Entity

Time filter

Source Type

Hughes Springs, MD, United States

Beidas B.F.,Advanced Development Group
IEEE Transactions on Communications | Year: 2011

The urgent objective of transmitting high data rates over satellite, coupled with the challenge to maximize satellite mass efficiency, has necessitated that multiple carriers share the same transponder high-power amplifier (HPA). This paper presents analytical framework that characterizes the resulting intermodulation distortion (IMD) by utilizing Volterra series representation to account for the memory within the carrier itself and those associated with other carriers. Also provided is analytical evaluation of nonlinear IMD which involves computing statistical averages of higher-order products of Volterra series containing complex-valued symbols from multiple carriers. Using this theoretical characterization, novel algorithms are developed to overcome IMD in highly distortion-limited environments by employing the powerful Turbo equalization method with linear minimum mean-squared error criterion. Further, the solution is adaptive so compensation does not require prior knowledge of the HPA characteristics and can be rapidly responsive to variations in the environment. Through extensive simulations, it is shown that the proposed multicarrier analysis and Turbo Volterra techniques can be used to substantially remove IMD resulting from operating the nonlinear transponder HPA, shared by multiple carriers, near saturation. By applying more iterations of joint equalization and decoding, the solution can approach the ideal performance when feeding back correct decisions. © 2011 IEEE. Source


Beidas B.F.,Advanced Development Group
IEEE Transactions on Communications | Year: 2016

Highly-efficient operation of communication systems requires effective compensation of nonlinear distortion with memory. The main contributor to nonlinearity is the high-power amplifier (HPA) when operated close to saturation. This results in two major detrimental effects: spectral regrowth causing interference in adjacent frequency bands and in-band distortion in the form of constellation warping and clustering. This paper introduces a novel family of adaptive digital signal predistortion schemes that successively modifies the HPA input to drive nonlinear distortion with memory toward zero. This family of schemes is capable of suppressing the spectral regrowth and in-band distortion simultaneously, while keeping the HPA operating efficiently close to saturation. In addition, the proposed solution offers the system designer a beneficial tunability feature to select the levels of suppression. Reduced-complexity Volterra model is adopted and is implemented on-The-fly to cope with systems with high degree of nonlinearity and large memory span. Furthermore, the proposed schemes are made adaptive by applying stochastic gradient method offline during training phase to effectively deal with nonlinear systems whose characteristics are unknown a priori. Extensive computer simulations demonstrate that the proposed adaptive family of predistortion schemes approaches the performance of the perfectly predistorted solution, and can considerably outperform techniques based on adaptive inverse, commonly adopted in the literature. © 1972-2012 IEEE. Source


Bowen R.M.,Rochester Institute of Technology | Sahin F.,Rochester Institute of Technology | Radomski A.,Advanced Development Group
Computers and Electrical Engineering | Year: 2016

We propose an application of specific machine learning techniques capable of evaluating systemic health of a Radio Frequency (RF) power generator. System signatures or fingerprints are collected from multivariate time-series data samples of sensor values under typical operational loads. These fingerprints are transformed into feature vectors using standard scaling/translation methods and the Fast Fourier Transform (FFT). The number of features per fingerprint are reduced by banding neighboring features and Principal Component Analysis (PCA). The reduced feature vectors are used with the Expectation Maximization (EM) algorithm to learn parameters for a Gaussian Mixture Model (GMM) to represent normal operation. One-class classification of normal fingerprints is achieved by thresholding the likelihood of a fingerprint feature vectors. Fingerprints were collected from normal operational conditions and seeded non-normal conditions. Preprocessing methods and algorithmic parameters have been selected using an iterative grid search. Average robust true positive rate achieved was 94.76% and best specificity reported is 86.56%. © 2016 Elsevier Ltd. Source


Beidas B.F.,Advanced Development Group
Proceedings - IEEE Military Communications Conference MILCOM | Year: 2011

The urgent objective of transmitting high data rates over satellite, coupled with the challenge to maximize satellite mass efficiency, has necessitated that multiple carriers share the same transponder high-power amplifier (HPA). This paper presents analytical framework that characterizes the resulting intermodulation distortion (IMD) by utilizing Volterra series representation to account for the memory within the carrier itself and those associated with other carriers. Using this theoretical characterization, novel algorithms are developed to overcome IMD in highly distortion-limited environments by employing the powerful Turbo equalization method with linear minimum mean-squared error criterion. Further, the solution is adaptive so compensation does not require prior knowledge of the HPA characteristics and can be rapidly responsive to variations in the environment. Through extensive simulations, it is shown that the proposed multicarrier analysis and Turbo Volterra techniques can be used to substantially remove IMD resulting from operating the nonlinear transponder HPA, shared by multiple carriers, near saturation. By applying more iterations of joint equalization and decoding, the solution can approach the ideal performance when feeding back correct decisions. © 2011 IEEE. Source


News Article | April 6, 2016
Site: http://boingboing.net

"The way we do technology development here is really hand-in-hand with the creative goals,” says (Lucasfilm CTO Rob) Bredow. “The R&D is always in service to the story.” For example, to port the Millennium Falcon from the Star Wars film universe into the interactive realm, the Advanced Development Group engineers first had to figure out how the VR hardware could render the massive 3D model in just milliseconds, compared with hours or days for a film shot. Then Skywalker Sound built a surround system that realistically rumbles and whooshes as a Corellian starship should. Meanwhile, game designers and the storytellers hashed out the most compelling way for a Jedi-in-training (you) to battle an army of Stormtroopers with a lightsaber.

Discover hidden collaborations