USOO943 0944B2
`
`(12) United States Patent
`Grimm et al.
`
`(10) Patent No.:
`(45) Date of Patent:
`
`US 9.430,944 B2
`Aug. 30, 2016
`
`(54) METHOD AND APPARATUS FOR
`DETERMINING TRAFFIC SAFETY EVENTS
`USING VEHICULAR PARTICIPATIVE
`SENSING SYSTEMS
`
`(71) Applicant: GM GLOBAL TECHNOLOGY
`OPERATIONS LLC, Detroit, MI (US)
`(72) Inventors: Donald K. Grimm, Utica, MI (US);
`Fan Bai, Ann Arbor, MI (US);
`Rozalina Ebrahimian, Ann Arbor, MI
`(US)
`(73) Assignee: GM Global Technology Operations
`LLC, Detroit, MI (US)
`Subject to any disclaimer, the term of this
`patent is extended or adjusted under 35
`U.S.C. 154(b) by 0 days.
`(21) Appl. No.: 14/539,764
`
`(*) Notice:
`
`(22) Filed:
`
`Nov. 12, 2014
`
`(65)
`
`Prior Publication Data
`US 2016/O13313O A1
`May 12, 2016
`
`(2006.01)
`(2006.01)
`(2009.01)
`(2006.01)
`(2009.01)
`(2006.01)
`(2006.01)
`
`(51) Int. Cl.
`G08G I/09
`G08G I/0967
`H0474/04
`H04L 29/08
`H0474/00
`G08G I/OI
`G08G I/27
`(52) U.S. Cl.
`CPC ...... G08G I/096766 (2013.01); G08G I/0104
`(2013.01); G08G I/091 (2013.01); G08G
`I/096716 (2013.01); G08G I/096741
`(2013.01); G08G I/096775 (2013.01); G08G
`1/127 (2013.01); H04L 67/12 (2013.01);
`H04W 4/008 (2013.01); H04 W4/046
`(2013.01)
`
`(58) Field of Classification Search
`CPC ... G08G 1/091; G08G 1/127; G08G 1/0104;
`G08G 1/096715; G08G 1/096741; G08G
`1/O96775
`USPC .................................. 340/902, 905: 701/117
`See application file for complete search history.
`
`(56)
`
`References Cited
`
`U.S. PATENT DOCUMENTS
`
`5,812,069 A * 9/1998 Albrecht et al. .............. 340,905
`7,825,824 B2 * 1 1/2010 Shrum, Jr. .................... 340,902
`8.498,775 B2
`7/2013 Yngve et al.
`2011/0012753 A1* 1/2011 Shrum, Jr. .................... 340,902
`
`FOREIGN PATENT DOCUMENTS
`
`1, 2010
`
`GB
`2461551 A
`* cited by examiner
`Primary Examiner — John A Tweel, Jr.
`(74) Attorney, Agent, or Firm — John A. Miller; Miller IP
`Group, PLC
`
`ABSTRACT
`(57)
`Methods and systems are disclosed for participative sensing
`of events and conditions by road vehicles, collection of this
`data from a large number of road vehicles by a central server,
`processing the data to identify events and conditions which
`may be of interest to other vehicles in a particular location,
`and sending notifications of the events and conditions to
`vehicles. A large number of vehicles use participative sens
`ing systems to identify a safety-related event or condition
`which should be reported to the central server—such as a
`large pothole, an obstacle in the roadway, an icy road
`Surface, a traffic accident, etc. The central server stores and
`aggregates the data, filters it and ages it. Vehicles requesting
`advisories from the central server will receive notices of
`safety-related events and conditions based on their location
`and heading. Driver warnings can be issued, and vehicle
`systems may respond to the notices.
`
`21 Claims, 9 Drawing Sheets
`
`
`
`IPR2025-00943
`Tesla EX1031 Page 1
`
`

