Vehicle control system
A vehicle control system may be provided for controlling adhesion of wheels to a route surface. The control system includes one or more processors configured to determine adhesion values representative of adhesion between the wheels of a vehicle and the route surface based on angular speeds of the wheels. An artificial intelligence neural network may generate a target slip value for the wheels that are coupled with at least two different axles of the vehicle by processing the adhesion values and modifying the target slip value to increase an average value of the adhesion values of the wheels. The one or more processors may control a torque applied to at least one of the axles based on the target slip value.
1. A vehicle control system comprising:
one or more processors configured to determine adhesion values representative of adhesion between wheels of a vehicle and a route surface based on angular speeds of the wheels,
an artificial intelligence (AI) neural network having artificial neurons arranged in layers and connected with each other by connections, the AI neural network configured to subsequently calculate a target slip value for the wheels that are coupled with at least two different axles of the vehicle using the adhesion values that were determined, the AI neural network also configured to repeatedly modify the target slip value to increase an average value of the adhesion values of the wheels,
the one or more processors configured to control a torque applied to at least one of the axles based on the target slip value that is calculated,
the AI neural network configured to receive feedback regarding the target slip value that is calculated and modified by the artificial neurons,
the AI neural network configured to be trained by changing one or more connections between the artificial neurons in the AI neural network based on the feedback that is received.
2. The vehicle control system of claim 1 , wherein the AI neural network are configured to reduce a derivative of the average adhesion value with respect to the target slip value by changing the torque applied to the at least one of the axles.
3. The vehicle control system of claim 2 , wherein the AI neural network are configured to determine an integral of the derivative of the average value of the adhesion values with respect to the target slip value, the AI neural network configured to determine or change the target slip value based on the integral that is determined.
4. The vehicle control system of claim 1 , wherein the at least two different axles include a leading axle and a trailing axle along a direction of movement of the vehicle, the AI neural network are configured to increase the target slip value associated with the leading axle to clean the mute surface with at least one of the wheels coupled with the leading axle prior to travel of at least another one of the wheels that is coupled with the trailing axle.
5. The vehicle control system of claim 4 , wherein the AI neural network are configured to increase the adhesion value for the at least another one of the wheels by increasing the target slip value associated with the leading axle.
6. The vehicle control system of claim 1 , wherein the AI neural network are configured to repeatedly change the torque applied to the at least one of the axles based on the target slip value that changes.
7. The vehicle control system of claim 1 , wherein the AI neural network are configured to determine an average value of the adhesion between the wheels and the mute surface using the adhesion values.
8. A method comprising:
Determining, using one or more processors, adhesion values representative of adhesion between wheels of a vehicle and a mute surface based on angular speeds of the wheels;
calculating a target slip value for the wheels that are coupled with at least two different axles of the vehicle, the target slip value calculated using the adhesion values that were determined;
calculating the target slip value using an artificial intelligence (AI) neural network having artificial neurons arranged in layers and connected with each other by connections;
controlling, using the one or more processors, a torque applied to at least one of the axles based on the target slip value;
repeatedly changing the target slip value as the vehicle moves based on changes in the adhesion values to increase an average value of the adhesion values, changing the target slip values using the AI neural network;
receiving feedback regarding the target slip value that is calculated and changed by the artificial neurons; and
training the AI neural network by changing one or more connections between the artificial neurons in the AI neural network based on the feedback that is received.
9. The method of claim 8 , further comprising reducing a derivative of an average value of the adhesion value with respect to the target slip value by changing the torque applied to the at least one of the axles.
10. The method of claim 9 , further comprising:
determining an integral of the derivative of the average value of the adhesion value with respect to the target slip value, wherein the target slip value is changed based on the integral that is determined.
11. The method of claim 8 , wherein the at least two different axles include a leading axle and a trailing axle along a direction of movement of the vehicle, and changing the target slip value includes increasing the target slip value associated with the leading axle to clean the mute surface with at least one of the wheels coupled with the leading axle prior to travel of at least another one of the wheels that is coupled with the trailing axle.
12. The method of claim 11 , further comprising:
increasing the adhesion value for the at least another one of the wheels by increasing the target slip value associated with the leading axle.
13. The method of claim 8 , wherein controlling the torque includes repeatedly changing the torque applied to the at least one of the axles based on the target slip value that changes.
14. The method of claim 8 , wherein the adhesion value is determined as an average value of the adhesion between the wheels and the mute surface as the adhesion value.
15. A system comprising:
one or more processors configured to determine adhesion values representative of adhesion between wheels of a vehicle and a mute surface based on angular speeds of the wheels,
an artificial intelligence (AI) neural network having artificial neurons arranged in layers and connected with each other by connections configured to calculate a target slip value for the wheels that are coupled with at least two different axles of the vehicle using the adhesion values that were determined,
are the AI neural network configured to repeatedly change the target slip value as the vehicle moves based on changes in the adhesion values to increase an average value of the adhesion value,
the artificial neurons configured to receive feedback regarding the target slip value that is calculated,
the AI neural network configured to be trained by changing one or more connections between the artificial neurons in the AI neural network based on the feedback that is received,
the one or more processors configured to control a torque applied to at least one of the axles based on the target slip value,
wherein the wheels include a leading wheel and a trailing wheel along a direction of movement of the vehicle, the AI neural network configured to increase the target slip value that is calculated associated with the leading wheel to clean the route surface with the leading wheel prior to travel of the trailing wheel.
16. The system of claim 15 , wherein the AI neural network are configured reduce a derivative of the average value of the adhesion values with respect to the target slip value by changing the torque applied to the at least one of the axles.
17. The system of claim 16 , wherein the AI neural network are configured to determine an integral of the derivative of the average value of the adhesion values with respect to the target slip value, the AI neural network configured to determine or change the target slip value based on the integral that is determined.
18. The system of claim 15 , wherein the AI neural network are configured to increase the adhesion value for the trailing by increasing the target slip value associated with the leading wheel.
19. The system of claim 15 , wherein the AI neural network also are configured to repeatedly change the torque applied to the at least one of the axles based on the target slip value that changes.
20. The system of claim 15 , further comprising a low-pass filter configured to reduce noise present outside a useful frequency band when determining the adhesion value.