WO2010011806A1 - System and method for evaluating and improving driving performance based on statistical feedback - Google Patents

Abstract

A system and method for evaluating and improving dnving performance based on statistical feedback Embodiments of the present invention allow for comparison of an individual driver's performance when compared to that of a group or population of drivers Embodiments can also be used to provide feedback to a particular driver as to their specific performance in comparison with a statistical sample culled either from an analysis of that particular driver's previous trips, or an analysis of the larger population of drivers To perform these comparisons, embodiments allow for the collection of profiles of driving-related and other data The data can be collected based on a location of a vehicle, in addition to other factors such as speed and recent historical data.

Description

SYSTEM AND METHOD FOR EVALUATING AND IMPROVING DRIVING PERFORMANCE BASED ON STATISTICAL FEEDBACK

Inventor: Christopher Wilson

COPYRIGHT NOTICE

A portion of the disclosure of this patent document contains material which is subject to copyright protection. The copyright owner has no objection to the facsimile reproduction by anyone of the patent document or the patent disclosure, as it appears in the Patent and Trademark Office patent file or records, but otherwise reserves all copyright rights whatsoever.

Field of Invention:

[0001] Embodiments of the invention are generally related to systems for using digital map and other location-related information in vehicles, and particularly to a system and method for evaluating and improving driving performance based on comparisons with behavior of other drivers collected from other vehicles.

Background:

[0002] In the United States and elsewhere, vehicle drivers are tested as to their driving skills before they can obtain a license to operate a vehicle. This testing is particularly important if the driver is to be licensed in a commercial setting or will be operating a commercial vehicle. Traditionally, an individual's driving skills are tested using a combination of a written test portion and a driving test portion. While the written portion is usually a list of scenario-related questions and can be objectively evaluated, the driving portion is usually subjectively evaluated by a human evaluator riding along with the driver and making notes of the driver's performance. Generally, no testing entity provides a means of testing or measuring a driver's actual driving skills in an objective manner.

[0003] Once a driver has been tested, licensed, and is actually operating a vehicle on a daily basis, there exists a number of techniques by which a driver's driving skills can be periodically tested. However, many of these techniques are again subjective, such as reevaluation of the driver by a human evaluator, or by the signage on commercial vehicles that invites a third-party observer to call a telephone number and provide comments on the driver's performance, or by similar means. This information can be tracked and used to review a driver over an extended period of time. However, such information is not generally focused on a driver's overall skill level, but concentrates more on the reporting of isolated incidents that warrant further specific review.

[0004] Some companies use a tachograph device to record general information about a vehicle's movements, particularly with delivery and public transportation vehicles. Generally, the tachograph monitors a vehicle's speed and the distance traveled over a period of time, for example a day or a week, and records the information in a storage medium where it can be subsequently accessed for analysis. Traditional tachographs stored their data in a paper chart format and were difficult to decipher. Modern systems store the data in a digital or computer format. Another commonly used in-vehicle systems is designed to record vehicle information as part of an event data recorder - a form of "black box". However, a tachograph is essentially a macro recording device, and while it may be useful for tracking a vehicle's overall usage, cannot provide an indication of the driver's overall skills, especially under specific conditions. Conversely, an event data recorder typically provides only a momentary snapshot of the actions of the driver and his/her vehicle while addressing or reviewing a specific situation, such as in the context of a vehicle accident, and is not reflective of the driver's overall skill level or historical usage of the vehicle.

[0005] Finally, even if currently available testing environments could adequately test a driver's skills, there is no readily available process to objectively determine a good set of driving skills, say within a particular environment, and to then teach those same driving skills to another driver. While basic driving skills can be learned using books and classes (for example, how to operate a vehicle at an appropriate speed, how to make a correct turn, or how to merge on or off a highway), it is difficult to learn in any objective manner how to then expand upon and adapt one's driving skills to best suit a particular environment (for example, a particular network of roads, or a particular route to be taken). There is currently no process for allowing a driver to learn from the experiences of other drivers in a similar environment; nor is there any process to determine a driver's existing set of skills within a particular environment and then teach that driver how to improve their driving skills in that same or another environment. It is estimated that an average driver might improve their fuel economy by up to 20% if they could be taught appropriately. It is also the case that novice drivers, those that lack the 'on the job' training, are much more frequently involved in accidents. This inability to easily quantify or share knowledge of driving skills generally leads to unsafe and/or inefficient driving practices. These are some of the areas that embodiments of the present invention are designed to address.

Summary:

[0006] Described herein is a system and method for evaluating and improving driving performance using feedback based on deviations from previously observed behaviors (possibly statistical behaviors). Embodiments of the present invention allow for comparison of an individual driver's performance when compared to that of a group or population of drivers. Embodiments can also be used to provide feedback to a particular driver as to their specific performance in comparison with a (possibly statistical) sample culled either from an analysis of that particular driver's previous trips, or an analysis of the larger population of drivers, or an analysis of some particular 'representative' population of drivers. To perform these comparisons, embodiments allow for the collection of profiles of driving-related and other data. The data can be collected based on a location of a vehicle, in addition to other factors such as speed, acceleration, fuel consumption, and/or recent historical data. A set of baseline data can be stored in a geographic database, map database, digital map, or other form of data storage or database. A measure of comparative performance can then be determined and fed back to the driver and/or the vehicle. Data describing an individual driver may also be archived and later reviewed by others to assess the driving skills of an individual driver, including for example to ascertain a driver's insurance premium, or to better teach other drivers to perform in an efficient manner within a particular environment.

Brief Description of the Drawings

[0007] Figure 1 shows an overview of the process used within a system and method for evaluating and improving driving performance, in accordance with an embodiment.

[0008] Figure 2 shows an illustration of the type of data can be collected during driving and used with the system, in accordance with an embodiment.