`

`U.S. Patent
`
`Aug. 30, 2016
`
`Sheet 1 of 9
`
`US 9.430,944 B2
`
`
`
`ru &
`
`IPR2025-00943
`Tesla EX1031 Page 2
`
`

`

`U.S. Patent
`
`Aug. 30, 2016
`
`Sheet 2 of 9
`
`US 9.430,944 B2
`
`
`
`IPR2025-00943
`Tesla EX1031 Page 3
`
`

`

`U.S. Patent
`
`Aug. 30, 2016
`
`Sheet 3 of 9
`
`US 9.430,944 B2
`
`
`
`CY)
`
`Y
`
`D CD
`
`-
`
`N
`
`8
`N
`
`IPR2025-00943
`Tesla EX1031 Page 4
`
`

`

`U.S. Patent
`
`Aug. 30, 2016
`
`Sheet 4 of 9
`
`US 9.430,944 B2
`
`
`
`S.
`
`IPR2025-00943
`Tesla EX1031 Page 5
`
`

`

`U.S. Patent
`
`Aug. 30, 2016
`
`Sheet 5 Of 9
`
`US 9.430,944 B2
`
`
`
`IPR2025-00943
`Tesla EX1031 Page 6
`
`

`

`U.S. Patent
`
`Aug. 30, 2016
`
`Sheet 6 of 9
`
`US 9.430,944 B2
`
`
`
`
`
`
`
`IPR2025-00943
`Tesla EX1031 Page 7
`
`

`

`U.S. Patent
`
`Aug. 30, 2016
`
`Sheet 7 Of 9
`
`US 9.430,944 B2
`
`
`
`IPR2025-00943
`Tesla EX1031 Page 8
`
`

`

`U.S. Patent
`
`Aug. 30, 2016
`
`Sheet 8 of 9
`
`US 9.430,944 B2
`
`s
`
`
`
`CO
`
`Y
`D
`CD
`L
`
`IPR2025-00943
`Tesla EX1031 Page 9
`
`

`

`U.S. Patent
`
`Aug. 30, 2016
`
`Sheet 9 Of 9
`
`US 9.430,944 B2
`
`
`
`3.
`
`IPR2025-00943
`Tesla EX1031 Page 10
`
`

