The system 200 also may include a driving analysis server 220, containing some or all of the hardware/software components as the computing device 101 depicted in FIG. 1. The driving analysis server 220 may include hardware, software, and network components to receive vehicle operation data from the vehicle operation computer system 225 and/or directly from a plurality of vehicles 210. The driving analysis server 220 and the vehicle operation computer system 225 may be implemented as a single server/system, or may be separate servers/systems. In some examples, the driving analysis server 220 may be a central server configured to receive vehicle operation data from a plurality of remotely located vehicle operation computer systems 225.

As shown in FIG. 2, driving analysis server 220 may include a driving analysis module 221 and a driver score calculation module 222. Modules 221 and 222 may be implemented in hardware and/or software configured to perform a set of specific functions within the driving analysis server 220. For example, the driving analysis module 221 and the driver score calculation module 222 may include one or more driving event analysis/driver score calculation algorithms, which may be executed by one or more software applications running on generic or specialized hardware within the driving analysis server 220. The driving analysis module 221 may use the vehicle operation data received from the vehicle operation computer system 225 and/or additional image data, video data, and/or object proximity data perform driving event analyses for vehicles 210. The driver score calculation module 222 may use the results of the driving event analysis performed by module 221 to calculate or adjust a driver score for a driver of a vehicle 210 based on specific driving events. Further descriptions and examples of the algorithms, functions, and analyses that may be executed by the driving analysis module 221 and the driver score calculation module 222 are described below in reference to FIGS. 3-5.

To perform driving event analyses and driver score calculations, the driving analysis server 220 may initiate communication with and/or retrieve data from one or more vehicles 210, vehicle operation computer systems 225, and additional computer systems 231-233 storing data that may be relevant to the driving event analyses and driver score calculations. For example, one or more traffic data storage systems 231, such as traffic databases, may store data corresponding to the amount of traffic and certain traffic characteristics (e.g., amount of traffic, average driving speed, traffic speed distribution, and numbers and types of accidents, etc.) at various specific locations and times. Traffic data storage systems 231 also may store image and video data recorded by traffic cameras various specific locations and times. One or more weather data storage systems 232, such as weather databases, may store weather data (e.g., rain, snow, sleet, hail, temperature, wind, road conditions, visibility, etc.) at different locations and different times. One or more additional driving databases/systems 233 may store additional driving data from one or more different data sources or providers which may be relevant to the driving event analyses and/or driver score calculations performed by the driving analysis server 220. Additional driving databases/systems 233 may store data regarding events such as road hazards and traffic accidents, downed trees, power outages, road construction zones, school zones, and natural disasters that may affect the driving event analyses and/or driver score calculations performed by the driving analysis server 220. As discussed below in reference to FIGS. 3-5, the driving analysis server 220 may retrieve and use data from databases/systems 231-233 to analyze and evaluate the driving events identified from the driving data of vehicles 210.

FIG. 3 is a flow diagram illustrating an example method of performing a driving event analysis and adjusting driver score based on a driving event identified for a vehicle, using image data, video data, and/or object proximity data associated with the driving event. This example method may be performed by one or more computing devices (e.g. driving analysis server 220, vehicle operation computer system 225, and vehicle telematics device 216) in the driving analysis system 200.