[0009] Figure 3 shows an illustration of a road network, a vehicle, and positions or time intervals in accordance with an embodiment.

[0010] Figure 4 shows an illustration of a road network and a speed profile for vehicles at a first location or time interval, in accordance with an embodiment. [0011] Figure 5 shows an illustration of a road network and a speed profile for vehicles at a second location or time interval, in accordance with an embodiment. [0012] Figure 6 shows an illustration of a road network and a speed profile for vehicles at a third location or time interval, in accordance with an embodiment.

[0013] Figure 7 shows an illustration of different speed profiles for vehicles at different locations, illustrating the statistical distribution of vehicle speeds, in accordance with an embodiment.

[0014] Figure 8 shows an illustration of a road network and speed measurements for a new vehicle as compared to the statistical profiles, in accordance with an embodiment. [0015] Figure 9 shows an illustration of how the system can be used for evaluating and improving driving performance, in accordance with an embodiment.

[0016] Figure 10 shows an illustration of the components of an in-vehicle system for evaluating and improving driving performance, in accordance with an embodiment. [0017] Figure 11 shows a flowchart of a method for evaluating and improving driving performance, in accordance with an embodiment.

[0018] Figure 12 shows an illustration of a road network and speed measurements for a new vehicle as compared to the statistical profiles, in accordance with an embodiment. [0019] Figure 13 shows an illustration of an embodiment for use as a speed warning system.

[0020] Figure 14 shows an illustration of an embodiment for use as a lane departure warning system.

[0021] Figure 15 shows an illustration of an embodiment for use as an environmental- conscious driving guide or game.

Detailed Description:

[0022] As described above, there is no readily available process to objectively determine a good set of driving skills, say within a particular environment, and to then teach those same driving skills to another driver. While basic driving skills can be learned using books and classes, it is difficult to learn in an objective manner how to adapt one's driving skills to best suit a particular environment. There is currently no process for allowing a driver to learn from the experiences of other drivers in a similar environment; nor is there any process to determine a driver's existing set of skills within a particular environment and then teach that driver how to improve their driving skills in that same or another environment. This inability to easily quantify or share knowledge of driving skills generally leads to unsafe and/or inefficient driving practices. These are some of the areas that embodiments of the present invention are designed to address.

[0023] Described herein is a system and method for evaluating and improving driving performance based on statistical feedback. Embodiments of the present invention allow for statistical comparison of an individual driver's performance when compared to that of a specific group of drivers (for example, a specific driver type, or a specific type of vehicle); or with that particular individual driver's previous trips; or with a larger population of drivers as a whole, (including random samples take from that population). Embodiments can also be used to provide feedback to a particular driver as to their specific performance in comparison with other drivers. To perform these comparisons, embodiments allow for the collection of driving-related data (for example, in-vehicle and sensor-based measurements), environmental data (for example, fuel consumption) and other relevant data, over different locations. As disclosed herein, a "road profile" captures and stores data about how a collection of vehicles have behaved a particular location, or within a particular area or region, or in a particular scenario. A 'driver profile' can then be used to describe how a particular driver ranks in comparison with other drivers (e.g. the road profile) at that location, region, or scenario. In accordance with an embodiment, the data can be collected based on a location of a vehicle, in addition to other factors such as speed and recent historical data (for example, vehicle dynamics, and configuration such as turning on/off the air conditioning). Information about a particular vehicle (for example, EPA mileage ratings, safety performance, type of vehicle, and/or tires) can also be used to assign weights to the data. The data can then be correlated based on the most relevant attributes, since depending on the performance metric different attributes may be more relevant than others. A set of baseline data can be stored in a geographic database, map database, digital map, or other form of data storage or database.

[0024] Depending on the embodiment, the database can be located within a vehicle in an onboard database. Alternatively, the data can be provided to the vehicle in real time from an external centralized or distributed data server. Some combination of on-board and off-board storage may also be used. In accordance with an embodiment the database can be stored in conjunction with, or as part of an electronic map or digital map. [0025] A measure of comparative performance can then be determined and fed back to the driver and/or the vehicle and/or some supervisory entity. Date describing an individual driver may also be archived and later reviewed by others to assess the driving skills of an individual driver, including for example to ascertain a driver's insurance premium, or to better teach other drivers to perform in an efficient manner within a particular environment. [0026] Figure 1 shows an overview of the process used within a system and method for evaluating and improving driving performance, in accordance with an embodiment. As shown in Figure 1 , in step 100, driving-related data is collected from an individual vehicle or from a plurality of vehicles within an environment. In step 104, the data is statistically analyzed to develop a profile of vehicle behavior at one or more locations in the environment. The accuracy by which a location is determined can depend on the particular embodiment or application to which the system is directed. For example, if the system is implemented as a means of providing an upcoming stop-sign warning, then the trigger for the warning can be the speed at a location, say 10 meters in front of the stop line on the road. For other applications, such as average fuel consumption, the location can be determined in a less exact manner, for example extended over many meters, or to a road segment such as "the stretch of Interstate I80 from San Francisco to Sacramento". In this manner the region of interest can be varied from small or local points to much larger continental-sized areas.

[0027] A road profile can be determined by monitoring and recording data for a single vehicle over a period of time, or a plurality of vehicles, or a combination of these techniques. In accordance with an embodiment the profile can be methodically collected by monitoring vehicles passing that location; or can be somewhat automatically generated based on probe tracks or probe vehicles. In step 108, this profile information is uploaded and/or stored in a road profile or other database, together with other profiles corresponding to this driver or other drivers.