`

`US 9,430,944 B2
`
`1.
`METHOD AND APPARATUS FOR
`DETERMINING TRAFFIC SAFETY EVENTS
`USING VEHICULAR PARTICIPATIVE
`SENSING SYSTEMS
`
`2
`Additional features of the present invention will become
`apparent from the following description and appended
`claims, taken in conjunction with the accompanying draw
`ings.
`
`BACKGROUND OF THE INVENTION
`
`BRIEF DESCRIPTION OF THE DRAWINGS
`
`10
`
`15
`
`25
`
`30
`
`1. Field of the Invention
`This invention relates generally to using crowd-sourced
`data from vehicles to determine traffic conditions and events
`and, more particularly, to a method and apparatus for deter
`mining traffic events using vehicular participative sensing
`systems, where data from multiple vehicles can be collected
`and analyzed on a central server and used to detect or infer
`various types of traffic safety-related conditions and events,
`including specific conditions and events detected in real time
`and chronic conditions that tend to recur regularly, and
`advisories of the traffic safety-related conditions and events
`are communicated to vehicles on the road.
`2. Discussion of the Related Art
`Many vehicles now include systems which can sense a
`wide range of parameters related to the vehicle's operating
`environment. For example, vehicle dynamics sensors can
`define a vehicle's dynamic state, object detection systems
`can detect objects and other vehicles on and around a
`roadway, the status of a vehicle's systems such as braking,
`steering, ABS and airbags is available, and traffic and road
`conditions can be determined by a variety of methods. Most
`of this data is evaluated and used by the host vehicle in real
`time, and discarded when it becomes Stale.
`At the same time, telematics systems are also available
`onboard many modern vehicles, where the telematics sys
`tems continuously or regularly communicate data from the
`vehicle to a centralized database system, which also com
`municates information back to vehicles. Although these
`35
`telematics systems have been used to gather some limited
`types of vehicle data for specific purposes—such as detect
`ing airbag deployment in a vehicle and automatically
`requesting emergency services—much more data could be
`collected from a large number of vehicles, and this data
`could be used to identify a wide range of traffic and road
`conditions which can be disseminated to and beneficial to
`other vehicles in a certain geographic locale.
`
`40
`
`SUMMARY OF THE INVENTION
`
`45
`
`In accordance with the teachings of the present invention,
`methods and systems are disclosed for participative sensing
`of events and conditions by road vehicles, collection of data
`regarding the events and conditions from a large number of
`road vehicles by a central server, processing the data to
`identify events and conditions which may be of interest to
`other vehicles in a particular location, and sending notifi
`cations of the events and conditions to vehicles as appro
`priate. A large number of vehicles use participative sensing
`systems to identify a safety-related event or condition which
`should be reported to the central server—Such as a large
`pothole which has been encountered, an obstacle in the
`roadway, an icy road Surface, a traffic accident, etc. The
`central server stores and aggregates the data, filters it and
`ages it. Vehicles requesting advisories from the central
`server, typically via a telematics system—will receive
`notices of safety-related events and conditions which may be
`significant based on their location and heading. Driver
`warnings can be issued, and vehicle systems can be adapted
`(e.g., Suspension tuning or transmission mode can be
`changed) in response to the notices.
`
`50
`
`55
`
`60
`
`65
`
`FIG. 1 is a schematic illustration of a vehicle with
`participative sensing systems and a telematics system for
`communicating data to a collection system;
`FIG. 2 is an illustration of several vehicles on a roadway,
`where some vehicles can provide road event data to a central
`server and the server can communicate alerts out to other
`vehicles which are approaching the event location;
`FIG.3 is a block diagram showing data flow in the central
`server and out to vehicles and other interested parties:
`FIG. 4 is a combined block diagram and flowchart dia
`gram showing a method used by a participative sensing
`vehicle, data flow to and processing in a cloud-based system,
`and a method used by a vehicle requesting advisories;
`FIG. 5 is a block diagram of a road surface condition
`classifier which can be used in a vehicle to determine road
`Surface friction conditions;
`FIG. 6 is a flowchart diagram showing a method for
`calculating an estimated coefficient of friction for a vehicle
`based on vehicle dynamic conditions;
`FIG. 7 is an illustration of a scenario with several par
`ticipative sensing vehicles providing road friction data to a
`central server, and the server communicating friction esti
`mations back out to the vehicles;
`FIG. 8 is a block diagram showing data flow in the central
`server of FIG. 7 and out to vehicles and other interested
`parties; and
`FIG. 9 is a combined block diagram and flowchart dia
`gram showing a friction estimation method used by a
`participative sensing vehicle, data flow to and processing in
`a cloud-based system, and a method used by a vehicle
`requesting road friction advisories.
`
`DETAILED DESCRIPTION OF THE
`EMBODIMENTS
`
`The following discussion of the embodiments of the
`invention directed to a method and apparatus for determin
`ing traffic safety events using vehicular participative sensing
`systems is merely exemplary in nature, and is in no way
`intended to limit the invention or its applications or uses.
`Many vehicles are now equipped with a wide range of
`sensors and systems which can provide data which is
`indicative of the conditions the vehicle is operating in and
`events which may have occurred in the vicinity of the
`vehicle. By collecting such data from a large number of
`vehicles and aggregating it to detect trends, a significant
`amount of information can be deduced which would be
`useful to and can be communicated to other vehicles in
`the vicinity.
`FIG. 1 is a schematic illustration of a vehicle 10 with
`participative sensing systems and telematics system capa
`bility for communicating data to a collection system. The
`vehicle 10 includes a vehicle dynamics module 20 for
`determining vehicle dynamic conditions and other related
`parameters. The vehicle dynamics module 20 receives data
`from at least one sensor 22. Typically, many of the sensors
`22 would be provided, including wheel speed sensors,
`longitudinal, lateral and vertical acceleration sensors, and a
`yaw rate sensor. The sensors 22 may also include wheel load
`sensors and other types of sensors. The vehicle dynamics
`
`IPR2025-00943
`Tesla EX1031 Page 11
`
`

