Autonomous vehicle operation based on real-time analytics
Systems and methods are disclosed for operating an autonomous vehicle based on real-time operating data. The operating data may be data about vehicles, drivers, passengers, as well as relevant environmental conditions and contextual data. In some cases, historical data for the preceding data types may be used. The systems and methods may obtain a set of real-time operating data indicative of one or more behaviors of an autonomous vehicle. One or more operations may be performed on the set of real-time operating data. An instruction to modify a particular vehicle operation may be generated based on output from the operations and the one or more behaviors of the autonomous vehicle, and the instruction to modify the particular vehicle operation may be provided to a particular processor that is on-board the vehicle and that controls the particular vehicle operation, to thereby automatically modify the particular vehicle operation.
1 . A method, comprising:
determining, by a processor, during a trip taken by a vehicle, and based on first data indicating an operating parameter of the vehicle during a first portion of the trip, that the operating parameter satisfies a first threshold, the first threshold corresponding to a first weather condition occurring during the first portion of the trip;
receiving, by the processor, data indicating a second weather condition occurring during a second portion of the trip, the second weather condition being different from the first weather condition;
selecting, by the processor and based on occurrence of the second weather condition during the second portion of the trip, a model trained to generate a threshold corresponding to the second weather condition;
generating, by the processor and using the model, a second threshold corresponding to the second weather condition;
determining, by the processor and based on second data indicating the operating parameter as exhibited by the vehicle during the second portion of the trip, that the operating parameter fails to satisfy the second threshold; and
providing, by the processor and based on determining that the operating parameter fails to satisfy the second threshold, an instruction to an additional processor, the instruction, when executed by the additional processor, causing modification of a vehicle operation corresponding to the operating parameter.
2 . The method of claim 1 , wherein the model is a context-specific model trained on historical operating data indicating past behaviors of vehicles operating in conditions substantially similar to the second weather condition.
3 . The method of claim 1 , wherein:
the model is characterized by a first plurality of weights, and
the first threshold is generated by an additional model characterized by a second plurality of weights, and different from the first plurality of weights.
4 . The method of claim 1 , wherein the first data and the second data include one or more of:
sensor data from one or more sensors carried by the vehicle,
indicator data comprising a current physical state of a component of the vehicle, or
status data indicative of a current status of oil levels, fuel levels, or cabin conditions of the vehicle.
5 . The method of claim 1 , wherein the vehicle operation comprises at least one of: braking, accelerating, steering, lane change or turning.
6 . The method of claim 1 , further comprising:
obtaining, by the processor, an indication that a driver of the vehicle has opted-in to automatic modification of vehicle operations,
wherein the instruction causes an automatic modification of the vehicle operation based on obtaining the indication.
7 . The method of claim 1 , wherein the first weather condition is determined based on one or more of:
data from one or more sensors carried by the vehicle,
third-party weather data based on a geographic location of the vehicle, or
information from an environmental monitoring application operating on a remote computing system communicatively connected to the processor.
8 . The method of claim 1 , further comprising:
obtaining, by the processor, a set of historical operating data associated with the vehicle;
identifying, by the processor and based on the set of historical operating data, one or more alert conditions; and
determining, by the processor, that the second data indicates an alert condition of the one or more alert conditions,
wherein the instruction causes the modification of the vehicle operation to alleviate the alert condition.
9 . The method of claim 1 , further comprising:
determining, by the processor, a vehicle response to a change in weather condition from the first weather condition to the second weather condition;
determining, by the processor and based at least in part on the vehicle response, a vehicle performance score; and
determining, by the processor and based at least in part on the vehicle performance score, an insurance premium amount.
10 . The method of claim 9 , wherein the vehicle response comprises one or more of:
a length of time from a change in the weather condition and a reaction by the vehicle,
a content of the reaction, and
a magnitude of the reaction.
