Esmaeil Seraj | Georgia Institute of Technology (original) (raw)

Papers by Esmaeil Seraj

American Control Conference, 2020

Fighting wildfires is a precarious task, imperiling the lives of engaging firefighters and those ... more Fighting wildfires is a precarious task, imperiling the lives of engaging firefighters and those who reside in the fire's path. Firefighters need online and dynamic observation of the firefront to anticipate a wildfire's unknown characteristics, such as size, scale, and propagation velocity, and to plan accordingly. In this paper, we propose a distributed control framework to coordinate a team of unmanned aerial vehicles (UAVs) for a human-centered active sensing of wildfires. We develop a dual-criterion objective function based on Kalman uncertainty residual propagation and weighted multi-agent consensus protocol, which enables the UAVs to actively infer the wildfire dynamics and parameters, track and monitor the fire transition, and safely manage human firefighters on the ground using acquired information. We evaluate our approach relative to prior work, showing significant improvements by reducing the environment’s cumulative uncertainty residual in firefront coverage performance to support human-robot teaming for firefighting. We also demonstrate our method on physical robots in a mock firefighting exercise.

2020 American Control Conference, 2020

Fighting wildfires is a precarious task, imperiling the lives of engaging firefighters and those ... more Fighting wildfires is a precarious task, imperiling the lives of engaging firefighters and those who reside in the fire's path. Firefighters need online and dynamic observation of the firefront to anticipate a wildfire's unknown characteristics, such as size, scale, and propagation velocity, and to plan accordingly. In this paper, we propose a distributed control framework to coordinate a team of unmanned aerial vehicles (UAVs) for a human-centered active sensing of wildfires. We develop a dual-criterion objective function based on Kalman uncertainty residual propagation and weighted multi-agent consensus protocol, which enables the UAVs to actively infer the wildfire dynamics and parameters, track and monitor the fire transition, and safely manage human firefighters on the ground using acquired information. We evaluate our approach relative to prior work, showing significant improvements by reducing the environments cumulative uncertainty residual by more than 10 2 and 10 5 times in firefront coverage performance to support human-robot teaming for firefighting. We also demonstrate our method on physical robots in a mock firefighting exercise.

The purpose of this document is to help individuals use the "Essential Motor Cortex Signal Proces... more The purpose of this document is to help individuals use the "Essential Motor Cortex Signal Processing MATLAB Toolbox". The toolbox implements various methods for three major aspects of investigating human motor cortex from Neuroscience view point: (1) ERP estimation and quanti cation, (2) Cortical Functional Connectivity analysis and (3) EMG quanti cation. The toolbox { which is distributed under the terms of the GNU GENERAL PUBLIC LICENSE as a set of MATLAB routines { can be downloaded directly at the address:
http://oset.ir/category.php?dir=Tools

or from the public repository on GitHub, at address below:
https://github.com/EsiSeraj/ERP Connectivity EMG Analysis

The purpose of this toolbox is threefold: 1. Extract the event-related-potential (ERP) from preprocessed cerebral signals (i.e. EEG, MEG, etc.), identify and then quantify the event-related synchronization/desynchronization (ERS/ERD) events. Both time-course dynamics and time-frequency (TF) analyzes are included. 2. Measure, quantify and demonstrate the cortical functional connectivity (CFC) across scalp electrodes. These set of functions can also be applied to various types of cerebral signals (i.e. electric and magnetic). 3. Quantify electromyogram (EMG) recorded from active muscles during performing motor tasks.

Automatic identification and categorization of Alzheimer's patients and the ability to distinguis... more Automatic identification and categorization of Alzheimer's patients and the ability to distinguish between different levels of this disease can be very helpful to the research community in this field, since other non-automatic approaches are very time-consuming and are highly dependent on experts' experience. Herein, we propose the utility of cerebral instantaneous phase and envelope information in order to discriminate between Alzheimer's patients, MCI subjects and healthy normal individuals from functional magnetic resonance imaging (fMRI) data. To this end, after performing the region-of-interest (ROI) analysis on fMRI data, different features covering power, entropy and coherency aspects of data are derived from instantaneous phase and envelope sequences of ROI signals. Various sets of features are calculated and fed to a sequential forward floating feature selection (SFFFS) to choose the most discriminative and informative sets of features. A Student's t-test has been used to select the most relevant features from chosen sets. Finally, a K-NN classifier is used to distinguish between classes in a three-class categorization problem. The reported performance in overall accuracy using fMRI data of 111 combined subjects, is 80.1% with 80.0% Sensitivity to both Alzheimer's and Normal categories distinction and is comparable to the state-of-the-art approaches recently proposed in this regard. The significance of obtained results was statistically confirmed by evaluating through standard classification performance indicators. The obtained results illustrate that introduced analytic phase and envelope feature indexes derived from the ROI signals are significantly discriminative in distinguishing between Alzheimer’s patients and Normal healthy subject.