`

`US 9,430,944 B2
`
`10
`
`15
`
`3
`module 20 collects data from all of the sensors 22 and
`performs calculations as necessary to provide a complete
`representation of the dynamic conditions of the vehicle
`10—including positions, Velocities, accelerations and forces
`affecting the vehicle 10.
`The vehicle 10 also includes an object detection module
`30. The object detection module 30 receives data from at
`least one object detection sensor 32 which could be a
`camera-based sensor or may use radar, lidar or some other
`type of object detection technology (including short range
`communications technologies such as Dedicated Short
`Range Communications DSRC or Ultra-Wide Band
`UWB). More than one of the object detection sensors 32
`may be provided, including forward view, rear view and side
`view sensors. Using data from the sensors 32, the object
`detection module 30 identifies objects in the vicinity of the
`vehicle 10, where the objects may include other vehicles,
`curbs and other roadway boundaries, pedestrians and any
`sort of objects that may be on or near the roadway. The
`object detection module 30 can distinguish between regular
`size cars and light trucks and other, larger vehicles Such as
`delivery trucks and semi-trailer trucks. The object detection
`module 30 can also determine the velocity of other vehicles
`on the roadway, and identify situations where vehicles are
`stopped that should ordinarily be moving (such as on a
`highway). In addition, the object detection module 30 can
`identify lane boundary markings and compute the position
`of the vehicle 10 relative to the lane or lanes on the roadway.
`The vehicle 10 also includes a system status module 40
`which collects data from a vehicle data communications bus
`regarding the status of virtually any vehicle system. For
`example, the system status module can determine conditions
`Such as; windshield wiperson, off or intermittent; headlights
`on or off throttle position; brake pressure; anti-lock brake
`system (ABS) activation; traction control system (TCS)
`activation; airbag deployment; seat occupancy; steering
`wheel position; ambient temperature; infotainment system
`usage including in-vehicle cell phone usage; HVAC system
`settings, etc. The data collected by the system status module
`40 can be used to identify many different types of driving
`situations and conditions, as will be discussed at length
`below.
`The vehicle 10 also includes a vehicle-to-vehicle (V2V)
`communications module 50, which communicates with
`other, similarly-equipped vehicle within communications
`range, using Dedicated Short Range Communications
`(DSRC) or other communications technology. The V2V
`communications module 50 can collect significant amounts
`of data from nearby vehicles, particularly including position,
`Velocity and acceleration data—as is needed for 'smart
`highway' or autonomous vehicle systems.
`Data from the vehicle dynamics module 20, the object
`detection module 30, the system status module 40 and the
`V2V communications module 50 are provided to a data
`collection module 60. The data collection module 60 is in
`communication with a telematics system 70, which commu
`nicates with a telematics central service via cellular com
`munication towers 80 or other technologies. The other
`communications technologies may include, but are not lim
`60
`ited to, DSRC or other vehicle-to-infrastructure (V2I) com
`munications, Wi-Fi, Satellite communications, etc.
`It is to be understood that the vehicle dynamics module
`20, the object detection module 30, the system status module
`40, the V2V communications module 50 and the data
`collection module 60 are comprised of at least a processor
`and a memory module, where the processors are configured
`
`45
`
`4
`with Software designed to perform data collection and
`computations as discussed above.