[0028] Depending on the particular implementation, the database can be stored onboard within the vehicle. Alternatively the database can be stored in an off-board, centralized or distributed infrastructure wherein the vehicle is able to obtain views of the database either in real-time or by downloading portions of road profile data from the database. In accordance with some embodiments the database can be stored in conjunction with, or as part of an electronic map or digital map, and can be stored and accessed using a GPS, navigation device, or other in-vehicle processing device. In those embodiments that access an off-board database, the data processing may be performed either on-board or off-board (for example, the vehicle can send an appropriate set of data or an observable to the centralized or distributed infrastructure; the comparisons performed off-board; and the results then either sent back to the vehicle or recorded for subsequent analysis). The road profile can also be updated as needed, for example when a new driver is added to the system, or when an existing driver makes new trips. Optionally a score may be developed for a driver based on their driving habits over the course of a trip, or a particular period of time. Trip scores can be generated by a weighted average of individual location scores.

[0029] In step 110, a driver profile can be generated to describe the relationship of a particular driver when compared to other drivers. In step 112, current trip data is collected from the vehicle for a location or a range of locations where the vehicle travels. In accordance with an embodiment the trip data corresponds to a profile, but generally represents the behavior of a single individual on a single or current trip. In step 116, the trip data is compared with a corresponding profile to determine where in that profile the current trip matches. A measure can also be added to indicate the confidence or "quality" of the road profile, i.e. to reflect that some road profiles may be based on the driving patterns of a small number of drivers using that road, and thus have a lower relative confidence; while other road profiles may be based on a larger number of drivers using that road, and thus have a higher relative level of confidence. This information is then used to determine the variance from the current trip. In those road profiles that have a higher relative level of confidence (i.e. the data is culled from a relatively larger number of drivers) then any deviation or variance outside of the distribution may of of particular significance. In step 120, depending on the variance (and optionally depending on the particular embodiment the driver's profile), the system can perform a number of tasks, including for example updating the driver's profile, provide immediate feedback to the driver or the vehicle, record variances for later analysis by a supervisor or as part of a training scheme, or calculate a score for use in an award scheme game, or for some other usage. [0030] Using the above-described technique, data is collected from a number of vehicles, an individual vehicle on multiple passes, a set of vehicles (for example a collective group of drivers), or any combination of these techniques. This data is then statistically analyzed to develop a profile of vehicle behavior at that location or a set of locations, and the profile is stored in a geographic database, where it can later be compared to actual or current trip data. Current data from the vehicle is collected from a location of range of locations where the vehicle travels, to reflect a current trip. The comparison is then used both to provide feedback to the vehicle and the driver about their trip, for information gathering about travel characteristics, and for further education of the driver, and other drivers.

[0031] In accordance with an embodiment, during either the creation of the profiles for storing in the database, and/or during a trip that will be used to compare with the profiles, information can be gathered from the vehicle or from other sources. Figure 2 shows an illustration of the type of data that can be collected during driving and used with the system, in accordance with an embodiment. As shown in Figure 2, the data set 124 collected by the vehicle (together with suggested applications that can utilize that data) can include some or all of the following in any variety of combinations:

• Speed profiles 125 (for use in such applications as speed warning, parental monitoring, and traffic control violation detection). Speed can be measured using digital speedometers, GPS and/ or in-vehicle sensors and systems.

• Instantaneous fuel consumption 126 (for use in applications such as an eco-driving 'competition', trip to trip comparison, or driver self training). Many modern vehicles have the ability to calculate and display instantaneous fuel consumption.

• Fuel consumption over a specific path, route, or trip 127 (similarly for use in applications such as an eco-driving 'competition', trip to trip comparison, or self training).

• Lateral lane position monitoring 128 (for use in applications such as a lane departure warning system).

• Turn indicators 129 (for use in navigation and turn indication assistance).

• Road position 130 (for use in navigation assistance based on the road position of other drivers who have taken the same route in the past).

• Route choice 131 (for use in applications such as route planning and optimization- based on actual driving by this driver, or by another driver on similar trips).

• Time over a specific route 132 (while generally not exceeding posted speed thresholds, for use in giving an indication of average speed and for use in supervision of the driver).

• Acceleration (both longitudinal and lateral) 133.

• Steering wheel and/or pedal movements 135 which may indicate drunk or impaired drivers, especially when compared with historical data for an individual driver and/or data from the reference population.

• Radar, camera, laser, or other sensor measurements 136. • Availability of GNSS or other signals 137.

• Additional forms of vehicle and driver information 138.

[0032] It will be evident that the examples of data and applications described above, including possible vehicle and driver information, are provided for purposes of illustration, and that other types of data, application and/or vehicle and driver information can be used or developed within the spirit and scope of the invention.

Generation and Use of Road Profiles

[0033] The following section describes how an embodiment of the invention can be used to create road profiles for a particular section or sections of road, and then use that information to match a driver profile and/or assess a driver's performance on a particular road section, and then provide feedback as necessary.

[0034] Figure 3 shows an illustration of a road network, a vehicle, and positions or time intervals in accordance with an embodiment. As shown in Figure 3, a road 146 is illustrated and having multiple road sections or locations along the road 150, 152, and a stop sign 153. A vehicle 148 travelling the road can be monitored for such attributes as road position, fuel consumption, or vehicle speed, and the information recorded. The process can then be repeated for other vehicles, again monitoring the other vehicles for such attributes as road position, fuel consumption, or vehicle speed, until a set of vehicle data for many vehicles travelling that road section is collected. As described above, in accordance with an embodiment, the vehicles can be probe vehicles.