The steps in the example method of FIG. 3 describe performing an analysis to determine whether or not to adjust a driver score in response to a potentially high-risk or unsafe driving event (e.g., an occurrence of sudden braking or swerving, an impact to the vehicle, or a moving violation, etc.), using image data, video data, and/or object proximity data associated with the driving event. For instance, an occurrence of sudden swerving or braking by a vehicle 210 may indicate a high-risk or unsafe driving behavior by a driver not paying attention to the road and the vehicle's immediate surroundings. However, if the image data, video data, and/or object proximity data shows that a pedestrian, animal, other vehicle, or object quickly and unexpectedly moved into the path of the vehicle 210, then an occurrence of sudden swerving or braking by the vehicle 210 may indicate that the driver of the vehicle 210 was paying proper attention and reacted appropriately to the situation. As another example, speeding or ignoring a posted traffic sign may be considered a potentially high-risk or unsafe driving behavior, and drivers that speed excessively and commit frequent moving violations may receive lower driver scores than other drivers. However, for a vehicle 210 speeding or committing a moving violation, image data, video data, and/or object proximity data may show that there was an external cause for the driving event (e.g., obscured traffic signs, road obstructions, malfunctioning traffic lights, etc.), or may show that a minor speeding offense or moving violation should not be considered high-risk or unsafe driving (e.g., based on road conditions, weather, visibility, etc.).

In step 301, a driving analysis server 220 may receive vehicle operation data (or driving data) for a vehicle 210. As described above in reference to FIG. 2, the driving analysis server 220 may receive vehicle operation data from one or more vehicle operation computer systems 225 and/or directly from telematics devices 216 or other systems on vehicles 210. The vehicle driving data may correspond to sensor data collected by sensors 212 and/or additional data collected by a telematics device 216 or other systems within a vehicle 210. In addition to the vehicle operation data, the driving analysis server 220 may receive location and time information corresponding to the vehicle operation data in step 301. Vehicle location data and time data may be received from the same sources as other vehicle operation data, or may be collected by different data sources or generated by the driving analysis server 220. For example, the driving analysis server 220 may receive vehicle operation data from a vehicle operation system 225, and then may initiate communication with the vehicle's telematics device 216, GPS servers, time servers, or other systems to determine the location and time that correspond to the received vehicle operation data.

In certain embodiments, telematics devices 216, vehicle operation systems 225, and other data sources may transmit vehicle operation data for a vehicle 210 to the driving analysis server 220 in real-time (or near real-time). The driving analysis server 220 may be configured to receive the vehicle operation data, and then perform real-time (or near real-time) driving analyses and driver score calculations for the vehicle 210. In other embodiments, vehicle operation data might not be transmitted in real-time but may be sent periodically (e.g., hourly, daily, weekly, etc.) by telematics devices 216 or vehicle operation systems 225. Periodic transmissions of vehicle operation data may include data for a single vehicle or single driver, or for multiple vehicles or drivers. The driving analysis server 220 may be configured to receive the periodic transmissions, and then to perform periodic driving event analyses and driver score calculations for one or more vehicles and drivers.

In step 302, the driving analysis server 220 may identify one or more potentially high-risk or unsafe driving events (or driving behaviors) within the operation data of the vehicle 210. The driving events identified in step 302 may correspond to specific occurrences or patterns of high-risk, unsafe, or illegal driving activities that have the potential to affect the driver score of the vehicle 210 or a driver of the vehicle 210.

For certain such driving events, the driving analysis server 220 may identify the driving event by analyzing only the vehicle operation data, for example, occurrences of sudden braking, accelerating, or swerving. Other driving events that may be identified based only on the vehicle operation data include failure to use seatbelts, phone usage while driving, loud noise levels inside the vehicle while driving (e.g., high stereo volume or passenger noises), or other distractions in the vehicle (e.g., animated passengers or pets in the vehicle, eating while driving, texting while driving, etc.). Additionally, impacts to the body of the vehicle 210 (e.g., minor accidents, driving fast over speed bumps or dips, etc.) may be identified based only on the operation data received for the vehicle 210.