arXiv preprint arXiv:1903.06847, 2019

Wildfires are destructive and inflict massive, irreversible harm to victims’ lives and natural re... more Wildfires are destructive and inflict massive, irreversible harm to victims’ lives and natural resources. Researchers have proposed commissioning unmanned aerial vehicles (UAVs) to provide firefighters with real-time tracking information; yet, these UAVs are not able to reason about a fire’s track, including current location, measurement, and uncertainty, as well as propagation. We propose a model-predictive, probabilistically safe distributed control algorithm for human-robot collaboration in wildfire fighting. The proposed algorithm overcomes the limitations of prior work by explicitly estimating the latent fire propagation dynamics to enable intelligent, time-extended coordination of the UAVs in support of on-the ground human firefighters. We derive a novel, analytical bound that enables UAVs to distribute their resources and provides a probabilistic guarantee of the humans’ safety while preserving the UAVs’ ability to cover an entire fire.

—During the past decades, a great body of research has been devoted to automatic sleep stage scor... more —During the past decades, a great body of research has been devoted to automatic sleep stage scoring using the electroencephalogram (EEG). However, the results are not yet satisfactory to be used as a standard procedure in clinical studies. In this study, using recent developments in robust EEG phase extraction, a novel set of EEG-based features containing the Shannon entropy of the instantaneous analytical form envelope and frequencies of the EEG are proposed for sleep stage scoring. The proposed feature set is used to construct a distributed decision-tree classifier, with binary K-nearest neighbor (KNN) classifiers at each decision node. The decision-tree structure is designed by brute-force-search over various combinations of the proposed feature set. The performance of the proposed approach is evaluated over two available sleep EEG datasets acquired using single-channel EEG. The first set contains 20 healthy young subjects containing equal number of male and female, and the second one has been acquired from 140 adult subjects from both genders, with sleep disorder. The performance of the proposed method is tested versus state-of-the-art classifiers. The results demonstrate that the proposed method, resulted in overall accuracies of 88.97% and 83.17% over the two datasets, respectively. Considering the high performance and simplicity of the proposed scheme, the method can be of interest for clinical sleep disorder studies.

Objective: The instantaneous phase (IP) and instantaneous frequency (IF) of the electroencephalog... more Objective: The instantaneous phase (IP) and instantaneous frequency (IF) of the electroencephalogram (EEG) are considered as notable complements for the EEG spectrum. The calculation of these parameters commonly includes narrow-band filtering, followed by the calculation of the signal’s analytical form. The calculation of the IP and IF is highly susceptible to the filter parameters and background noise level, especially in low analytical signal amplitudes. The objective of this study is to propose a robust statistical framework for EEG IP/IF estimation and analysis. Approach: Herein, a Monte Carlo estimation scheme is proposed for the robust estimation of the EEG IP and IF. It is proposed that any EEG phase-related inference should be reported as an average with confidence intervals obtained by repeating the IP and IF estimation under infinitesimal variations (selected by an expert), in algorithmic parameters such as the filter’s bandwidth, center frequency and background noise level. In the second part of the paper, a stochastic model consisting of the superposition of narrow-band foreground and background EEG is used to derive analytically probability density functions of the instantaneous envelope (IE) and IP of EEG signals, which justify the proposed Monte Carlo scheme. Main results: The instantaneous analytical envelope of the EEG, which has been empirically used in previous studies, is shown to have a fundamental impact on the accuracy of the EEG phase contents. It is rigorously shown that the IP/IF estimation quality highly depends on the IE and any phase/frequency interpretations in low IE are statistically unreliable and require a hypothesis test. Significance: The impact of the proposed method on previous studies, including time-domain phase synchrony, phase resetting, phase locking value and phase amplitude coupling are studied with examples. The findings of this research can set forth new standards for EEG phase/frequency estimation and analysis techniques.