[0035] Figure 4 shows an illustration of a road network and a speed profile for vehicles at a first location or time interval, in accordance with an embodiment. As shown in Figure 4, when a sufficient number of vehicles have travelled the road, and have provided data, a profile 154 can be generated that reflects the speed of vehicles at that point or region of the road. As shown in Figure 4, the profile for location (2) indicates that the speed of vehicles at that location have a reasonable rate of speed and an approximately normal distribution. [0036] Figure 5 shows an illustration of a road network and a speed profile for vehicles at a second location or time interval, in accordance with an embodiment. As shown in Figure 5, again when a sufficient number of vehicles have travelled that particular portion of the road, and have provided data, a profile 156 can be generated that reflects the speed of vehicles at that point or region of the road. As shown in Figure 5, the profile for location (3) indicates that the speed of vehicles at that location have a much lower rate of speed as they approach the stop sign, and a relatively wide distribution pattern.

[0037] Figure 6 shows an illustration of a road network and a speed profile for vehicles at a third location or time interval, in accordance with an embodiment. As shown in Figure 6, a profile 158 can be similarly generated that reflects the speed of vehicles approaching a corner in the road. As shown in Figure 6, the profile for location (6) indicates that the speed of vehicles at that location have a lower rate of speed than for location (2) above, and a relatively narrow distribution pattern.

[0038] When sufficient number of vehicles have been tracked in the location or road segments (or in those embodiments that use a single vehicle over multiple trips then when sufficient number of trips have been undertaken), a database of different speed profiles can be created for those locations. Figure 7 shows an illustration of different speed profiles for vehicles at different locations, illustrating the statistical distribution of vehicle speeds, in accordance with an embodiment. It will be evident that the illustration of Figure 7 shows different speed profiles as charts for illustrative purposes, but that in practical implementations the profile data can be stored in the database in a number of different ways. As shown in Figure 7, the profile database includes a plurality of road profiles 160, as collected using the steps described above. In accordance with the particular embodiment shown, each road profile indicates the speed distribution and a relative number of vehicle (which may be normalized or non-normalized) for particular locations and/or travel times. Standard interpolation techniques can be used to estimate distributions for other locations and/or travel times. This information then represents the statistical distribution of driver or vehicle performance (in this instance, vehicle speed) for a particular range of road sections or regions.

[0039] Figure 8 shows an illustration of a road network and speed measurements for a new vehicle as compared to the statistical profiles, in accordance with an embodiment. As a new (or existing) driver travels the road, information is received from the vehicle 166 and/or the driver or other sources. This information can then be compared 167 with statistical information from the road profile database, and used to provide driver feedback, guidance and/or automatic vehicle maneuvers 168. As shown in Figure 8, the arrows 165 indicate the speed of a particular driver within the overall population. Pictorially, at point (6) the driver is clearly exceeding the normal range for speed at this location. This information may be used to generate a warning based solely on the variance from the norm, without any knowledge of the reasons that the reference population may have slowed down at this particular point. In this case the slowing is presumed to be due to the curve in the road, but it might also be due to a school zone, construction work being performed during the reference period, the presence of an interesting advertising billboard, or any number of other reasons. Regardless of the underlying reason, the system can use statistical comparison to at least discern that the current driver may be driving outside of the norm when compared to other drivers. This approach can also be used in reverse to identify places where the road infrastructure should be checked as an unmotivated slowing may be a surrogate for accidents. Road authorities may wish to adjust speed limits or post curve warnings etc. in part on the basis of such profiles. It can also be used to mitigate driver distraction by for example, not allowing cell phone calls to be made at 'difficult' spots, as identified from the road profile. The information from 165 can also be combined to define or to update an individual's driver profile.

[0040] It will be evident that the example implementation described above with regard to speed measurement and approaching curve warning is merely one possible embodiment, and that additional embodiments and implementations can be developed, some of which are illustrated in further detail below.

System Implementation

[0041] Figure 9 shows an illustration of how the system can be used for evaluating and improving driving performance, in accordance with an embodiment. As shown in Figure 9, a vehicle 170 travels a roadway 174, that includes one or more curbs 178, lane and or road markings 182 , road signs 206, and other road objects 210 or structures. The road may also include, or be divided into, different locations, sections, or regions of interest 211 , here indicated as boxes (1 ), (2) and (3). Together, all of these road markings and objects, or a selection of the road markings and objects, can be considered a scene 202 or an environment for possible interpretation and profiling by the system. The road network, vehicle, and objects may be considered in terms of a coordinate system, including placement, orientation and movement in the x 190, y 194, and z 198 directions or axes.

[0042] Generally the system need only know about the vehicle, vehicle sensors (i.e. their location etc), and the driver's actions; it need not derive what objects in the "real world" those sensor readings may actually correspond with. In accordance with an embodiment, a map database that is stored either on-board within the vehicle, or at an off-board location, can be used to store the objects, or their mathematical equivalents, in addition to the traditional road network and road attributes. In accordance with some embodiments the database can be stored in conjunction with, or as part of a digital map, and can be stored and accessed using a GPS, navigation device, or other in-vehicle processing device. In accordance with some embodiments the system can make use of the additional information a digital map can provide, such as precise information about "real world" objects; and intangible information such as legally-imposed speed limits and other traffic restrictions.

