Systems and methods for managing assignments of tasks for mining equipment using machine learning

Patent metadata

Field	Value
Publication number	US-12158765-B2
Application number	US-202217652926-A
Country	US
Kind code	B2
Filing date	Feb 28, 2022
Priority date	Feb 28, 2022
Publication date	Dec 3, 2024
Grant date	Dec 3, 2024

Abstract

Systems and methods are disclosed for managing task assignments for a fleet of haul trucks at a mine site. An assignment engine may: receive state data for a haul truck including haul weight data and second state data for the mine site that is indicative of a plurality of available tasks and associated task material weight data; assign a task to the haul truck by inputting the state data into a trained reinforcement-learning model, wherein: the model has been trained to learn an assignment policy that optimizes a reward function, such that the learned policy accounts for vehicle performance variance due to changing vehicle haul weight and/or road conditions; and cause the at least one haul truck to be operated according to the at least one task assignment.

First claim

What is claimed is:

1. A computer-implemented method for managing task assignments for a fleet of haul trucks at a mine site, comprising:
receiving, via one or more processors of an assignment engine system, first state data that includes state information for at least one haul truck of a fleet used to operate the mine site, the state information for the at least one haul truck including haul weight data;
receiving via the one or more processors of the assignment engine system, second state data that includes state information for the mine site, wherein the state information for the mine site is indicative of a plurality of tasks available in the mine site and task material weight data associated with the plurality of tasks;
assigning, via the one or more processors of the assignment engine system, at least one task from amongst the plurality of tasks to the at least one haul truck of the fleet by inputting the first state data and the second state data into a trained reinforcement-learning model, wherein:
the trained reinforcement-learning model has been trained, based on training first state data and training second state data, to learn an assignment policy that optimizes a reward function for the mine site, such that the trained reinforcement-learning model is configured to apply the learned assignment policy to the input first state data and second state data to select the at least one task; and
the training first state data includes training haul truck haul weight data, and the training second state data includes training task material weight data, such that the learned policy accounts for haul truck performance variance due to changing haul truck haul weight; and

causing, via the one or more processors of the assignment engine system, the at least one haul truck to be operated according to the at least one task assignment.

2. The computer-implemented method of claim 1 , wherein the state information for the mine site includes a material movement plan for the mine site and state information for one or more of a static charging station, a dynamic charging mechanism, a material source site, a material sink site, a material processing site, a weather condition associated with the mine site, or a road condition associated with the mine site.

3. The computer-implemented method of claim 1 , wherein the first state data includes a respective learning model for each haul truck in the fleet that is updated based on operating characteristics of the respective haul truck over time.

4. The computer-implemented method of claim 1 , wherein the at least one task assignment includes route instructions and speed instructions.

5. The computer-implemented method of claim 1 , wherein:
the at least one haul truck is a green-house-gas free vehicle;
the state information for the at least one haul truck further includes a state of charge for the at least one haul truck; and
the training first state data further includes training haul truck state of charge data, and the training second state data includes training task charge requirement data, such that the learned policy accounts for capacity for the at least one haul truck to complete the at least one task assignment.

6. The computer-implemented method of claim 1 , the at least one haul truck is an autonomous vehicle.

7. The computer-implemented method of claim 1 , wherein the reward function accounts for at least one of total cost for operating the mine site, total material processed by the mine site, power consumption of the mine site, balance of charge across the fleet, state of charge of the fleet, an amount of opportunistic charging for the fleet, or capacity for the at least one haul truck to complete the at least one task assignment.

8. A computer-implemented method for training a machine-learning model for managing task assignments for a fleet of haul trucks at a mine site, comprising:
obtaining, via one or more processors of an assignment engine system, first training state data that includes state information for at least one haul truck of a fleet used to operate a mine site, the state information for the at least one haul truck including haul weight data;
obtaining via the one or more processors of the assignment engine system, second training state data that includes state information for the mine site, wherein the state information for the mine site is indicative of a plurality of tasks available in the mine site and task material weight data associated with the plurality of tasks;
assigning, via the one or more processors of the assignment engine system, at least one task from amongst the plurality of tasks to the at least one haul truck of the fleet by inputting the first training state data and the second training state data into a reinforcement-learning model that includes a task assignment policy and a reward function;
causing, via the one or more processors of the assignment engine system, the at least one haul truck to be operated according to the at least one task assignment;
obtaining further first training state data and further second training state data from a period of time after the at least one task assignment is complete;
generating a score using the reward function based on the further first training state data and the further second training state;
updating the task assignment policy based on the generated score.

9. The computer-implemented method of claim 8 , wherein the method is iterated until one or more of:
the score at least meets a predetermined threshold;
a rate of change of the score across iterations is less than a further predetermined threshold;
a predetermined number of iterations have occurred; or
a predetermined period of time has passed.

10. The computer-implemented method of claim 8 , further comprising:
receiving, from a user via a user device, one or more objective selections; and
generating or tuning the reward function based on the one or more objective selections.

11. The computer-implemented method of claim 8 , wherein at least a portion of the fleet is simulated, such that at least a portion of the first training state data is simulated information.

12. The computer-implemented method of claim 8 , wherein at least a portion of the mine site is simulated, such that at least a portion of the second training state data is simulated information.

13. The computer-implemented method of claim 8 , wherein the state information for the mine site includes a material movement plan for the mine site and state information for one or more of a static charging station, a dynamic charging mechanism, a material source site, a material sink site, a material processing site, a weather condition associated with the mine site, or a road condition associated with the mine site.

14. The computer-implemented method of claim 8 , wherein the at least one task assignment includes route instructions and speed instructions.

15. The computer-implemented method of claim 1 , wherein:
the at least one haul truck is a green-house-gas free vehicle; and
the at least one haul truck is an autonomous vehicle.

16. The computer-implemented method of claim 8 , wherein the reward function is configured to account for at least one of total cost for operating the mine site, total material processed by the mine site, power consumption of the mine site, balance of charge across the fleet, state of charge of the fleet, or capacity for the at least one haul truck to complete the at least one task assignment.

17. A system for operating a mine site, comprising:
a fleet of haul trucks that are green-house-gas free an

Assignees

• Caterpillar Inc

Inventors

• Wulf Stefan Jacob
• Lane Cameron T
• Yun Andrew Seungyul
• Davis Greg
• Doherty Thomas Frank
• Collins Darryl

Classifications

• G05D1/661

  Docking at a base station (delivering or retrieving payloads G05D1/667) · CPC title

• B60L58/12

  responding to state of charge [SoC] · CPC title

• B60W2720/10

  Longitudinal speed · CPC title

• B60W2530/10

  Weight · CPC title

• B60W60/00256

  Delivery operations · CPC title