公开(公告)号：US20160096272A1

公开(公告)日：2016-04-07

申请号：US14588168

申请日：2014-12-31

Applicant: BRAIN CORPORATION

Inventor： Andrew T. Smith ,  Vadim Polonichko

IPC: B25J9/16

CPC classification number: B25J9/163 ,  B25J9/1607 ,  B25J9/1666 ,  B25J9/1697 ,  G05D1/0088 ,  G05D1/0246 ,  G05D2201/02 ,  G06N3/00 ,  G06N3/008 ,  G06N3/049 ,  Y10S901/01 ,  Y10S901/03 ,  Y10S901/09 ,  Y10S901/47

Abstract: A random k-nearest neighbors (RKNN) approach may be used for regression/classification model wherein the input includes the k closest training examples in the feature space. The RKNN process may utilize video images as input in order to predict motor command for controlling navigation of a robot. In some implementations of robotic vision based navigation, the input space may be highly dimensional and highly redundant. When visual inputs are augmented with data of another modality that is characterized by fewer dimensions (e.g., audio), the visual data may overwhelm lower-dimension data. The RKNN process may partition available data into subsets comprising a given number of samples from the lower-dimension data. Outputs associated with individual subsets may be combined (e.g., averaged). Selection of number of neighbors, subset size and/or number of subsets may be used to trade-off between speed and accuracy of the prediction.

Abstract translation: 随机k最近邻（RKNN）方法可用于回归/分类模型，其中输入包括特征空间中最接近的k个训练样本。 RKNN过程可以利用视频图像作为输入，以便预测用于控制机器人的导航的马达命令。 在基于机器人视觉的导航的一些实现中，输入空间可以是高度尺寸和高度冗余的。 当视觉输入用另一种具有较少尺寸（例如，音频）特征的模态的数据进行增强时，视觉数据可能会压倒较低维度的数据。 RKNN过程可以将可用数据划分为包含来自较低维数据的给定数量样本的子集。 与各个子集相关联的输出可以组合（例如，平均）。 选择邻居数量，子集大小和/或子集数可用于在速度和预测精度之间进行权衡。

公开(公告)号：US09687984B2

公开(公告)日：2017-06-27

申请号：US14588168

申请日：2014-12-31

Applicant: BRAIN CORPORATION

Inventor： Andrew T. Smith ,  Vadim Polonichko

IPC: G05B19/04 ,  G05B19/18 ,  B25J9/16 ,  G05D1/00 ,  G05D1/02 ,  G06N3/00 ,  G06N3/04

CPC classification number: B25J9/163 ,  B25J9/1607 ,  B25J9/1666 ,  B25J9/1697 ,  G05D1/0088 ,  G05D1/0246 ,  G05D2201/02 ,  G06N3/00 ,  G06N3/008 ,  G06N3/049 ,  Y10S901/01 ,  Y10S901/03 ,  Y10S901/09 ,  Y10S901/47

Abstract: A random k-nearest neighbors (RKNN) approach may be used for regression/classification model wherein the input includes the k closest training examples in the feature space. The RKNN process may utilize video images as input in order to predict motor command for controlling navigation of a robot. In some implementations of robotic vision based navigation, the input space may be highly dimensional and highly redundant. When visual inputs are augmented with data of another modality that is characterized by fewer dimensions (e.g., audio), the visual data may overwhelm lower-dimension data. The RKNN process may partition available data into subsets comprising a given number of samples from the lower-dimension data. Outputs associated with individual subsets may be combined (e.g., averaged). Selection of number of neighbors, subset size and/or number of subsets may be used to trade-off between speed and accuracy of the prediction.