Personalized route indices via crowd-sourced data
Traffic incident information may be clustered to determine geographical locations in which a statistically-increased level of traffic incidents occur. Indices indicative of vehicle operating conditions within the locations may be generated according to crowd-sourced event information factors and an index function including coefficients defining contributions of the factors to the indices. The coefficients may be optimized to minimize error between the index-predicted statistically-increased level areas and the clustered traffic incident information.
1. A system comprising:
a vehicle processor programmed to
cluster traffic incident information to determine geographical locations in which a statistically-increased level of traffic incidents occur;
generate traffic incident likelihood indices predictive of traffic incidents levels within the geographic locations according to crowd-sourced event information factors and an index function including coefficients defining contributions of the information factors to the traffic incident likelihood indices, the index function being of the form
Ri ( k )= f _1(geometry( k ), a _1 , . . . ,a _ n )+ f _2(weather( k ), a _( n+ 1), . . . , a _ m )+ f _3(crowd data( k ), a _( m+ 1), . . . , a _ l ),
where Ri represents the traffic incident likelihood indices, k represents the geographic locations for which the indices Ri are being calculated, f 1 represents a function of road geometry information, f 2 represents a function of weather conditions, f 3 represent a function of the crowd-sourced event information, and a 1 , . . . , a l represent the coefficients defining contributions used to weight the factors in creation of the indices Ri; and
optimize the coefficients to minimize error between the indices Ri and the clustered traffic incident information; and
generate traffic incident likelihood alerts responsive to the traffic incident likelihood indices exceeding a threshold level of traffic incident likelihood.
2. The system of claim 1 , wherein the crowd-sourced event information includes at least two of traffic data, environment data, road type data, human factor data, and vehicle parameter data.
3. The system of claim 1 , wherein the factors include hard braking events and hard steering events.
4. The system of claim 1 , wherein the vehicle processor is further configured to optimize the coefficients according to historical crowd-sourced event information of a specific vehicle driver to determine coefficients configured to generate indices optimized for the driver.
5. The system of claim 1 , wherein the traffic incident information further includes traffic incidents identified according to the crowd-sourced event information.
6. The system of claim 1 , wherein the vehicle processor is included in an index server device further configured to receive the crowd-sourced event information over a wide area network from a plurality of vehicles.
7. The system of claim 1 , wherein the vehicle processor is further configured to: receive a request for an index for a geographic location from a vehicle; and provide the index to vehicle according to the index function and coefficients.
8. The system of claim 7 , wherein the geographic location is a current geographic area in which the vehicle is located.