Line qualification with neural networks
AssigneeTeradyne, Inc.
InventorLee F. Holeys
Life of this patent
Estimating the term…
Abstract
A method tests a subscriber line. The method includes determining values of electrical line features from electrical measurements on the subscriber line and processing a portion of the values of the electrical features with a neural network. The neural network predicts whether the line qualifies to support one or more preselected data services from the portion of the values.
Claims39
A method of testing a subscriber line, comprising:
- determining values of electrical line features from electrical measurements on the subscriber line;
- and
- processing a portion of the values of the electrical features with a neural network that predicts whether the line qualifies to support one or more preselected data services from the portion of the values,
- wherein the measurements are one-ended electrical measurements, and further comprising
- compressing a vector whose components are the electrical features into a subspace;
- and
- predicting whether the line qualifies based on the compressed vector.
A method of constructing a test for qualifying subscriber lines for data transmissions, comprising:
- obtaining electrical feature and qualification data for sample lines of a training set;
- determining parameters defining a neural network from the data of the training set, the network is configured to use values of electrical properties of a subscriber line to predict whether the subscriber line qualifies for one or more preselected data services, wherein determining comprises finding a projected subspace in a space of vectors, components of each vector being the electrical features of an associated one of the sample lines.
A data storage medium storing a computer executable program for a method of testing a subscriber line, the method comprising:
- determining values of electrical line features from electrical measurements on the subscriber line;
- and
- processing a portion of the values of the electrical features with a neural network that predicts whether the line qualifies to support one or more preselected data services from the portion of the values;
- compressing a vector whose components are the electrical features into a subspace;
- and
- predicting whether the line qualifies based on the compressed vector.
A method of testing a subscriber line, comprising:
- determining values of electrical features of the line from electrical measurements on the subscriber line;
- and
- forming a vector having the values as components;
- determining whether the vector is a member of a cluster of feature vectors, feature vectors of the cluster being associated with a sample lines of a training set;
- and
- predicting whether the line qualifies for a data service based in part on the determination of cluster membership, and
- projecting out a portion of the components of the vector having features of the subscriber line as components;
- and
- wherein the act of determining membership is based on the compressed vector.
Dependencies are inferred from the claim text.