To identify other types of driving events, the driving analysis server 220 may analyze the vehicle operation data, as well as additional data retrieved from one or more external data sources. For example, to identify an occurrence of speeding by the vehicle 210, the driving analysis server 220 may receive vehicle speed data from a telematics device 216 or vehicle operation system 225, and may receive the vehicle's location from the telematics device 216, vehicle operation system 225, or a separate GPS system associated with the vehicle 210. Then, the driving analysis server 220 may access a speed limit database to determine the legal speed limit at the location of the vehicle 210, and may compare the speed limit to the detected speed of the vehicle. The driving analysis server 220 may identify other moving violations using similar techniques. For example, the driving analysis server 220 may identify a failure to use proper turn signals by analyzing the turn signal usage of the vehicle 210, as compared to the location/driving route of the vehicle. Stop sign violations, illegal turns, and U-turn violations may be identified by comparing the driving route of the vehicle 210 to a database of the traffic regulations and posted traffic signs at different streets and intersections along the driving route.

Additional driving behaviors that may be identified in step 302 include occurrences of risky or aggressive driving under adverse driving conditions or in safe driving areas. For example, it may be deemed high-risk or unsafe to drive a vehicle 210 at the maximum speed limit during a rainstorm, snowstorm, or on slick or icy road conditions. To detect an occurrence of this type of driving event, the driving analysis server 220 may analyze the speed of the vehicle 210, the speed limit at the vehicle's location, and the weather or road conditions at the time the vehicle was being driven at that location. Additionally, it may be deemed high-risk or unsafe to drive aggressively in safe driving areas, such as construction zones and school zones. To identify an occurrence of aggressive driving in a safe driving area, the driving analysis server 220 may analyze certain vehicle operation data (e.g., sudden acceleration or braking, phone usage while, other driver distractions, etc.), and compare the vehicle's location to a database of safe driving areas.

In step 303, the driving analysis server 220 may retrieve additional image data, video data, and/or object proximity data associated with the driving event(s) identified for the vehicle 210 in step 302. As discussed below, the image, video, and proximity data retrieved in step 303 may allow the driving analysis server 220 to identify external causes and/or justifications for potentially high-risk or unsafe driving events or behaviors. For example, if the driving event identified in step 302 is an occurrence of sudden braking or swerving by the vehicle 210, or an impact to the vehicle 210, then in step 303 the driving analysis server 220 may retrieve image, video, and object proximity data for the time and location of the event in order to determine if there was an external cause for the swerving, braking, or vehicle impact.

The image, video, and object proximity data received by the driving analysis server 220 in step 303 may correspond to various different types of cameras or sensors, and may be received from one or more different data sources. For example, the image, video, and object proximity data may include data recorded by the cameras and proximity sensors 214 of same vehicle 210 at which the driving event was identified in step 302. The data received in step 303 also may include data recorded by cameras and proximity sensors 214 in one or more other vehicles 210a-n that were in the immediate proximity of the vehicle 210 at the time of driving event. Additionally, the data received in step 303 may include image, video, and proximity data from other data sources, such traffic camera data received from one or more traffic databases/systems 231.

In certain embodiments, the driving analysis server 220 may identify a potentially high-risk or unsafe driving event of the vehicle 210 in real-time or near real-time (in step 302), and may retrieve image, video, or object proximity data associated with the driving event or near the same time (in step 303). For example, after identifying a potentially high-risk or unsafe driving event at the vehicle 210, one or more of the telematics devices 216 of the vehicle, the vehicle operation system 225, or the driving analysis server 220 may be used to identify the cameras and proximity sensors 214 of the vehicle 212 or other nearby vehicles 210a-n, or images and video feeds from traffic cameras in the vicinity of the vehicle 210 at the time of the identified driving event. GPS systems or proximity sensors within the vehicle 210 and/or vehicle location data from other sources may be used to identify traffic cameras or other data sources for image, video, and object proximity data in the vicinity of the vehicle 210 at the time of the driving event. Alternatively, the cameras or proximity sensors near the vehicle 210 at the time of the driving event may be identified at a later time, based on previously stored location data (e.g., GPS coordinates, driving route data) and time data for the vehicle 210, one or more other vehicles 210a-n, and traffic cameras and other data sources 231.