[0043] As the vehicle travels in the environment at different locations, information about the vehicle and/or the driver is collected 220, and associated with a trip 224. , This information is then used to either create or update a new profile 230. The profile information for a driver over several trips, or for a group of drivers is stored as a plurality of profiles 236-260 in a profile database 234. In accordance with an embodiment the database and its profiles can be located within the vehicle in an onboard database. Alternatively, the database can be stored at an external centralized or distributed data server and accessed as needed by the vehicle. [0044] Figure 10 shows an illustration of the components of an in-vehicle system for evaluating and improving driving performance, in accordance with an embodiment. As shown in Figure 10, in accordance with some embodiments a vehicle profiling and feedback system and/or sensors can be embedded with or connected to software and a micro-processor in the vehicle. As shown in Figure 10, the system comprises a vehicle profiling and feedback system 280 that can be placed in a vehicle, such as a car, truck, bus, or any other moving vehicle. Alternative embodiments can be similarly designed for use in shipping, aviation, handheld navigation devices, and other activities and uses. The navigation system comprises a digital map or map database 2902, which in turn includes a plurality of object information. Alternately, some or all of the map database can be stored off-board and selected parts communicated to the device as needed. The system further comprises a positioning sensor subsystem 302. In accordance with an embodiment, the positioning sensor subsystem includes a combination of one or more on-board vehicle sensors 306, absolute positioning logics 310 and/or relative positioning logics 314, which together serve to obtain an initial estimate as to the absolute position or trajectory of the vehicle and/or provide relative position or bearing of the vehicle compared to an object. The navigation system further comprises a navigation logic 294. A vehicle feedback interface 298 receives the information about the status and position of the vehicle. This information can be used by the driver, or automatically by the vehicle. In accordance with an embodiment, the information can be used for driver feedback (in which case it can also be fed to a driver's navigation display 284. This information can include position and orientation feedback, and detailed route guidance. In accordance with an embodiment the system further comprises a profile interface 318 for accessing profile data, a profiled generation and update logic 322, and a vehicle and driver information interface 326 for collecting vehicle and driver information. In usage, the road profile database 234 can be accessed and an initial profile 330 retrieved or generated for a particular location in which the vehicle is travelling. In accordance with an embodiment, the initial profile can be inferred from a map or other data. Depending on the quantity of data in the profile (reflecting how well the profile represents the true population) the profile may or may not be suitable for certain applications. For example, for a speed advisory application, real time advice would not be provided to a driver unless the road profile was sufficiently accurate to ensure that the drivers rank against the entire population was accurate to within, say, five percentage points. Standard statistical methods can be used to determine the statistical accuracy of the road profile, and this data then stored with the profile. The profile information, together with vehicle-sensed and other data, can be used by the system to provide driver suggestions and/or feedback 344, and/or perform automatic vehicle actions and maneuvers, example of which are described in further detail below. During or after a trip the profile information can be communicated back to the database, where it is used to improve the overall accuracy of the profile information stored therein, provide a means of analyzing the driver's skills, and provide a learning source for subsequent trips by the same driver, or trips by other drivers in similar environments.

[0045] Figure 11 shows a flowchart of a method for evaluating and improving driving performance in accordance with an embodiment. As shown in Figure 11 , in step 360, the in- vehicle system determines a location within the environment in which the current trip is taking place. In step 364, a profile is retrieved from the profile database (if one is available). In step 368, while the driver is driving the system collects vehicle and/or driver information, such as the data described in Figure 2 or any combination of speed profiles, instantaneous fuel consumption, fuel consumption over a specific route, lateral lane position, turn indicators, road position, route choice, or time over a specific route. In step 372 the current statistics for the driver (and/or their vehicle) are compared with the road profile information. In step 376 the system can provide real-time driver suggestions and/or automatic vehicle actions and maneuvers based on the comparison with the road profile and, possibly, the driver's historical profile. The data can also be recorded, reported, or archived for subsequent use in various applications, examples of which are described below. In step 380 the data can be used to update the profile information in the database for this driver, or a group of drivers, for this location or environment.

Applicability to different Driver Types

[0046] Figure 12 shows an illustration of a road network and speed measurements for a new vehicle as compared to the statistical profiles, in accordance with an embodiment, and illustrates how different drivers may or may not be suitable for different applications. [0047] As shown in Figure 12, three driver profiles 169 are shown. Driver (1 ) is considered a very normal driver, practically the definition of average. If the system detects this person driving at the extremes then it may be indicative of some type of problem. This type of driver is an ideal candidate for receiving warnings or advisories at the ends of the distribution. Driver (2) tends to drive excessively fast. This type of driver is likely to be a good candidate for receiving training and feedback. They may not be as amenable to real-time warnings since they would likely receive many such warnings, and those might be more of a distraction than a benefit. Driver (3) only complies in an erratic fashion with the statistical distribution. This type of driver would be interesting to analyze further, to determine if the spot on the trimodal distribution is a function of other factors, such as impairment of the driver, the presence of other people in the vehicle, or environmental conditions.

Speed Warning System

[0048] In accordance with a particular embodiment, the systems and techniques described above can be used in a particular embodiment that functions as a speed warning system, and that allows a driver (or more correctly their vehicle) speed to be statistically compared to a stored profile or profiles. Figure 13 shows an illustration of an embodiment for use as a speed analysis, or speed warning system 398. Generally described, a speed analysis system can determine a drivers speed for a set of road, location, or traffic controls, including whether their speed is excessive, and provide some reporting after-the fact to the driver, including enforcing penalties as appropriate. A speed warning system attempts to warn drivers when their speed is excessive for the road, location, or traffic controls ahead, i.e. it performs a similar functionality to a speed analysis system, but in real-time, and with immediate feedback to the driver so that they cab take action. The same basic approach described above can be used for other systems, such as a stop sign warning system, a curve speed warning system, or a generalized speeding warning system.

[0049] As shown in Figure 13, the vehicle profiling and feedback system uses profile information together with the vehicle's current location and other collected vehicle and driver information to determine 406 where the current speed ranks on the speed profiles stored in the database for this location, time, and environment (taking into account any attributes of the vehicle, such as the vehicle type or weight, that may have a bearing on the most suitable speed), and whether this rank exceeds a speed threshold that has been established by fiat or by previous behavior (such as a driver profile). The system can then provide suggestions for feedback 418 to the driver to, say, reduce speed, or make automatic adjustments 414 to the vehicle to reduce the speed. The information can also be used to update the profile 410 for this driver, location and environment.