`It is to be further understood that the features and calcu
`lations of the vehicle dynamics module 20, the object
`detection module 30, the system status module 40, the V2V
`communications module 50 and the data collection module
`60 could be allocated differently than described herein
`without departing from the spirit of the disclosed invention.
`For example, although the functions of the modules 20-60
`are described as being distinct throughout this disclosure,
`they could in fact all be programmed on the same processor,
`or more or fewer than the five distinct modules shown.
`FIG. 2 is an illustration of a scenario 100 with several
`vehicles on a roadway 102, where some vehicles can provide
`road event data to a central server and the server can
`communicate advisories out to other vehicles which are
`approaching the event location. The scenario 100 includes
`vehicles 110-150, driving on the 2-lane road 102, where the
`vehicles 110, 120 and 130 are driving in one direction, and
`the vehicles 140 and 150 are driving in the other direction.
`An event location 160 is indicated with the dashed box,
`where the event location 160 could be a bad pothole, a patch
`of slippery road, a tree or other object on the road Surface,
`or any of a variety of other conditions. The vehicles 120 and
`130 have already passed through the event location 160, and
`have collected data indicative of the event or condition. For
`example, a large pothole could be detected by a wheel load
`spike in one vehicle and an evasive steering maneuver in
`another vehicle, a slippery road surface could be detected by
`traction control system and/or anti-lock braking system
`activation, and an object on the road Surface could be
`detected by the object detection module 30.
`The vehicles 120 and 130 communicate data regarding the
`event location 160 to a central server 170. The central server
`170 is shown as a cloud-based device, meaning that it could
`be one or more servers existing anywhere on a globally
`connected network. The central server 170 may be part of a
`telematics service, such as the service which is used by the
`telematics system 70 of the vehicle 10. The central server
`170 may instead be operated by any business or government
`entity that can collect and disseminate data from a large
`number of vehicles with participative sensing systems.
`In the case of a tree or other obstacle on the roadway 102,
`as an example, the vehicles 120 and 130 would both have
`detected the large, static object in an unexpected location on
`the road surface. The vehicles 120 and 130 may also have
`performed braking and/or steering maneuvers in response to
`the presence of the obstacle. This data is communicated to
`the central server 170, in the manner discussed relative to the
`vehicle 10 of FIG.1. In some instances, even a single vehicle
`reporting an event or condition may be compelling enough
`for the central server 170 to issue advisories out to other
`vehicles. But there is power in large numbers, and the server
`170 can determine the existence of more conditions, and
`with greater accuracy, by aggregating data from many
`vehicles.
`In the scenario 100, based on the report of an obstacle on
`the road by the vehicles 120 and 130, the server 170 issues
`advisories to the vehicle 110, which is going to encounter the
`condition imminently. The vehicle 110 can take action in a
`number of different ways in response to the information it
`receives, including issuing an alert to the driver, terminating
`cruise control if it is activated, slowing down the vehicle 110
`by applying the brakes, taking evasive steering action, and
`re-focusing object detection sensors onboard the vehicle 110
`to attempt to locate the obstacle. Similar types of actions,
`and others (e.g., modifying the navigation route, adapting
`
`25
`
`30
`
`35
`
`40
`
`50
`
`55
`
`65
`
`IPR2025-00943
`Tesla EX1031 Page 12
`
`