In some instances, the driving analysis server 220 may already have, or may already be configured to receive, the associated image data, vehicle data, and object proximity data from the vehicle cameras and proximity sensors 214 of the vehicle 210. For example, the driving analysis server 220 may receive a stream or batch of vehicle operation data from one or more vehicles 210a-n or vehicle operation systems 225 containing all of the data recorded by all vehicle operation sensors 212, including the cameras and proximity sensors 214, in the set of vehicles 210a-n. Thus, the driving analysis server 220 may receive the associated image, video, and proximity data for a vehicle 210 at the time same that it receives the speed, braking, acceleration, and steering data, and other sensor data from which it may identify a driving event. In other instances, the driving analysis server 220 may specifically request image, video, or object proximity data from the vehicle operation system 225, vehicle telematics devices 216, or other data sources corresponding to the time and location of the driving event, in response to identifying the driving event for the vehicle 210. Thus, the driving analysis server 220 may be configured to locate, initiate communication with, and request specific sensor/camera data from vehicle operation systems 225, vehicle telematics devices 216, and/or traffic camera data sources 231 corresponding to the time and location of the driving event.

As an example, if the driving event identified in step 302 is an impact to the front of a first vehicle 210, then the driving analysis server 220 may be configured to specifically request and retrieve image, video, and object proximity data from the front-facing cameras and sensors 214 of first vehicle 210. Additionally, the driving analysis server 220 may use the location and time data associated with the driving event to identify one or more other vehicles 210a-n in the vicinity of the first vehicle 210 at the time of the driving event. For instance, the driving analysis server 220 may access GPS system data, location and time data from telematics devices 216, or location and time data from vehicle operation systems 225 for a plurality of other vehicles. After identifying one or more other vehicles 210a-n near the first vehicle 210 at the time of the driving event, the driving analysis server 220 may request and retrieve image, video, and object proximity data from the relevant cameras and sensors 214 of those vehicles. For instance, data from rear-facing cameras and proximity sensors may be requested from vehicles that were in front of the first vehicle 210 at the time of the driving event, data from front-facing cameras and proximity sensors may be requested from vehicles that were behind the first vehicle 210 at the time of the driving event, etc. The driving analysis server 220 also may use the location and time data associated with the driving event to identify one or more traffic cameras or other traffic data sources that may have recorded image, video, or proximity data in the vicinity of the first vehicle 210 at the time of the driving event (e.g., traffic cameras at the road or intersection where the driving event occurred). The driving analysis server 220 may request and retrieve traffic camera images or video recordings from one or more traffic data sources 231 corresponding to the time and location of the driving event.

In step 304, the driving analysis server 220 may analyze the image, video, and/or object proximity data retrieved in step 303 to determine one or more causes of the driving event identified in step 302. For example, if the identified driving event is an occurrence of sudden braking or swerving, or an impact to a vehicle 210, then the image, video, and/or object proximity data may be analyzed to determine that the braking, swerving, or impact was caused by one or more of: an obstruction in the path in the vehicle 210; adverse road conditions; an electrical or mechanical malfunction within the vehicle; and/or high-risk or unsafe driving by the driver of the vehicle. As another example, if the identified driving event is an occurrence of speeding or another type of moving violation, then the image, video, and/or object proximity data may be analyzed to determine that the speeding or moving violation was caused by one or more of: a disabled vehicle or other road obstruction, a road construction zone, an obscured or missing traffic sign, and/or a malfunctioning traffic signal. FIGS. 4 and 5, discussed below, respectively describe examples of driving event analyses and determinations of a driver score adjustments based on a braking, swerving, or vehicle impact driving event (FIG. 4), or based on a speeding or moving violation driving event (FIG. 5).