[0050] In accordance with an embodiment, speed information is collected for a large number of vehicles at a particular location on the roadway, as a function of direction of travel, and in some instances as a function of lane, weather, time-of-day, or any other parameters. The speeds from multiple vehicles are combined to provide a histogram of speed against number of vehicles. This is the road speed profile at a given point on the roadway, and represents the distribution of speeds for all vehicles at that location. In accordance with an embodiment the road speed profile can be stored in a digital map database and associated with the identified point on the road. At some points the road speed profile can be a function of the lane, time of day, or other factor. An individual vehicle speed at a given point on the roadway can be compared to the road speed profile at that point on the roadway. The output of this step is then a measure of the cumulative distribution of vehicles that drive faster and slower than the individual vehicle at that location (for example, it may be determined that at this location only 5% of vehicles drive faster while 95% of vehicles drive slower). This number represents a driver's speed percentile. Training of the system for an individual driver (possibly in association with a particular vehicle) can be effected by observing the drivers speed percentile over a period of time. This assumes that, on the average, a driver will tend to drive at the same percentile region of the speed distribution. The distribution of a particular driver's speed percentile over time is their driver speed profile. The speed profile may also be a function of time, weather, road classification or geometry, vehicle they are driving, or other input. The driver speed profile may also be constrained by various authorities (for example, parents, or licensing requirements) so that the profile does not contain relatively high-speed driving. The system can compare the road speed profile with the driver's speed at a particular location. If the driver's speed percentile falls outside of the driver's speed profile (for example, the driver is driving with a 95% speed percentile, but their profile shows them very rarely above an 80% percentile) then a warning can be provided to the driver. In accordance with an embodiment a threshold can be set to determine when a warning is given based on the percentage of time an individual driver spends in various parts of their speed profile.

[0051] In accordance with an embodiment, the system is not required to use a predefined "safe speed" for any particular location or corner (which would need to take into account a variety of weather-related, vehicle, driver and other safety-related parameters). Nor is the system required to use a "legally-permitted speed" (that might be set by some legal speed limit rather than for mechanical safety). The system can determine a suitable speed as a statistical indication of safer speed based on previously recorded and generally acceptable driver behavior, i.e. the suitable speed depends on the individual driver and current conditions, as compared to the attributes that other drivers use to determine their speed. The system- determined safer speed is then the speed that most people drive, because most people presumably did fine using those speeds.

[0052] It will also be evident that the profiles may be weather/lighting dependent, for example the road profile may be a function of weather and lighting. This becomes apparent, and can be compensated for, by comparing the profile derived from sample populations under those various different operating conditions.

[0053] Notwithstanding the above, in accordance with some embodiments the system can take into account additional weather-related, vehicle, driver, or legal speed limits information in supplementing its determination as to what may be a safer speed and/or what constitutes acceptable driver behavior.

[0054] Driver training may also be based upon examination of an individual's speed profile. A more aggressive speed profile may not be appropriate for beginning drivers, whereas a very slow profile may be a warning sign for elderly drivers. In accordance with some embodiments, the system can use a more complex measure than speed alone; for example a combination of speed and acceleration, or any combination of measurements from the environment and the particular vehicle.

Lane Departure Warning System

[0055] In accordance with a particular embodiment, the systems and techniques described above can be used in a particular embodiment that functions as a lane departure warning system, and that allows a driver (or more correctly their vehicle) position on the road to be statistically compared to a stored profile or profiles. Figure 14 shows an illustration of an embodiment for use as a lane departure warning system 430. As shown in Figure 14, profile information, location, vehicle and driver information, and sensor information 434 is continuously retrieved and used to determine vehicle position information and to detect a possible lane intervention 438. This information can be used to adjust 414 the vehicle's movement, and can also be used to increase the accuracy of the profiles stored in the database. [0056] In accordance with an embodiment, a lane departure warning system attempts to warn drivers when they would leave the normal lane of travel. This can be for lane change, or can represent an accidental road departure. The following example assumes a vehicle departing from a single lane road with a single direction of travel. The problem becomes more complicated with additional lanes and directions of travel; however the approach is essentially the same. In accordance with an embodiment, position information is collected from a large number of vehicles as they travel along a particular lane. Once the data has been collected a centerline can be computed using a least-squares technique or other technique. The variance of the position data about the centerline can be computed as a function of distance along the line. This variance derives from three sources: error in the vehicle positioning system, differences in the detailed tracks chosen by individual drivers, and errors in the driver's ability to maintain their chosen track. In accordance with an embodiment the system generally determines the magnitude of the chosen differences in the tracks, or track variation. This can easily be performed by subtracting the two errors from the overall variance. The positioning system error can be estimated from the covariance matrices of the positioning system and is assumed to be zero mean. The driver tracking error can be measured statistically (usually around 25 cm) and is also zero mean. This provides the location of the centerline for increments along the lane, along with the measurement of acceptable deviations from that centerline, and the propensity for a given deviation within the sample population. [0057] In accordance with an embodiment the data can then be stored in an on-board digital map database. This map database is made available in a lane departure warning system in the vehicle, and can be stored in the vehicle, or communicated into the vehicle as it approaches a given area. The vehicle should also have a positioning system with accuracy significantly less than the lane width. The deviation of this particular vehicle is derived by comparing the vehicle location to that of the computed centerline, and the offset is then compared to the statistical deviation of the sample population. A warning can then be given to a driver if their deviation exceeds an appropriate statistical threshold (for example, if 95% of the sample population has a larger deviation then a warning is not given, however if only 5% of the sample population has a larger deviation then a warning is given). If the driver is known to always drive in a '5%' manner, then there would likely be many warnings issued, and there is also a high chance that any warning issued by the system would be ignored by this particular driver. In these instances the system may selectively suppress warnings, set a higher threshold for warnings, and/or take alternate actions such as reporting the driver's consistent poor driving performance.

