Learn More
Semantic understanding of environments is an important problem in robotics in general and intelligent autonomous systems in particular. In this paper, we propose a semantic segmentation algorithm which effectively fuses information from images and 3D point clouds. The proposed method incorporates information from multiple scales in an intuitive and(More)
We propose split-brain autoencoders, a straightforward modification of the traditional autoencoder architecture, for unsupervised representation learning. The method adds a split to the network, resulting in two disjoint sub-networks. Each sub-network is trained to perform a difficult task – predicting one subset of the data channels from another. Together,(More)
NOTICE This report was prepared as an account of work sponsored by an agency of the United States government. Neither the United States government nor any agency thereof, nor any of their employees, makes any warranty, express or implied, or assumes any legal liability or responsibility for the accuracy, completeness, or usefulness of any information,(More)
This paper describes the experimental results of a medium-voltage modular 35-MW drive for oil and gas applications realized by interconnecting voltage-source three-level converters equally sharing the motor current. Drives rated for tens of megawatts are increasingly needed as torque helpers for gas turbines in the oil and gas industry, and one envisions(More)
Tommaso Toma and is the Team Leader of the Variable Speed Drive System and New Product Introduction team, with GE Oil & Gas, in Florence, Italy. He is presently working on large VSDS applications and product development for the LNG and NG markets. Mr. Toma worked as an Electrical Design Engineer on GT/ST generation plants, and spent two years in a quality(More)
In this paper, a high-performance 22 MW/27 MVA test bench using a pumpback topology is introduced for testing electrical machines. It proposes a cost-effective concept that works reliably even with a weak grid. A paralleling concept of two three-level integrated gate-commutated thyristor (IGCT) converters is presented. I generates a five-level pulse pattern(More)
In this paper, we propose an algorithm to automatically identify window regions on exterior facing facades of buildings using interior 3D point cloud resulting from an ambulatory backpack sensor system, outfitted with multiple LiDAR sensors and cameras. We develop a set of discriminative features for the task, namely visual brightness, infrared opaqueness,(More)
We propose a deep learning approach for user-guided image colorization. The system directly maps a grayscale image, along with sparse, local user "hints" to an output colorization with a Convolutional Neural Network (CNN). Rather than using hand-defined rules, the network propagates user edits by fusing low-level cues along with high-level semantic(More)
This paper demonstrates the impact of modulation schemes on the power capability of a high-power converter with low pulse ratios. This integrated gate-commutated thyristor (IGCT) converter uses a five-level neutral-point-clamped H-bridge topology. It is concluded that phase-shifted carrier modulators are not attractive for such converters at low pulse(More)