To determine a cause for a driving event in step 304, the driving analysis server 220 may analyze and measure the image, video, and/or proximity data retrieved in step 303, calculate relevant data metrics relating to the event, and compare the data metrics to predetermined thresholds. For example, image, video, and/or proximity data may show that a vehicle 210 swerved to avoid hitting an object that was in the vehicle's path (e.g., a pedestrian, cyclist, animal, disabled vehicle, etc.), and then hit another object (e.g., a tree or a parked car) as a result. In this case, the driving analysis server 220 may analyze the image, video, and/or proximity data to determine the distance between the object and the vehicle when the object entered the vehicle's path, or the distance at which the object was first visible to the driver of the vehicle 210. From this distance, the driving analysis server 220 may use the vehicle's speed to calculate the amount of time that the driver had to react to the obstruction, and may compare that amount of time to a reaction time threshold. For instance, if the driver had very little time to react to the obstruction (e.g., (0-2 seconds), the driving analysis server 220 may determine that the obstruction was the cause of the swerving/accident event, rather than any unsafe or high-risk driving on the part of the driver. On the other hand, if the driver had ample time (e.g., 5-10 seconds) to observe and avoid the obstruction, but only swerved at the last second (hitting the parked car or tree), the driving analysis server 220 may determine that the obstruction was not the cause of the swerving/accident event.

In addition to calculating a reaction time data metric and comparing it to a reaction time threshold, as described in the example above, the driving analysis server 220 may calculate and use additional data metrics and thresholds for determining the causes of driving events. As another example, after an accident between two vehicles, the driving analysis server 220 may use the image, video, and/or proximity data to determine the following distance of the trailing vehicle just before the accident. Based on the following distance, the driving analysis server 220 may use the speed of the vehicles to calculate a tailgating time metric for the trailing vehicle, which may be compared to a tailgating safety threshold (e.g., 2 seconds) to determine whether or not tailgating by the trailing vehicle will be classified as a cause of the accident. Additional data metrics relating to weather conditions, road conditions, and visibility may be calculated using image data and video data received in step 303, and may be used in combination with other data metrics to determine causes of driving events. For instance, a first tailgating safety threshold (e.g., 2 seconds) may be applied in good weather and high visibility conditions, while a second tailgating safety threshold (e.g., 5 seconds) may be applied in poor weather and low visibility conditions such as rain, fog, icy roads, etc. After a traffic accident, if the analysis of the image, video, and proximity data indicates that the trailing vehicle was tailgating closer than the tailgating safety threshold for the current road/weather/visibility conditions, then the driving analysis server 220 may determine that tailgating by the trailing vehicle was a cause of the accident.

As another example, a first speeding safety threshold (e.g., 10 MPH over the posted speed limit), a first turn signal safety threshold (e.g., 100 feet before turning), and/or a first braking safety threshold (e.g., 80% of maximum braking pressure) may be applied for daytime driving on good road conditions, while a second speeding safety threshold (e.g., 5 MPH under the posted speed limit), a second turn signal safety threshold (e.g., 250 feet before turning), and/or a second braking safety threshold (e.g., 40% of the maximum braking pressure) may be applied for specific roads/locations or specific ranges of adverse road conditions (e.g., low visibility ranges, poor road traction ranges, etc.). In addition to these examples, other types of safe/unsafe driving thresholds (e.g., vehicle speed thresholds, vehicle maintenance/operational condition thresholds, tailgating thresholds, lane change thresholds, lane departure thresholds, reaction time thresholds, weather condition thresholds, road condition thresholds, traffic condition thresholds, road visibility thresholds, and thresholds relating to the driver distractions and the use of vehicle controls, etc.) may be used in the analyses of step 304.

The determination in step 304 may identify a single cause or multiple causes for a driving event. As illustrated in the examples above, a driving event such as a sudden occurrence of swerving or braking, or an impact to the vehicle, may have multiple causes. Additionally, a moving violation may have multiple possible causes, for example, a traffic sign that is blocked by an obstruction, bad road conditions, a road construction project, and/or an impatient or high-risk driver. Thus, the driving analysis server 220 may identify multiple causes for a driving event in step 304. In certain examples, the driving analysis server 220 may use the calculated driving metrics and thresholds to calculate ranks and/or percentages associated with one or more determined causes of a driving event. For instance, a traffic accident between two vehicles may be determined by the driving analysis server 220 to have been 30% caused by tailgating by the trailing vehicle, 15% caused by unexpectedly slick road conditions, 5% caused by old and worn brake pads in the trailing vehicle 210, and 50% caused by a sudden occurrence of jaywalking by a pedestrian in front of the leading vehicle.