[0058] Vehicle position alone may not be enough to identify lane departure, and it may also be necessary to include speed or direction of travel in making the assessment. This can be satisfied by requiring those variables to be built into the statistical profile stored in the database. [0059] The above process is advantageous in that it obviates the needs for any judgment on the part of the lane departure warning application as to the position of this vehicle with respect to the lane marking. The determination becomes instead a comparison of a particular behavior to a statistical distribution of a normal population, and the only judgment involves setting an appropriate threshold for an individual. This can be based on previous driving performance (e.g. a data-gathering of a few miles of driving), inference from other driving parameters such as speed, or assigned by a parent or other authority. Another processing step that can be included in the lane departure system is to remove the effects of drivers cutting corners. This can be performed by deriving the vehicle lateral distribution across the lane from straight sections of the road where it is assumed that drivers track fairly well. This is then taken to be the proper distribution as the lane goes through a curve (the implicit assumption being the lane width is the same). In order to get the true curve shape one matches the profile derived from a straight section to the outside portion of the profile in the curve. The presumption is that since there are people cutting the lane, but not everyone does, the profile in the curve will be anomalously wide, however all of the extra bit will be on the inside of the curve and none of it on the outside. Therefore by matching the outside of the curve to the profile it is possible to reconstruct an ideal lane profile.

Environmental-Conscious Driving Guide or Game

[0060] In accordance with a particular embodiment, the systems and techniques described above can be used in a particular embodiment that functions as an environmental- conscious driving guide or game, and that allows a driver operation of the vehicle to be statistically compared to a stored profile or profiles. Figure 15 shows an illustration of an embodiment for use as an environmental-conscious driving guide or game 440. [0061] The eco-driving guide or game is designed to allow drivers to compare their driving habits from the perspective of fuel savings, environmental efficiency, and vehicle performance including reduction of particulates and chemicals. It can be used to compare individual drivers to each other, a particular driver's performance over many trips, or compare a driver to a larger driving population. Fuel consumption data can be collected for a large number of vehicles over a particular segment of road. Segments can be composed of several other segments, and will largely reflect the common driving routes in a particular area. The fuel consumption data from a vehicle driving across a particular segment is weighted according to the EPA -determined efficiency of that vehicle. Thus the weighted fuel consumption for a road segment is the total fuel consumed by the vehicle multiplied by the EPA fuel efficiency rating for that particular vehicle (in essence the EPA rating becomes the 'driver profile' for this application since it is assumed that all drivers want to drive as efficiently as practically possible). This weighting tends to minimize any effects of the vehicle, and instead emphasizes those of the driver. The weighted fuel consumption data for a large number of vehicles over a particular segment of road is combined to create a fuel consumption profile for that segment, and the profile stored in the database. This profile can be a function of time of day, weather, or other factors. In accordance with an embodiment, this profile is stored in a map database in association with the road segment. An individual driver's weighted consumption data 450 is compared 454 with the profile developed for a given segment of road and the driver is then given a score 458 representing his percentile performance over that segment. Scores for individual road segments can be combined to provide overall efficiency scores for the entire range of segments. Scores for an individual driver can also reflect the average of many transits of a particular road segment.

[0062] In accordance with an embodiment the scores can be normalized by vehicle type and average driving patterns and used to compare individuals or groups of individuals driving on various road segments. This can also form the basis of a game wherein users seek a better score which can be compared to other drivers, sets of drivers, and drivers on other road segments. The system can also be used to provide targeted feedback to individual drivers based on whether their scores are high or low, and inform drivers about the implications of their driving behavior (for example on their gas bill, or on the environment) as reflected in their scores; or provide suggestions to them as to which is the most fuel efficient path between two points (for example, by making optimal use of highways, and avoiding traffic congestion or hills).

[0063] It will be evident that the examples of various embodiments that utilize the systems and processes described herein are provided for purposes of illustration, and that other embodiments and uses of the systems and processes can be developed within the spirit and scope of the invention. [0064] Examples of additional various embodiments include but are not limited to:

• Embodiments wherein the profile is stored as attributes in a geographic database or map.

• Embodiments wherein the action is elicited from the vehicle itself, such as a control function (e.g. the vehicle slows down).

• Embodiments wherein the action is elicited from the driver (e.g. A warning that they are going faster than expected.

• Embodiments wherein the action is a recommendation, advice, or score provided after the fact, either automatically or by some human reviewer (parent or supervisor as part of training or performance feedback).

• Embodiments wherein the action is to provide a 'score' for a game allowing drivers to compete.

• Embodiments wherein the observed variance is compared to an expected variance for the vehicle or individual and the action is only elicited based on individualized tolerances (e.g. A driver with a considerable driving history would be allowed tolerate a higher variance before a safety warning is given to that driver, whereas a relatively new or beginning driver may have a lower tolerance to warning and as such would receive more warnings for similar driving scenarios).

Embodiments wherein the data is collected from an individual vehicle over many encounters with the same environment.

Embodiments wherein the statistical profile is based on a single instance of a single vehicle.

Embodiments wherein the statistical profile is developed from an aggregate of many vehicles and times.

Embodiments wherein the profile exists in the vehicle and the profile is performed locally without communications.

Embodiments wherein the profile is computed at a central facility.

Embodiments wherein the profile is modified or weighted according to the configuration of the vehicle (e.g. EPA mileage ratings).

Embodiments wherein the environment consists of a location or geographic region.

Embodiments wherein the environment consists of certain types of terrain.

Embodiments wherein the environment consists of vehicle dynamics.

Embodiments wherein the environmental conditions occur over time (e.g. extended driving up hills or in a dust storm.

Embodiments wherein the activities of the driver include their chosen speed.

Embodiments wherein the activities of the driver include their chosen lateral position on the road.

Embodiments wherein the activities of the driver include their chosen acceleration and deceleration.

Embodiments wherein the activities of the driver include their chosen use of vehicle accessories such as air conditioning, or lights.