`

`US 9,430,944 B2
`
`10
`
`15
`
`25
`
`30
`
`35
`
`40
`
`5
`the safety system warning timings), could be taken by the
`vehicle 110 in the event of other types of reports—such as
`potholes, slippery road Surface, traffic accident, etc.—by the
`vehicles 120 and 130 via the central server 170.
`The server 170 also issues advisories of the obstacle in the
`roadway 102 to the vehicles 140 and 150, which are headed
`in the direction of the event location 160. Although the
`vehicles 140 and 150 are travelling in the opposite direction
`and lane of travel from the vehicles 120 and 130 which
`reported the obstacle, it is apparent from FIG. 2 that they
`may benefit from the advisory. Many factors can be consid
`ered by the server 170 in determining to which vehicles
`advisories should be issued—including the nature of the
`reported event or condition, the specific location of the event
`or condition on the road surface (center of lane, left shoulder,
`etc.), whether the roadway 102 is divided and how many
`lanes of travel are available in each direction, etc. These
`factors will be discussed further below.
`The three types of conditions (pothole, slippery road,
`obstacle) described above and shown in FIG. 2 are merely
`exemplary; many other types of safety-related roadway and
`vehicle conditions may be detected by vehicles and com
`municated to the central server 170. Other conditions which
`may be reported by the vehicles 120/130 include one or
`more vehicles exceeding the speed limit by a significant
`amount, vehicles travelling significantly slower than the
`speed limit, rain, Snow or fog, any significant or unusual
`usage of vehicle controls such as steering, throttle or brakes,
`airbag deployment, etc. Furthermore—although the scenario
`100 is described in terms of the vehicles 120 and 130
`communicating data to the server 170 and the vehicles 110.
`140 and 150 receiving data from the server 170 in reality,
`all of the vehicles 110-150 would be in continuous 2-way
`communications with the server 170.
`As can be understood from the above discussion, each of
`the vehicles 110-150 will continuously gather data from
`onboard systems—such as the vehicle dynamics module 20,
`the object detection module 30, the system status module 40
`and the V2V communications module 50 of the vehicle 10.
`However, it is not likely to be practical for every vehicle to
`communicate all of this raw data to the server 170. Rather,
`in one preferred embodiment, each of the participative
`sensing system vehicles 110-150 performs calculations
`locally to determine what threats—or hazardous events or
`conditions—exist which warrant sending a report to the
`Server 170.
`The calculations by each participative sensing system
`vehicle, which may be performed on the data collection
`module 60, may include several parts. For example, an
`obstacle or object on the roadway greater than a certain
`predetermined size, detected via object detection, may trig
`ger an immediate report to the server 170. Likewise, a traffic
`accident, or a pothole strike resulting in a wheel load greater
`than a certain threshold, may also trigger an immediate
`report to the server 170.
`However, other types of threats or hazardous conditions
`may only be determined by evaluating multiple parameters.
`An example of this would be determining that a particular
`Surrounding vehicle is driving in a dangerous manner. A
`threat level TL of a particular vehicle i may be calculated as:
`60
`(1)
`Where w, is a weighting value associated with a specific
`property j, and p, is the property (such as braking, accel
`eration, or speed) for the vehicle i. The property p, is in turn
`calculated as:
`
`45
`
`50
`
`55
`
`65
`
`(2)
`
`6
`Where x,", x,,..., x," are vehicle parameters obtained from
`raw serial data from the vehicle dynamics module 20, the