In step 305, the driving analysis server 220 may determine if the cause (or one of the causes) of the driving event identified in step 302 was high-risk or unsafe driving by the driver of the vehicle 210. In this example, if high-risk or unsafe driving by the driver of the vehicle 210 was a cause of the driving event (305: Yes), then the driving analysis server 220 may negatively adjust a driver score associated with the vehicle 210 or a driver of the vehicle 210 in step 306. On the other hand, if high-risk or unsafe driving by the driver of the vehicle 210 was not a cause of the driving event (306: No), then the driving analysis server 220 might not adjust the driver score associated with the vehicle 210 or driver in step 307.

As discussed above, the analysis in step 304 may determine a single cause or multiple causes for a driving event, and may calculate ranks or percentages associated with each determined cause. Thus, the determination performed by the driving analysis server 220 in step 305 may take into account each of determined causes and the associated ranks or percentages, in deciding whether or not to adjust a driver score and by how much. For example, the driving analysis server 220 may negatively adjust a driver score only if unsafe or high-risk driving was a primary cause of the driving event, or was greater than a predetermined event cause threshold (e.g., >30% cause, >40% cause, >50% cause, etc.). In the case of a driving event affecting two vehicles 210a and 210b, the driving analysis server 220 may receive and analyze the data as described above, and then may adjust a driver score associated with one or both of the vehicles 210a and 210b. For example, an analysis of the image, video, and proximity data for an accident between a first vehicle 210a and a second vehicle 210b may show that the accident was caused partially by erratic and distracted driving by the first driver in vehicles 210a, and partially by tailgating too closely by the second driver in vehicle 210b, resulting in vehicle 210b rear-ending vehicle 210a. In this case, the driving event may be analyzed and driver scores for both vehicles may be adjusted accordingly.

Although the example of FIG. 3 shows a negative driver score adjustment (step 306), in other examples a driver score may be positively adjusted based on the image, video, and proximity data analysis associated with the driving event. For instance, if an image, video, and proximity data analysis in step 304 indicates that a driver was not speeding or tailgating based on the current road conditions, and that the driver reacted quickly and appropriately to an unforeseen driving event (e.g., by quickly swerving or braking to avoid a pedestrian), then the driver score may be positively adjusted (e.g., raised) based on the driving event. Both positive and negative driver score adjustments may be made in varying degrees/magnitudes based on the image, video, and proximity data analysis in step 304. For instance, the driver scores for vehicles/drivers performing in an exceptionally competent and safe manner may be positively adjusted by a greater amount, and vice versa. The magnitude of the driver score adjustments may be based on the magnitude of the data metrics calculated (e.g., speed of driver reaction time, degree of tailgating, degree of speeding, degree of poor driving conditions, degree of low visibility, degree of driver distractions, etc.), and may be determined using additional threshold comparisons in step 304.

In the above examples, the identification of driving events (steps 301-302), retrieval of image, video, and proximity data (step 303), driving event analyses (step 304), and driver score determinations (steps 305-307), may be performed by a driving analysis server 220 external to the vehicle 210. In other examples, a driving analysis server 220 (or other devices) may be located partially or entirely within a vehicle 210 and may be configured to perform some or all of the same steps described above in steps 301-307. For instance, one or more devices within the vehicle 210 (e.g., a telematics device 216 or a driver's mobile device) may be configured as a driving analysis server 220. In certain examples, a driving analysis server 220 within the vehicle 210 may be configured to report the driving event to a vehicle operation system