Embodiments wherein the activities of the driver include their use of in vehicle devices such as phones, PNDs and PDAs.

Embodiments wherein the feedback is provided based on a statistical profile derived from methodically collected data rather than probe data.

Embodiments wherein the road profiles are weather or environmentally dependent. • Embodiments wherein the comparisons of the driver behavior to profiles determines a level of impairment for a driver (eg alcohol or sleep).

[0065] Embodiments of the present invention include a computer program product which is a storage medium (media) having instructions stored thereon/in which can be used to program a computer to perform any of the processes of the present invention. The storage medium can include, but is not limited to, any type of disk including floppy disks, optical discs, DVD, CD-ROMs, microdrive, and magneto-optical disks, ROMs, RAMs, EPROMs, EEPROMs, DRAMs, VRAMs, flash memory devices, magnetic or optical cards, nanosystems (including molecular memory ICs), or any type of media or device suitable for storing instructions and/or data.

[0066] Stored on any one of the computer readable medium (media), embodiments of the present invention include software for controlling both the hardware of the general purpose/specialized computer or microprocessor, and for enabling the computer or microprocessor to interact with a human user or other mechanism utilizing the results of the present invention. Such software may include, but is not limited to, device drivers, operating systems, and user applications. Ultimately, such computer readable media further includes software for performing the present invention, as described above.

[0067] Appropriate software coding can readily be prepared by skilled programmers based on the teachings of the present disclosure, as will be apparent to those skilled in the software art. The invention may also be implemented by the preparation of application specific integrated circuits or by interconnecting an appropriate network of conventional component circuits, as will be readily apparent to those skilled in the art.

[0068] The foregoing description of the present invention has been provided for the purposes of illustration and description. It is not intended to be exhaustive or to limit the invention to the precise forms disclosed. Many modifications and variations will be apparent to the practitioner skilled in the art. The embodiments were chosen and described in order to best explain the principles of the invention and its practical application, thereby enabling others skilled in the art to understand the invention for various embodiments and with various modifications that are suited to the particular use contemplated. It is intended that the scope of the invention be defined by the following claims and their equivalence.

Claims

Claims:What is claimed is:

1. A system for providing driver information or improving driving performance based on statistical feedback, comprising: a database and a plurality of statistical profiles stored therein, wherein the profiles describes the actions of a plurality of drivers, vehicles, and/or trips in one or more particular locations and environments; a vehicle profiling and feedback system for use with a moving vehicle that determines information about the vehicle's current location, environment and operating conditions; a vehicle interface that allows the vehicle to communicate and receive information to match the vehicle's current location with profile information; and wherein the system uses the information about the vehicle's current location, environment and operating conditions together with the profile information for the current location to determine additional information about the vehicle and/or the driver when compared to the database of profiles, and then either communicate information to the driver, make actions to the vehicle, or store in a storage for analysis and use in providing driver feedback information or improving driver performance.

2. The system of claim 1 wherein the profile is stored as attributes in a geographic database or map.

3. The system of claim 1 wherein the action is elicited by the vehicle itself, such as a control function or automatic maneuver.

4. The system of claim 1 wherein the action is elicited by the driver in response to a signal from the vehicle.

5. The system of claim 1 wherein the action is a recommendation or advice provided after the fact, either automatically or by a human reviewer, including an optional score.

6. The system of claim 1 wherein the action includes providing a score for a game allowing drivers to compete.

7. The system of claim 1 wherein the observed variance is compared to an expected variance for the vehicle or individual, and the action is only elicited when the behavior exceeds an individualized threshold that reflects the driving history of the individual driver, or their tolerance for receiving warnings.

8. The system of claim 1 wherein the data is collected from an individual vehicle over many encounters with the same environment.

9. The system of claim 1 wherein the statistical profile is based on a single driver and a single vehicle over many trips.

10. The system of claim 1 wherein the statistical profile is developed from an aggregate of many vehicles over a common location.

11. The system of claim 1 wherein the profile is stored in the vehicle and the profile matching is performed locally without external communications.

12. The system of claim 1 wherein the profile is stored and/or computed at a central facility and communicated to and from the vehicle.

13. The system of claim 1 wherein the profile is modified or weighted according to the configuration of the vehicle.

14. The system of claim 1 wherein the environment consists of a location or geographic region.

15. The system of claim 1 wherein the environment consists of certain types of terrain.

16. The system of claim 1 wherein the environment consists of vehicle dynamics.

17. The system of claim 1 wherein the environmental conditions occur over time.

18. The system of claim 1 wherein the activities of the driver include their chosen speed.

19. The system of claim 1 wherein the activities of the driver include their chosen lateral position on the road.

20. The system of claim 1 wherein the activities of the driver include their chosen acceleration and deceleration.

21. The system of claim 1 wherein the activities of the driver include their chosen use of vehicle accessories such as air conditioning, or lights.

22. The system of claim 1 wherein the activities of the driver include their use of in vehicle devices such as phones, PNDs and PDAs.

23. The system of claim 1 wherein the feedback is provided based on a statistical profile derived from methodically collected data rather than probe vehicle data.

24. The system of claim 1 wherein the system is used in a curve warning system, and uses information about the vehicles current speed, and optionally a driver profile, to match against road profiles for regions approaching a curve in the road, and to determine if the current speed exceeds the statistical variance for the curve approach, and if so issuing a warning of excessive speed, subject to any further information specified by the driver profile that may limit or suppress such warning.

25. The system of claim 1 wherein the system is used in a lane departure warning system , and uses information about the vehicles current road position, and optionally a driver profile, to match against road profiles for that road regions, and to determine if the current road position speed exceeds the statistical variance for road position, and if so issuing a warning of potential lane departure, subject to any further information specified by the driver profile that may limit or suppress such warning.