Vehicle-data analytics

Abstract

Provided is a system configured to determine and push adjustments to vehicle operations using machine-learning systems across multiple computing layers.

1 - 20 . (canceled) 
     
     
         21 . A non-transitory, machine-readable medium storing instructions that, when executed by one or more processors, effectuate operations comprising:
 receiving a vehicle profile and an operator profile corresponding to an active vehicle and an operator of the active vehicle, wherein the active vehicle includes one or more of a first set of processors executing an active vehicle application operating as part of a vehicle computing layer, wherein the vehicle computing layer comprises a set of vehicle applications operating on the first set of processors;   receiving a set of vehicle sensor data obtained by a set of sensors of the active vehicle;   receiving a set of geolocation data corresponding to a current geographic region of the active vehicle;   processing the vehicle sensor data and geolocation data using a trained neural network executing on a vehicle application;   performing a machine-learning prediction or inference operation using the trained neural network to update the vehicle profile and the operator profile; and   updating one or more vehicle control metrics in accordance with the operator profile and vehicle profile.   
     
     
         22 . The medium of  claim 21 , comprising:
 receiving a second set of vehicle sensor data obtained by the set of sensors of the active vehicle;   identifying a second active vehicle distinct from the active vehicle using the second set of vehicle sensor data;   generating a second vehicle profile corresponding to the second active vehicle;   transmitting the second vehicle profile to a computing system.   
     
     
         23 . The medium of  claim 22 , wherein in response to transmitting the second vehicle profile, receiving, from the computing system, an indication the second active vehicle satisfies a vehicle risk warning threshold, causing the active vehicle to display a warning to occupants of the active vehicle. 
     
     
         24 . The medium of  claim 21 , wherein the vehicle sensor data includes a video stream, the operations further comprise:
 using the trained neural network:
 detecting an object within the video stream; 
 determining the object is an anomalous object using the video stream and the set of geolocation data; and 
   updating the geolocation data to include an indication and location of the anomalous object.   
     
     
         25 . The medium of  claim 24 , the operations further comprising:
 using the trained neural network, determining the anomalous object is a dangerous anomalous object;   sending, to a computing device, a visual indicator of the dangerous anomalous object from the video stream, and a location of the dangerous anomalous object; and   updating one or more vehicle control metrics to avoid collision with the dangerous anomalous object.   
     
     
         26 . The medium of  claim 21 , further comprises:
 receiving an indication of a destination of the active vehicle;   determining a navigation route from a current location of the active vehicle to the destination using the geolocation data, the navigation route comprising multiple segments;   retrieving one or more road risk values associated with each of the multiple segments of the navigation route;   calculating, a route risk value based on the one or more road risk values;   in accordance with the route risk value meeting a threshold value, generating an alternative navigation route to the destination by removing one or more segments of the navigation route.   
     
     
         27 . The medium of  claim 26 , wherein determining the route risk value further comprises:
 assigning a vehicle risk value based on one or more vehicle features including: size, length, weight, number of occupants, average speed, tire tread depth, tire pressure, and drivetrain.   
     
     
         28 . The medium of  claim 21 , wherein the operations further comprise:
 obtaining a publicly available user profile of the operator of the active vehicle;   updating the operator profile based on data from the publicly available user profile of the operator; and   linking the publicly available user profile of the operator with the operator profile.   
     
     
         29 . The medium of  claim 21 , wherein the operations further comprise:
 determining the active vehicle is idle;   in accordance with determining the active vehicle is idle, performing a self-diagnostic test on the set of sensors of the active vehicle;   determining a sensor of the set of sensors is inactive;   determining the sensor is defective; and   in response to determining the sensor is defective, causing the active vehicle to display a warning to the operator that the respective sensor is offline.   
     
     
         30 . The medium of  claim 21 , wherein the active vehicle includes a control system of the active vehicle, the operations comprising:
 obtaining a set of control-system data comprising data indicating a use of the control system of the active vehicle, the control-system data including a measure of a force at which the vehicle operator presses on a brake pedal;   using the trained neural network to infer a braking response rate based on using the measure of the force, a number of detected other moving objects, and a number of recorded accidents proximate to the active vehicle's current geolocation as inputs; and   adjusting the active vehicle response including changing the brake response rate, wherein an increase in the active vehicle response rate increases the vehicle brake rate with respect to a push in the vehicle brake pedal.   
     