`object detection module 30, the system status module 40 and
`the V2V communications module 50.
`Using the above calculations, a threat level TL, for a
`vehicle i can be computed. In an exemplary embodiment, a
`“watch” report could be issued to the server 170 if the threat
`level exceeds a first threshold. A watch report would be
`indicative of a vehicle exhibiting moderately hazardous
`driving behavior, which could be followed at the server 170
`to see if other corroborating reports are received. A “warn
`ing report could be issued to the server 170 if the threat
`level of a vehicle exceeds a second, higher threshold. A
`warning report would be indicative of a vehicle exhibiting
`severely dangerous driving behavior, which could trigger the
`server 170 to immediately issue advisories out to surround
`ing vehicles.
`The above example describes calculating a threat level of
`a surrounding vehicle based on several different properties
`of the particular Surrounding vehicle. A similar approach
`could be used to calculate a threat level of a location on a
`roadway, based on properties of multiple vehicles (such as
`how many vehicles are braking unexpectedly on a freeway).
`The participative sensing system vehicle 120 (as an
`example) can thus send hazardous condition reports to the
`server 170 based on individual or cumulative data about
`itself (Such as a pothole Strike or a loss of traction), based on
`calculations focused on another vehicle (such as dangerous
`driving behavior), or based on calculations focused on a
`location on a roadway (Such as a traffic slowdown).
`FIG. 3 is a block diagram 200 showing data flow in the
`central server 170 and out to vehicles and other interested
`parties. At box 210, data is collected from many participa
`tive sensing vehicles, such as the vehicle 10 of FIG. 1 and
`the vehicles 110-150 of FIG. 2. Although a single vehicle
`report of an event Such as a traffic accident can be Sufi
`ciently definitive to result in advisories being issued to other
`vehicles in the vicinity, the real power of the disclosed
`methods lies in continuous data collection from large num
`bers of vehicles. For example, a single vehicle driver tapping
`the brakes on a freeway would not, in and of itself, be
`noteworthy. But if many vehicles report a brake tap at a
`certain location on a freeway, it is likely indicative of a
`developing heavy traffic condition, which could quickly
`degenerate into a stop-and-go traffic situation. If reduced
`visibility or wet/icy road conditions also exist in the heavy
`traffic area, and if vehicle speeds are still high, advisories of
`the braking activity ahead may well be warranted for
`vehicles approaching the braking Zone. This is just one
`example of how data from many vehicles can be used to
`identify conditions that could not be deduced from a single
`vehicle or a small number of vehicles.
`The server 170 will continuously receive data from many
`thousands, or millions, of vehicles. Therefore, methods must
`be employed to analyze the data to detect or infer various
`types of potential hazardous driving conditions, and deter
`mine to whom the hazardous driving conditions should be
`communicated. One technique for doing this is to segregate
`the potential hazardous driving situations into three types;
`those that relate to the behavior of specific other vehicles and
`their drivers, those that relate to chronic conditions that
`occur at a particular fixed location on a roadway, and those
`that relate to transient conditions at various locations on the
`roadway.
`At box 220, hazardous driving situations related to the
`behavior of specific other vehicles and their drivers are
`identified from the data collected at the box 210. As dis
`
`IPR2025-00943
`Tesla EX1031 Page 13
`
`