11 . A system, comprising:
a processor;
a plurality of sensors configured to interface with the processor; and
a non-transitory memory storing computer-executable instructions that, when executed, cause the processor to:
determine, during a trip taken by a vehicle, and based on first data from the plurality of sensors, that an operating parameter of the vehicle during a first portion of the trip satisfies a first threshold, the first threshold corresponding to a first weather condition occurring during the first portion of the trip;
receive data indicating a second weather condition occurring during a second portion of the trip, the second weather condition being different from the first weather condition;
select, based on occurrence of the second weather condition during the second portion of the trip, a model trained to generate a threshold corresponding to the second weather condition;
generate, using the model, a second threshold corresponding to the second weather condition;
determine, based on second data from the plurality of sensors, that the operating parameter as exhibited by the vehicle during the second portion of trip fails to satisfy the second threshold; and
provide, based on determining that the operating parameter fails to satisfy the second threshold, an instruction to an additional processor, the instruction, when executed by the additional processor, causing modification of a vehicle operation corresponding to the operating parameter.
12 . The system of claim 11 , wherein the second data comprises a set of time-series data, the time-series data representing (i) the operating parameter at each time instance and (ii) the weather condition at the respective time instance.
13 . The system of claim 11 , the instructions further causing the processor to:
determining, based on the second data, that the second data indicates an alert condition of one or more pre-determined alert conditions,
wherein the instruction causes the modification of the vehicle operation to alleviate the alert condition.
14 . The system of claim 13 , wherein the pre-determined alert conditions are specified by a driver of the vehicle or by an insurance provider of the vehicle.
15 . The system of claim 11 , wherein the model is a context-specific model trained on historical operating data indicating past behaviors of vehicles operating in conditions substantially similar to the second weather condition.
16 . A tangible, non-transitory computer-readable medium storing executable instructions for operating an autonomous vehicle based on real-time operating data that, when executed by a processor of a system, cause the processor to:
determine, during a trip taken by a vehicle, and based on first data indicating an operating parameter of the vehicle during a first portion of the trip, that the operating parameter satisfies a first threshold, the first threshold corresponding to a first weather condition occurring during the first portion of the trip;
receive data indicating a second weather condition occurring during a second portion of the trip, the second weather condition being different from the first weather condition;
select, based on occurrence of the second weather condition during the second portion of the trip, a model trained to generate a threshold corresponding to the second weather condition;
generate, using the model, a second threshold corresponding to the second weather condition;
determine, based on second data indicating the operating parameter as exhibited by the vehicle during the second portion of the trip, that the operating parameter fails to satisfy the second threshold; and
provide, based on determining that the operating parameter fails to satisfy the second threshold, an instruction to an additional processor, the instruction, when executed by the additional processor, causing modification of a vehicle operation corresponding to the operating parameter.
17 . The tangible, non-transitory computer-readable medium of claim 16 , wherein the executable instructions further cause the processor to:
obtain an indication that a driver of the vehicle has opted-in to automatic modification of vehicle operations,
wherein the instruction causes an automatic modification of the vehicle operation based on obtaining the indication.
18 . The tangible, non-transitory computer-readable medium of claim 16 , wherein the executable instructions further cause the processor to:
obtain a set of historical operating data associated with the vehicle;
identify, based on the set of historical operating data, one or more alert conditions; and
determine, based on the second data, that the second data indicates an alert condition of the one or more alert conditions,
wherein the instruction causes the modification of the vehicle operation to alleviate the alert condition.
19 . The tangible, non-transitory computer-readable medium of claim 16 , wherein:
the model is a context-specific model characterized by a first plurality of weights trained on historical operating data indicating past behaviors of vehicles operating in conditions substantially similar to the second weather condition, and
the first threshold is generated by an additional model characterized by a second plurality of weights, different from the first plurality of weights.
20 . The tangible, non-transitory computer-readable medium of claim 16 , wherein the first data and the second data include one or more of:
sensor data from one or more sensors carried by the vehicle,
indicator data comprising a current physical state of a component of the vehicle, or
status data indicative of a current status of oil levels, fuel levels, or cabin conditions of the vehicle.