Online machine learning for determining soil parameters
AssigneeBuilt Robotics Inc.
InventorsGaurav Jitendra Kikani, Noah Austen Ready-Campbell, Andrew Xiao Liang, Joonhyun Kim
Life of this patent
Estimating the term…
Abstract
When an EMV performs an action comprising moving a tool of the EMV through soil or other material, the EMV can measure a current speed of the tool through the material and a current kinematic pressure exerted on the tool by the material. Using the measured current speed and kinematic pressure, the EMV system can use a machine learned model to determine one or more soil parameters of the material. The EMV can then make decisions based on the soil parameters, such as by selecting a tool speed for the EMV based on the determined soil parameters.
Claims20
A method comprising:
- accessing data representative of historical movements of tools of one earth moving vehicle (EMV) through a material;
- generating a training set of data based on the accessed data, the training set of data describing one or more parameters of the material based on, for each tool, a speed and kinematic pressure associated with the tool as the tool moves through the material;
- and
- training a machine-learning model using the generated training set of data, the machine-learning model configured to output a target movement of the tool of the target EMV based on a predicted set of current parameters of a current material that the tool is moving through and one or more of:
- a set of movements performed by the tool within a preceding time interval and movements performed by tools of other EMVs.
A non-transitory computer-readable storage medium storing executable instructions that, when executed by a hardware processor, cause the hardware processor to:
- access data representative of historical movements of tools of one earth moving vehicle (EMV) through a material;
- generate a training set of data based on the accessed data, the training set of data describing one or more parameters of the material based on, for each tool, a speed and kinematic pressure associated with the tool as the tool moves through the material;
- and
- train a machine-learning model using the generated training set of data, the machine-learning model configured to output a target movement of the tool of the target EMV based on a predicted set of current parameters of a current material that the tool is moving through and one or more of:
- a set of movements performed by the tool within a preceding time interval and movements performed by tools of other EMVs.
A method comprising:
- accessing data representative of historical movements of tools of one earth moving vehicle (EMV) through a material;
- accessing a machine-learning model, wherein the machine-learning model is trained using a training set of data generated based on the accessed data, the training set of data describing parameters of the material based on a speed and kinematic pressure associated with a tool as the tool moves through the material and one or more of:
- a set of movements performed by the tool within a preceding time interval and movements performed by tools of other EMVs;
- and
- inputting a set of current parameters of a current material that the tool is moving through to the machine-learning model to output a target movement of a tool of a target EMV.
Dependencies are inferred from the claim text.