     
         31 . The medium of  claim 30 , wherein an application operating on a computing layer applies the trained neural network to determine a vehicle accident has occurred;
 determining rescue scores for each of a plurality of neighboring rescue vehicles based on activity and statuses of the neighboring rescue vehicles;   selecting a neighboring rescue vehicle with a highest rescue score; and   causing display on a display of the neighboring rescue vehicle with the highest rescue score,   a location of the vehicle accident, visual imagery of the vehicle accident, and a navigation route to the vehicle accident location.   
     
     
         32 . The medium of  claim 21 , wherein targeted content is provided to the active vehicle for display based on the active vehicle profile and the operator profile. 
     
     
         33 . The medium of  claim 32 , wherein the targeted content is provided to the active vehicle by way of routing the active vehicle through a route calculated to provide the targeted content. 
     
     
         34 . The medium of  claim 29 , wherein causing the display of a warning includes:
 causing display of the warning on a heads-up-display of the active vehicle.   
     
     
         35 . The medium of  claim 21 , wherein the operations comprise inferring a predictive adjustment package including adjusting gear ratios, suspension systems, headlights in anticipation of a vehicle occurrence. 
     
     
         36 . The medium of  claim 21 , further comprising:
 predicting a vehicle operation score based on the active vehicle profile and the operator profile;   updating the vehicle operation score based on a trip safety score;   calculating an updated vehicle operation score to store in memory with the operator profile; and   providing, to a computing system, the updated vehicle operation score.   
     
     
         37 . The medium of  claim 21 , further comprising:
 selecting a road network graph from a road network graph repository based on the set of geolocations;   receiving, from the neural network, an indication that a projected position along a navigational path includes a high-risk region; and   providing, to the operator of the active vehicle, an option to re-route the navigational path to avoid the high-risk region.   
     
     
         38 . The medium of  claim 37 , the operations further comprising: in accordance with receiving an indication to continue along the navigational path, updating one or more vehicle settings including activating door locks, adjusting suspension levels, decreasing volume of an entertainment system, and increasing a sensitivity level of computer assisted braking. 
     
     
         39 . The medium of  claim 21 , the operations further comprising:
 providing, to a computing system, a concatenated profile including the active vehicle profile and the operator profile;   receiving, from the computing system, an emergency indication that the operator is high risk and;   in accordance with a determination using sensor data of the active vehicle that it is safe to disable the active vehicle, disabling the active vehicle.   
     
     
         40 . A method, comprising:
 receiving a vehicle profile and an operator profile corresponding to an active vehicle and an operator of the active vehicle, wherein the active vehicle includes one or more of a first set of processors executing an active vehicle application operating as part of a vehicle computing layer, wherein the vehicle computing layer comprises a set of vehicle applications operating on the first set of processors;   receiving a set of vehicle sensor data obtained by a set of sensors of the active vehicle;   receiving a set of geolocation data corresponding to a current geographic region of the active vehicle;   processing the vehicle sensor data and geolocation data using a trained neural network executing on a vehicle application;   performing a machine-learning prediction or inference operation using the trained neural network to update the vehicle profile and the operator profile; and   updating one or more vehicle control metrics in accordance with the operator profile and active vehicle profile.   
     
     
         41 . The method of  claim 40 , further comprises:
 steps for determining derived sensor data;   steps for securing sensitive data;   steps for obtaining roadside sensor results; and   steps for transmitting results of the machine-learning prediction or inference operation.   
     
     
         42 . the method of  claim 41 , further comprises:
 steps for training the trained neural network;   steps for performing a second machine learning prediction or inference operation on the trained neural network;   steps for determining control system adjustment values;   steps for generating an online profile;   steps for updating the active vehicle profile, operator profile, or online profile; and   steps for adjusting control system values.