predefined threshold, one or more control units 236 can warn the driver as described above.

[0094] Under further circumstances, the one or more control units 236 can base the warning threshold on the gap time (discussed above) between the vehicle 232 and the vehicle in front of it. If the gap time is less than a threshold, the one or more control units 236 can, for example, activate (vibrate) haptic elements of the vehicle input and / or vehicle control systems 240, etc.

[0095] Alternatively, one or more control units 236 can access tables of distance and speed / relative speed to determine when to warn the driver. For example, if the distance and speed fall within a certain range, one or more control units 236 can activate a haptic accelerator pedal. Other scenarios are also possible.

[0096] The intensity (frequency and / or amplitude) with which haptic elements of the vehicle input and / or vehicle control systems 240 are activated can depend on the distance, the time to collision, the gap time, etc., between the vehicle 232 and the vehicle in front of it. For example, the intensity can increase when these parameters decrease. This increasing intensity can signal increasing urgency.

[0097] The predefined thresholds can depend on the type of driver. That is, the one or more control units 236 can implement the driver characterization algorithms discussed above and, based on this characterization, increase or decrease the warning threshold. For example, a threshold can be decreased for an experienced driver, as they are less likely to experience an accident as a result of tailgating. A threshold can be increased for a reckless or aggressive driver, as they are more likely to experience an accident as a result of tailgating, and so on.

[0098] The predefined thresholds can be modified based on whether the driver heeds the warning they provide. For example, if a driver does not increase their following distance to the vehicle in front after activating a haptic accelerator pedal, the predefined following distance used to trigger the haptic pedal activation can be reduced to prevent it from becoming a distraction for the driver. The predefined thresholds associated with the time to collision and the gap time can also be reduced. However, a minimum threshold can be set below which the reduction will not occur.

[0099] As is evident to those skilled in the art, the algorithms disclosed herein can be supplied to a processing device, which may comprise any existing electronic control unit or purpose-built electronic control unit, in many forms, including, but not limited to, information permanently stored on non-writable storage media such as ROM devices, and information modifiably stored on writable storage media such as floppy disks, magnetic tapes, CDs, RAM devices, or other magnetic and optical media. The algorithms can also be implemented in an executable software object. Alternatively, the algorithms can be embodied wholly or partially using suitable hardware components, such as…application-specific integrated circuits (ASICs), state machines, control units or other hardware components or devices, or a combination of hardware, software and firmware components.

[0100] Although embodiments of the invention have been presented and described, it is not intended that these embodiments explain and describe all possible forms of the invention. Rather, the words used in the patent description are descriptive rather than limiting, and of course, various modifications can be made without deviating from the concept and scope of protection of the invention.

Claims

[1] Vehicle comprising the following: an identification system located inside the vehicle and configured to capture information from a token in the vehicle's environment and to classify a driver of the vehicle based on that information; a sensor array located inside the vehicle and configured to measure several parameters representing the vehicle's current handling condition and the vehicle's limit handling condition; and at least one control unit in communication with the identification system and the sensor array, configured to determine a range between the current handling condition of the vehicle and the limit handling condition, to record a history of the range if the driver classification is of a special type, and to characterize the long-term behavior of the driver, combining several driver characterization possibilities and calculating fuzzy probabilities that the driver classified as being of the special type falls into one of several categories. [2] Vehicle according to claim 1, wherein the at least one control unit is further configured to characterize the dynamic control of the vehicle by the driver on the basis of the span and to record a course of characterization. [3] Vehicle according to claim 2, wherein characterizing the dynamic control of the vehicle by the driver on the basis of the span comprises determining a time period during which the span falls into a predetermined range and recording a time series of the time period during which the span falls into the predetermined range. [4] Vehicle according to claim 1, further comprising an interface configured to provide access to the recorded span history. [5] Vehicle according to claim 1, wherein the identification system is configured to acquire information from a high-frequency identification chip and / or a mobile phone. [6] Vehicle comprising the following: an identification system located within the vehicle and configured to capture information from a token in the vehicle's environment and to classify a driver of the vehicle based on that information; and at least one control unit in communication with the identification system, configured to define categories of driver behavior based on a history of driver torque demands and associated rates of change thereto, to classify the current driver behavior into one of the defined categories based on the current driver torque demands and associated rates of change thereto, to record a history of the classification if the driver classification is of a special type, and to characterize the long-term behavior of the driver, combining several driver characterization possibilities and calculating fuzzy probabilities that the driver classified as being of the special type falls into one of several categories. [7] Vehicle according to claim 6, further comprising an interface configured to provide access to the recorded classification history. [8] Vehicle according to claim 6, wherein the identification system is configured to acquire information from a high-frequency identification chip and / or a mobile phone. [9] Vehicle according to claim 6, wherein defining categories of driver behavior based on a history of driver torque demands and associated rates of change therether comprises defining category thresholds based on a mean and standard deviation of the history of driver torque demands and associated rates of change therether. [10] Vehicle comprising the following: a single-position accelerator pedal; a brake pedal with one position; an identification system located within the vehicle and configured to capture information from a token in the vehicle's environment and to classify a driver of the vehicle based on that information; and at least one control unit configured to determine a gap time between the vehicle and another vehicle, to determine a variability of a rate of change of the accelerator and / or brake pedal position, to characterize the longitudinal control of the vehicle by a driver based on the gap time and the variability of the rate of change, to record a history of the characterization if the driver classification is of a special type, and to characterize the long-term behavior of the driver, combining several driver characterization possibilities and calculating fuzzy probabilities that the driver classified as being of the special type falls into one of several categories. [11] Vehicle according to claim 10, further comprising a forward detection system configured to detect a distance between the vehicles. [12] Vehicle according to claim 11, wherein determining a gap time between the vehicle and another vehicle is based on the distance between the vehicles and a speed of the vehicle. [13] Vehicle according to claim 10, wherein characterizing the longitudinal control of the vehicle by a driver on the basis of the gap time and the variability of the rate of change comprises generating a driver index by evaluating the classifications of the gap time and the variability of the rate of change on a rule basis. [14] Vehicle according to claim 13, wherein a value of the driver index represents a type of longitudinal control of the vehicle. [15] Vehicle according to claim 10, further comprising an interface configured to provide access to the recorded history of the characterization. [16] Vehicle according to claim 10, wherein the identification system is configured to acquire information from a high-frequency identification chip and / or a mobile phone. [17] Vehicle comprising the following: an identification system located within the vehicle and configured to capture information from a token in the vehicle's environment and to classify a driver of the vehicle based on that information; and at least one control unit in communication with the identification system, configured to characterize the longitudinal control of the vehicle by a driver based on a distance and relative speed between the vehicle and another vehicle and a longitudinal acceleration of the vehicle, to record a history of the characterization if the driver classification is of a special type, and to characterize the long-term behavior of the driver, combining several driver characterization possibilities and calculating fuzzy probabilities that the driver classified as being of the special type falls into one of several categories. [18] Vehicle according to claim 17, further comprising an interface configured to provide access to the recorded history of the characterization. [19] Vehicle according to claim 17, wherein the identification system is configured to acquire information from a high-frequency identification chip and / or a mobile phone. [20] Vehicle according to claim 17, wherein characterizing the longitudinal control of the vehicle by a driver on the basis of the distance and relative speed between the vehicles and the longitudinal acceleration of the vehicle includes determining the reaction time of a driver to changes in the distance between the vehicles.

Images