`

`US 9,430,944 B2
`
`10
`
`15
`
`25
`
`30
`
`7
`cussed previously, vehicles such as the vehicle 10 of FIG. 1
`can collect and report to the central server 170 a significant
`amount of data about other vehicles in their immediate
`vicinity—with this data being collected at least by the object
`detection module 30 and the V2V communications module
`50. Analysis of velocity and acceleration data from other
`vehicles, in particular, can reveal potential driving threats
`Such as dangerous driving behavior, distracted driving,
`intoxicated or impaired driving, etc. These behaviors can be
`detected by vehicle speeds significantly higher than the
`speed limit, speed significantly lower than the speed limit
`where not caused by heavy traffic, accelerations above a
`threshold (such as 0.1 g) and/or hard braking events above
`a threshold (such as 0.3 g), especially where the acceleration
`or braking events occur repeatedly, extreme tailgating, hard
`steering activity above a threshold (such as 10 deg? sec),
`wandering off-center in lane and/or partially across lane
`boundaries, etc. The idea that a specific dangerous driver/
`vehicle (or dangerous types of crowd behavior) can be
`identified, and other vehicles in the vicinity warned of the
`hazard, is extremely powerful. Of course, the location of the
`dangerous driver/vehicle is constantly changing, and the
`anticipated location can be taken into account when issuing
`hazard warnings to other vehicles. For example, a hazard
`warning could be issued for "dangerous driver may be
`encountered at next intersection approaching from right'.
`Furthermore, the identification of the dangerous driver/
`vehicle or dangerous driving area is made much more robust
`by aggregating participative sensing data from many
`vehicles on the roadway.
`At box 222, hazardous driving situations related to
`chronic conditions that occur at a particular fixed location on
`a roadway are identified from the data collected at the box
`210. These chronic or static conditions are the types of
`things that occur repeatedly and regularly—such as traffic
`35
`jams at a particular intersection or stretch of freeway at rush
`hour on weekdays. These chronic conditions may be caused
`by poor roadway designs such as complex merges or overly
`tight road curvature, poorly timed traffic signals, road con
`struction, or simply roads or intersections that can’t handle
`the traffic volume due to insufficient lanes or other factors.
`Chronic conditions can easily be identified at the box 222 by
`monitoring data from many vehicles over a period of days or
`months and detecting densely packed traffic traveling at
`speeds significantly below the posted speed limit. Similarly,
`locations where an inordinate number of traffic accidents
`occur can be identified. When these conditions are detected
`regularly, a chronic hazardous traffic location has been
`identified. Notification of the chronic hazardous traffic loca
`tion can be provided to approaching vehicles, and also to
`whatever governmental Transportation Department or Road
`Commission has responsibility for the roadway in question.
`At box 224, hazardous driving situations related to tran
`sient conditions at various locations on the roadway are
`identified from the data collected at the box 210. The
`transient conditions which are identified at the box 224 are
`temporary in nature, unlike the chronic conditions identified
`at the box 222. Transient hazardous driving conditions may
`be caused by weather conditions, a traffic accident, poor road
`condition, a traffic signal outage or other event, and may
`include poor visibility, wet or icy road surface, pothole or
`debris on the road, accident vehicles and/or emergency
`vehicles on the road or the shoulder, etc. These conditions
`may be identified by many different types of data provided
`by the participative sensing system vehicles—including low
`vehicle speeds, object detection data (stopped vehicles or
`other objects where they don’t belong on the roadway),
`
`55
`
`8
`wheel load data indicative of a pothole, anti-lock brake or
`traction control system activations indicative of a slippery
`road Surface, and others. Hazard warnings such as “severe
`pothole ahead in right lane' or “disabled vehicle ahead on
`left shoulder” can be issued based on the data identified at
`the box 224.
`At box 230, data fusion of the safety metrics from the
`boxes 220/222/224 is performed. The fusion of the safety
`metrics combines the three types of hazardous driving
`conditions described above-along with their associated
`communications parameters—into a single database for dis
`semination. The fusion also identifies correlations between
`the three types of hazardous driving conditions—such as
`traffic accidents from the box 224 and chronic rush hour
`congestion at the box 222.
`The data at the boxes 210-230 will preferably have one or
`more decay function applied to it. For example, the raw
`event data from individual vehicles at the box 210 may have
`certain rules for half-life and eventual purging, where each
`individual event report may carry full weight for a prede
`termined amount of time, and then decay in weight factor
`after that. Similarly, the hazardous conditions which are
`determined at the boxes 220/222/224 (and fused at the box
`230) may have different decay functions, where dangerous
`driver conditions detected at the box 220 may decay very
`quickly, chronic conditions detected at the box 22 may decay
`very slowly, and transient conditions detected at the box 224
`may decay at an intermediate rate.
`Finally, the data which has been collected, aggregated and
`analyzed by the server 170 results in advisories which can be
`issued to vehicles such as the vehicles 110-150. These
`advisories take two general forms.
`At box 240, advisories are issued in what can be referred
`to as “relaxed real time. Whereas “real time' would imply
`advisories being issued within milliseconds of occurrence of
`an event, relaxed real time refers to advisories being issued
`generally within a matter of seconds to the vehicles which
`can benefit from the information. This is not to imply that
`advisories cannot be issued in real time by the server 170. A
`real time advisory may be issued, for example, in a situation
`where vehicles are travelling at a high speed and an accident
`has just occurred immediately ahead. On the other hand,
`relaxed real time advisories may be issued in many instances
`where warranted by traffic conditions or road conditions
`a