Current Search: Vosoughi, Azadeh (x)
View All Items
Pages
- Title
- On Distributed Estimation for Resource Constrained Wireless Sensor Networks.
- Creator
-
Sani, Alireza, Vosoughi, Azadeh, Rahnavard, Nazanin, Wei, Lei, Atia, George, Chatterjee, Mainak, University of Central Florida
- Abstract / Description
-
We study Distributed Estimation (DES) problem, where several agents observe a noisy version of an underlying unknown physical phenomena (which is not directly observable), and transmit a compressed version of their observations to a Fusion Center (FC), where collective data is fused to reconstruct the unknown. One of the most important applications of Wireless Sensor Networks (WSNs) is performing DES in a field to estimate an unknown signal source. In a WSN battery powered geographically...
Show moreWe study Distributed Estimation (DES) problem, where several agents observe a noisy version of an underlying unknown physical phenomena (which is not directly observable), and transmit a compressed version of their observations to a Fusion Center (FC), where collective data is fused to reconstruct the unknown. One of the most important applications of Wireless Sensor Networks (WSNs) is performing DES in a field to estimate an unknown signal source. In a WSN battery powered geographically distributed tiny sensors are tasked with collecting data from the field. Each sensor locally processes its noisy observation (local processing can include compression,dimension reduction, quantization, etc) and transmits the processed observation over communication channels to the FC, where the received data is used to form a global estimate of the unknown source such that the Mean Square Error (MSE) of the DES is minimized. The accuracy of DES depends on many factors such as intensity of observation noises in sensors, quantization errors in sensors, available power and bandwidth of the network, quality of communication channels between sensors and the FC, and the choice of fusion rule in the FC. Taking into account all of these contributing factors and implementing a DES system which minimizes the MSE and satisfies all constraints is a challenging task. In order to probe into different aspects of this challenging task we identify and formulate the following three problems and address them accordingly:1- Consider an inhomogeneous WSN where the sensors' observations is modeled linear with additive Gaussian noise. The communication channels between sensors and FC are orthogonal power and bandwidth-constrained erroneous wireless fading channels. The unknown to be estimated is a Gaussian vector. Sensors employ uniform multi-bit quantizers and BPSK modulation. Given this setup, we ask: what is the best fusion rule in the FC? what is the best transmit power and quantization rate (measured in bits per sensor) allocation schemes that minimize the MSE? In order to answer these questions, we derive some upper bounds on global MSE and through minimizing those bounds, we propose various resource allocation schemes for the problem, through which we investigate the effect of contributing factors on the MSE.2- Consider an inhomogeneous WSN with an FC which is tasked with estimating a scalar Gaussian unknown. The sensors are equipped with uniform multi-bit quantizers and the communication channels are modeled as Binary Symmetric Channels (BSC). In contrast to former problem the sensors experience independent multiplicative noises (in addition to additive noise). The natural question in this scenario is: how does multiplicative noise affect the DES system performance? how does it affect the resource allocation for sensors, with respect to the case where there is no multiplicative noise? We propose a linear fusion rule in the FC and derive the associated MSE in closed-form. We propose several rate allocation schemes with different levels of complexity which minimize the MSE. Implementing the proposed schemes lets us study the effect of multiplicative noise on DES system performance and its dynamics. We also derive Bayesian Cramer-Rao Lower Bound (BCRLB) and compare the MSE performance of our porposed methods against the bound.As a dual problem we also answer the question: what is the minimum required bandwidth of thenetwork to satisfy a predetermined target MSE?3- Assuming the framework of Bayesian DES of a Gaussian unknown with additive and multiplicative Gaussian noises involved, we answer the following question: Can multiplicative noise improve the DES performance in any case/scenario? the answer is yes, and we call the phenomena as 'enhancement mode' of multiplicative noise. Through deriving different lower bounds, such as BCRLB,Weiss-Weinstein Bound (WWB), Hybrid CRLB (HCRLB), Nayak Bound (NB), Yatarcos Bound (YB) on MSE, we identify and characterize the scenarios that the enhancement happens. We investigate two situations where variance of multiplicative noise is known and unknown. Wealso compare the performance of well-known estimators with the derived bounds, to ensure practicability of the mentioned enhancement modes.
Show less - Date Issued
- 2017
- Identifier
- CFE0006913, ucf:51698
- Format
- Document (PDF)
- PURL
- http://purl.flvc.org/ucf/fd/CFE0006913
- Title
- Biophysical Sources of 1/f Noises in Neurological Systems.
- Creator
-
Paris, Alan, Vosoughi, Azadeh, Atia, George, Wiegand, Rudolf, Douglas, Pamela, Berman, Steven, University of Central Florida
- Abstract / Description
-
High levels of random noise are a defining characteristic of neurological signals at all levels, from individual neurons up to electroencephalograms (EEG). These random signals degrade the performance of many methods of neuroengineering and medical neuroscience. Understanding this noise also is essential for applications such as real-time brain-computer interfaces (BCIs), which must make accurate control decisions from very short data epochs. The major type of neurological noise is of the so...
