ONLINE MACHINE LEARNING FOR AUTONOMOUS EARTH MOVING VEHICLE CONTROL

    公开(公告)号:US20220413495A1

    公开(公告)日:2022-12-29

    申请号:US17558845

    申请日:2021-12-22

    IPC分类号: G05D1/00 G06N20/00 G05D1/02

    摘要: An autonomous earth moving system can determine a desired state for a portion of the EMV including at least one control surface. Then the EMV selects a set of control signals for moving the portion of the EMV from the current state to the desired state using a machine learning model trained to generate control signals for moving the portion of the EMV to the desired state based on the current state. After the EMV executes the selected set of control signals, the system measures an updated state of the portion of the EMV. In some cases, this updated state of the EMV is used to iteratively update the machine learning model using an online learning process.

    Obstacle detection and manipulation by a vehicle within a dig site

    公开(公告)号:US10761537B1

    公开(公告)日:2020-09-01

    申请号:US15996408

    申请日:2018-06-01

    IPC分类号: G05D1/02 G01C21/20

    摘要: An autonomous or semi-autonomous excavation vehicle is capable of determining a route between a start point and an end point in a site and navigating over the route. Sensors mounted on the excavation vehicle collect any or more of spatial, imaging, measurement, and location data to detect an obstacle between two locations within the site. Based on the collected data and identified obstacles, the excavation vehicle generates unobstructed routes circumventing the obstacles, obstructed routes traveling through the obstacles, and instructions for removing certain modifiable obstacles. The excavation vehicle determines and selects the shortest route of the unobstructed and obstructed route and navigates over the selected path to move within the site.

    ONLINE MACHINE LEARNING FOR CALIBRATION OF AUTONOMOUS EARTH MOVING VEHICLES

    公开(公告)号:US20220412057A1

    公开(公告)日:2022-12-29

    申请号:US17748999

    申请日:2022-05-19

    IPC分类号: E02F9/26 G05D1/00 G06N3/08

    摘要: In some implementations, the EMV uses a calibration to inform autonomous control over the EMV. To calibrate an EMV, the system first selects a calibration action comprising a control signal for actuating a control surface of the EMV. Then, using a calibration model comprising a machine learning model trained based on one or more previous calibration actions taken by the EMV, the system predicts a response of the control surface to the control signal of the calibration action. After the EMV executes the control signal to perform the calibration action, the EMV system monitors the actual response of the control signal and uses that to update the calibration model based on a comparison between the predicted and monitored states of the control surface.