Norfolk Southern Corporation

Mitsubishi Electric Research Laboratories, Inc.

Wei Liu

Ilya A Lavrik

Mahbod Amouie

Gongli Duan

Evan T Gebhardt

Myles Grayson Akin

Tianchen Wang

Mayuresh Manoj Sardesai

Chunjie Duan

Yige Wang

Toshiaki Koike-Akino

Ramesh Annavajjala

A computer-implemented method in which one or more processing devices perform operations may include obtaining a field image of a railcar collected from a field camera system and applying a machine-learning algorithm to the field image to generate a machine-learning algorithm output. The method may also include performing a post-processing operation on the machine-learning algorithm output to generate a filtered machine-learning algorithm output. Further, the method may include detecting a defect of the railcar using the filtered machine-learning algorithm output.

Machine-learning framework for detecting defects or conditions of railcar systems

Systems, devices, media, and methods are presented for training a predictive model to detect objects in digital images, such as detecting whether a cotter key is present or absent in a digital photograph of a rail car coupler. The training system includes curating a plurality of training datasets, each including a number of raw images, together with a number of adjusted, augmented, and duplicate images. The predictive model includes a localization algorithm and an ensemble of models for classification. The localization algorithm is a deep convolutional neural network (CNN) which identifies a region of interest. One of more of the deep CNN classification models generates a plurality of candidate regions associated with each region of interest, thereby generating a large number of additional regions useful for training. In use, the trained predictive model is part of a detection and notification system that processes new images from the field and broadcasts a notice when an anomaly is detected.

Two-stage deep learning framework for detecting the condition of rail car coupler systems

Data are communicated in a wireless network, between a transmitter and a receiver. The transmitter estimates a first channel response between the receiver and the transmitter, and generates a first key based on the first channel response. The data are encoded at the transmitter using a rate-adaptive code to produce encoded data, and scrambled using the first key before broadcasting. Subsequently, the receiver can estimate a second channel response to generate a second key to be used to descramble the broadcast data.

Secure wireless communication using rate-adaptive codes

The first and second nodes in a wireless network estimate first and second channel response. The first node quantizes the first channel response to produce a first bit sequence, and a feed-forward message, which is transmit as a feed-forward message to the second node. The second node quantizes the second channel response using the feed-forward message to produce and an estimate of the first bit sequence, a second bit sequence and a feed-back message, which is transmitted to the first node. Then, the first and second nodes delete bits in the respective bit sequences using the feed-back and feed-forward message to generate first and second private keys with low bit mismatch rate.

Method for generating private keys in wireless networks

Growing community of inventors

Syracuse, NY, United States of America

Wei Liu