Adaptive sensor fusion
A computer, including a processor and a memory, the memory including instructions to be executed by the processor to determine performance of a plurality of vehicle data sources used to operate a vehicle by evaluating each vehicle data source output data and train a deep neural network to determine reliability for each of the vehicle data sources based on the performance using reinforcement learning. The instructions can further include instructions to combine output data from the vehicle data sources based on the reliability including using the deep neural network to correlate output data from one or more vehicle data sources to the performance to determine how accurately the output data from each vehicle data source corresponds to vehicle performance and operate the vehicle based on combined output data.
1. A computer, comprising a processor; and a memory, the memory including instructions executable by the processor to:
determine performance of a plurality of vehicle data sources used to operate a vehicle by evaluating each vehicle data source output data;
train a deep neural network to determine reliability for each of the vehicle data sources based on the performance using reinforcement learning;
combine output data from the vehicle data sources based on the reliability including using the deep neural network to correlate output data from one or more vehicle data sources to the performance to determine how accurately the output data from each vehicle data source corresponds to vehicle performance; and
operate the vehicle based on combined output data.
2. The computer of claim 1 , the instructions including further instructions to determine which data sources to include in the vehicle based on the reliability.
3. The computer of claim 1 , the instructions including further instructions to determine which data sources are operating correctly based on the reliability.
4. The computer of claim 1 , the instructions including further instructions to determine performance of each vehicle data source based on measures of vehicle operation including at least one of collision probability, fuel consumption, or travel time.
5. The computer of claim 1 , wherein each of the one or more vehicle data sources includes a vehicle sensor and a computing device to process sensor data.
6. The computer of claim 5 , wherein the vehicle sensors include one or more of a camera, a lidar sensor, an ultrasound sensor and a radar sensor.
7. The computer of claim 1 , the instructions including further instructions to combine output data from the vehicle data sources by determining weights that determine how output data are combined.
8. The computer of claim 1 , wherein performance includes missing data and sensor noise levels.
9. The computer of claim 1 , wherein output data from the vehicle data sources includes a location of an object in an environment around the vehicle.
10. The computer of claim 9 , wherein operating the vehicle includes determining a vehicle path based on the location of the object in the environment around the vehicle.
11. A method, comprising:
determining performance of a plurality of vehicle data sources used to operate a vehicle by evaluating each vehicle data source output data;
training a deep neural network to determine reliability for each of the vehicle data sources based on the performance using reinforcement learning;
combining output data from the vehicle data sources including using the deep neural network to correlate output data from one or more vehicle data sources to the performance to determine how accurately the output data from each vehicle data source corresponds to vehicle performance; and
operating the vehicle based on the combined output data.
12. The method of claim 11 , further comprising determining which data sources to include in the vehicle based on the reliability.
13. The method of claim 11 , further comprising determining which data sources are operating correctly based on the reliability.
14. The method of claim 11 , further comprising determining performance of each vehicle data source based on measures of vehicle operation including at least one of collision probability, fuel consumption, or travel time.
15. The method of claim 11 , wherein each of the vehicle data sources includes a vehicle sensor and a computing device to process sensor data.
16. The method of claim 15 , wherein the vehicle sensors include one or more of a camera, a lidar sensor, an ultrasound sensor and a radar sensor.
17. The method of claim 11 , further comprising combining output data from the vehicle data sources by determining weights that determine how output data are combined.
18. The method of claim 11 , wherein performance includes missing data and sensor noise levels.
19. The method of claim 11 , wherein output data from the vehicle data sources includes a location of an object in an environment around the vehicle.
20. The method of claim 19 , wherein operating the vehicle includes determining a vehicle path based on the location of the object in the environment around the vehicle.