Show moreHigh levels of random noise are a defining characteristic of neurological signals at all levels, from individual neurons up to electroencephalograms (EEG). These random signals degrade the performance of many methods of neuroengineering and medical neuroscience. Understanding this noise also is essential for applications such as real-time brain-computer interfaces (BCIs), which must make accurate control decisions from very short data epochs. The major type of neurological noise is of the so-called 1/f-type, whose origins and statistical nature has remained unexplained for decades. This research provides the first simple explanation of 1/f-type neurological noise based on biophysical fundamentals. In addition, noise models derived from this theory provide validated algorithm performance improvements over alternatives.Specifically, this research defines a new class of formal latent-variable stochastic processes called hidden quantum models (HQMs) which clarify the theoretical foundations of ion channel signal processing. HQMs are based on quantum state processes which formalize time-dependent observation. They allow the quantum-based calculation of channel conductance autocovariance functions, essential for frequency-domain signal processing. HQMs based on a particular type of observation protocol called independent activated measurements are shown to be distributionally equivalent to hidden Markov models yet without an underlying physical Markov process. Since the formal Markov processes are non-physical, the theory of activated measurement allows merging energy-based Eyring rate theories of ion channel behavior with the more common phenomenological Markov kinetic schemes to form energy-modulated quantum channels. These unique biophysical concepts developed to understand the mechanisms of ion channel kinetics have the potential of revolutionizing our understanding of neurological computation.To apply this theory, the simplest quantum channel model consistent with neuronal membrane voltage-clamp experiments is used to derive the activation eigenenergies for the Hodgkin-Huxley K+ and Na+ ion channels. It is shown that maximizing entropy under constrained activation energy yields noise spectral densities approximating S(f) = 1/f, thus offering a biophysical explanation for this ubiquitous noise component. These new channel-based noise processes are called generalized van der Ziel-McWhorter (GVZM) power spectral densities (PSDs). This is the only known EEG noise model that has a small, fixed number of parameters, matches recorded EEG PSD's with high accuracy from 0 Hz to over 30 Hz without infinities, and has approximately 1/f behavior in the mid-frequencies. In addition to the theoretical derivation of the noise statistics from ion channel stochastic processes, the GVZM model is validated in two ways. First, a class of mixed autoregressive models is presented which simulate brain background noise and whose periodograms are proven to be asymptotic to the GVZM PSD. Second, it is shown that pairwise comparisons of GVZM-based algorithms, using real EEG data from a publicly-available data set, exhibit statistically significant accuracy improvement over two well-known and widely-used steady-state visual evoked potential (SSVEP) estimators.
Show less - Date Issued
- 2016
- Identifier
- CFE0006485, ucf:51418
- Format
- Document (PDF)
- PURL
- http://purl.flvc.org/ucf/fd/CFE0006485
- Title
- Prototype Development in General Purpose Representation and Association Machine Using Communication Theory.
- Creator
-
Li, Huihui, Wei, Lei, Rahnavard, Nazanin, Vosoughi, Azadeh, Da Vitoria Lobo, Niels, Wang, Wei, University of Central Florida
- Abstract / Description
-
Biological system study has been an intense research area in neuroscience and cognitive science for decades of years. Biological human brain is created as an intelligent system that integrates various types of sensor information and processes them intelligently. Neurons, as activated brain cells help the brain to make instant and rough decisions. From the 1950s, researchers start attempting to understand the strategies the biological system employs, then eventually translate them into machine...
Show moreBiological system study has been an intense research area in neuroscience and cognitive science for decades of years. Biological human brain is created as an intelligent system that integrates various types of sensor information and processes them intelligently. Neurons, as activated brain cells help the brain to make instant and rough decisions. From the 1950s, researchers start attempting to understand the strategies the biological system employs, then eventually translate them into machine-based algorithms. Modern computers have been developed to meet our need to handle computational tasks which our brains are not capable of performing with precision and speed. While in these existing man-made intelligent systems, most of them are designed for specific purposes. The modern computers solve sophistic problems based on fixed representation and association formats, instead of employing versatile approaches to explore the unsolved problems.Because of the above limitations of the conventional machines, General Purpose Representation and Association Machine (GPRAM) System is proposed to focus on using a versatile approach with hierarchical representation and association structures to do a quick and rough assessment on multitasks. Through lessons learned from neuroscience, error control coding and digital communications, a prototype of GPRAM system by employing (7,4) Hamming codes and short Low-Density Parity Check (LDPC) codes is implemented. Types of learning processes are presented, which prove the capability of GPRAM for handling multitasks.Furthermore, a study of low resolution simple patterns and face images recognition using an Image Processing Unit (IPU) structure for GPRAM system is presented. IPU structure consists of a randomly constructed LDPC code, an iterative decoder, a switch and scaling, and decision devices. All the input images have been severely degraded to mimic human Visual Information Variability (VIV) experienced in human visual system. The numerical results show that 1) IPU can reliably recognize simple pattern images in different shapes and sizes; 2) IPU demonstrates an excellent multi-class recognition performance for the face images with high degradation. Our results are comparable to popular machine learning recognition methods towards images without any quality degradation; 3) A bunch of methods have been discussed for improving IPU recognition performance, e.g. designing various detection and power scaling methods, constructing specific LDPC codes with large minimum girth, etc.Finally, novel methods to optimize M-ary PSK, M-ary DPSK, and dual-ring QAM signaling with non-equal symbol probabilities over AWGN channels are presented. In digital communication systems, MPSK, MDPSK, and dual-ring QAM signaling with equiprobable symbols have been well analyzed and widely used in practice. Inspired by bio-systems, we suggest investigating signaling with non-equiprobable symbol probabilities, since in bio-systems it is highly-unlikely to follow the ideal setting and uniform construction of single type of system. The results show that the optimizing system has lower error probabilities than conventional systems and the improvements are dramatic. Even though the communication systems are used as the testing environment, clearly, our final goal is to extend current communication theory to accommodate or better understand bio-neural information processing systems.
Show less - Date Issued
- 2017
- Identifier
- CFE0006758, ucf:51846
- Format
- Document (PDF)
- PURL
- http://purl.flvc.org/ucf/fd/CFE0006758