中科微点-全球专利检索

  1. Login
  2. Sign Up

Search Results

8 records (took 0.016 seconds)

    Autonomous UAV obstacle avoidance using machine learning from piloted UAV flights
    3.
    发明授权
    Autonomous UAV obstacle avoidance using machine learning from piloted UAV flights 有权

    公开(公告)号：US11087632B1

    公开(公告)日：2021-08-10

    申请号：US16884312

    申请日：2020-05-27

    Applicant: Amazon Technologies, Inc.

    Inventor： Pradeep Krishna Yarlagadda

    IPC: G08G5/00 ,  H04L29/08

    Abstract: A machine learning engine may correlate characteristics of obstacles identified during remotely piloted UAV flights with manual course deviations performed for obstacle avoidance. An obstacle detection application may access computer vision footage to determine notable characteristics (e.g. a direction of travel and/or velocity) of obstacles identified during the piloted UAV flights. A deviation characteristics application may access flight path information identify course deviations performed by a pilot in response to the obstacles. A machine learning engine may use the obstacle characteristic data and the deviation characteristics data as training data to generate an optimal course deviation model to use by an autopilot module to autonomously avoid obstacles during autonomous UAV flights. In creating the optimal deviation model, the training data may be processed by the machine learning engine to identify correlations between certain types of manual course deviations performed to avoid certain types of obstacles.

    Active sensor fusion systems and methods for object detection
    5.
    发明授权
    Active sensor fusion systems and methods for object detection 有权

    公开(公告)号：US11017513B1

    公开(公告)日：2021-05-25

    申请号：US16367857

    申请日：2019-03-28

    Applicant: Amazon Technologies, Inc.

    Inventor： Pradeep Krishna Yarlagadda ,  Jean-Guillaume Dominique Durand

    IPC: G06T5/50 ,  G06T7/70 ,  B64C39/02

    Abstract: Active sensor fusion systems and methods may include a plurality of sensors, a plurality of detection algorithms, and an active sensor fusion algorithm. Based on detection hypotheses received from the plurality of detection algorithms, the active sensor fusion algorithm may instruct or direct modifications to one or more of the plurality of sensors or the plurality of detection algorithms. In this manner, operations of the plurality of sensors or processing of the plurality of detection algorithms may be refined or adjusted to provide improved object detection with greater accuracy, speed, and reliability.

    Autonomous UAV obstacle avoidance using machine learning from piloted UAV flights
    6.
    发明授权
    Autonomous UAV obstacle avoidance using machine learning from piloted UAV flights 有权

    公开(公告)号：US10679509B1

    公开(公告)日：2020-06-09

    申请号：US15271107

    申请日：2016-09-20

    Applicant: Amazon Technologies, Inc.

    Inventor： Pradeep Krishna Yarlagadda

    IPC: G08G5/00 ,  H04L29/08

    Abstract: A machine learning engine may correlate characteristics of obstacles identified during remotely piloted UAV flights with manual course deviations performed for obstacle avoidance. An obstacle detection application may access computer vision footage to determine notable characteristics (e.g. a direction of travel and/or velocity) of obstacles identified during the piloted UAV flights. A deviation characteristics application may access flight path information identify course deviations performed by a pilot in response to the obstacles. A machine learning engine may use the obstacle characteristic data and the deviation characteristics data as training data to generate an optimal course deviation model to use by an autopilot module to autonomously avoid obstacles during autonomous UAV flights. In creating the optimal deviation model, the training data may be processed by the machine learning engine to identify correlations between certain types of manual course deviations performed to avoid certain types of obstacles.

    Identifying anomalous sensors
    7.
    发明授权
    Identifying anomalous sensors 有权

    公开(公告)号：US11544161B1

    公开(公告)日：2023-01-03

    申请号：US16428048

    申请日：2019-05-31

    Applicant: Amazon Technologies, Inc.

    Inventor： Pradeep Krishna Yarlagadda ,  Jean-Guillaume Dominique Durand

    IPC: G06F11/16 ,  G01D18/00 ,  G05D1/00 ,  B64C39/02

    Abstract: A sensor system may include first and second sensors configured to be coupled to a vehicle and generate respective first and second sensor signals indicative of operation of the vehicle. The sensor system may also include a sensor anomaly detector including an anomalous sensor model configured to receive the first and second sensor signals and determine that one or more of the first sensor or the second sensor is an anomalous sensor generating inaccurate sensor data. The sensor system may also be configured to identify one or more of the first sensor or the second sensor as the anomalous sensor generating inaccurate sensor data.

    Producing training data for machine learning
    8.
    发明授权
    Producing training data for machine learning 有权

    公开(公告)号：US11113567B1

    公开(公告)日：2021-09-07

    申请号：US16259783

    申请日：2019-01-28

    Applicant: Amazon Technologies, Inc.

    Inventor： Jean-Guillaume Durand ,  Pradeep Krishna Yarlagadda ,  Ishay Kamon ,  Francesco Callari

    IPC: G06K9/62 ,  G01S19/03 ,  G06N20/00

    Abstract: Described are systems and methods for generating training data that is used to train a machine learning system to detect moving objects represented in sensor data. The system and methods utilize position data received from a target vehicle to determine data points within sensor data that represents that target vehicle. For example, a station at a known location may receive Automatic Dependent Surveillance-Broadcast (“ADS-B”) data (position data) corresponding to a target vehicle that is within the field of view of a station sensor, such as a camera. The position data may then be correlated with the sensor data and projected into the sensor data to determine data points within the sensor data that represent the target vehicle. Those data points are then labeled to indicate the location, size, and/or shape of the target vehicle as represented in the sensor data, thereby producing training that may be provided to train a machine learning algorithm or system to detect moving objects, such as aircraft.

Patent Agency Ranking