Objective: In this study, a robust method is developed for frequency specific electroencephalogra... more Objective: In this study, a robust method is developed for frequency specific electroencephalogram (EEG) phase extraction using the analytic representation of the EEG. Based on recent theoretical findings in this area, it is shown that some of the phase variations—previously associated to the brain response—are systematic side-effects of the methods used for EEG phase calculation, especially during low analytical amplitude segments of the EEG.
Approach: With this insight, the proposed method generates randomized ensembles of the EEG phase using minor perturbations in the zero-pole loci of narrow-band filters, followed by phase estimation using the signal’s analytical form and ensemble averaging over the randomized ensembles to obtain a robust EEG phase and frequency. This Monte Carlo estimation method is shown to be very robust to noise and minor changes of the filter parameters and reduces the effect of fake EEG phase jumps, which do not have a cerebral origin.
Main results: As proof of concept, the proposed method is used for extracting EEG phase features for a brain computer interface (BCI) application. The results show significant improvement in classification rates using rather simple phase-related features and a standard K-nearest neighbors and random forest classifiers, over a standard BCI dataset.
Significance: The average performance was improved between 4–7% (in absence of additive noise) and 8–12% (in presence of additive noise). The significance of these improvements was statistically confirmed by a paired sample t-test, with 0.01 and 0.03 p-values, respectively. The proposed method for EEG phase calculation is very generic and may be applied to other EEG phase-based studies.

arXiv preprint arXiv:1610.02249, 2018

This document is meant to help individuals use the Cerebral Signal Phase Analysis toolbox which i... more This document is meant to help individuals use the Cerebral Signal Phase Analysis toolbox which implements different methods for estimating the instantaneous phase and frequency of a signal and calculating some related popular quantities. The toolbox - which is distributed under the terms of the GNU GENERAL PUBLIC LICENSE as a set of MATLAB R routines - can be downloaded at the address:

http://oset.ir/category.php?dir=Tools
.
The purpose of this toolbox is to calculate the instantaneous phase and frequency sequences of cerebral signals (EEG, MEG, etc.) and some related popular features and quantities in brain studies and Neuroscience such as Phase Shift, Phase Resetting, Phase Locking Value (PLV), Phase Difference and more, to help researchers in these fields.

Motor imagery based brain-computer interface (BCI) systems translate the motor-intention of an in... more Motor imagery based brain-computer interface (BCI) systems translate the motor-intention of an individual into a control signal. For this BCI system, most of previous studies are based on power changes of mu and beta rhythms. In this paper, we employ both phase and envelope features of EEG signals to cover a comprehensive set of required informa tion for intention detection. For this purpose we use narrow band channelization in combination with a recently proposed Monte Carlo based statistical approach for EEG instantaneous parameters estimation known as transfer function perturbation (TFP) to calculate time/frequency measures of these parameters. The estimated time/frequency measures of instantaneous envelope (IE) and instantaneous phase (IP) using TFP are then utilized to elicit two set of robust features containing Shannon entropy of IE and phase lag index of IP. Extracted features are trained and tested by a KNN classifier to discriminate between classes in a three class motor imagery based BCI system. Results show that the combination of proposed robust analytic phase and envelope features outperforms either solely phase or envelope features and improves the classification rate in three class BCI problems.

Within the neuroscience and computational neuroscience communities, applications such as evaluati... more Within the neuroscience and computational neuroscience communities, applications such as evaluating different cognitive responses of the brain, brain-computer interface (BCI) systems and brain connectivity studies have increasingly been using EEG phase information during the past few decades. The utility of EEG phase can be directly linked to the neural propagation and synchronized firing of neuronal populations during different cognitive states of the brain. Nevertheless, it has previously been shown that phase of narrow-band (frequency specific) foreground EEG (desired) is prone to contain fake spikes and variations (unrelated to brain activity) in the presence of background spontaneous EEG and low SNRs of foreground EEG (the low-amplitude analytic signals or LAAS problem). Accordingly, extracting the instantaneous EEG phase sequence for further utilization upon interpreting the cognitive states of the brain using phase related quantities, such as instantaneous frequency, phase shift, phase resetting and phase locking value, is a very sensitive and rigorous process. In this study, a simple procedure is proposed to illustrate the effects of LAAS problem on the utility of EEG phase related quantities in aforementioned applications, also to investigate the reliability of interpretations of the brain's cognitive states based on such quantities. Results show that, without a proper and effective solution strategy, such potential flaws lead to incorrect physiological and pathological interpretations.

arXiv preprint arXiv:1612.04295, 2018

The human brain is ultimately responsible for all thoughts and movements that the body produces. ... more The human brain is ultimately responsible for all thoughts and movements that the body produces. This allows humans to successfully interact with their environment. If the brain is not functioning properly many abilities of human can be damaged. The goal of cerebral signal analysis is to learn about brain function.

The idea that distinct areas of the brain are responsible for specific tasks, the functional segregation, is a key aspect of brain function. Functional integration is an important feature of brain function, it is the concordance of multiple segregated brain areas to produce a unified response. There is an amplified feedback mechanism in the brain called reentry which requires specific timing relations. This specific timing requires neurons within an assembly to synchronize their firing rates. This has led to increased interest and use of phase variables, particularly their synchronization, to measure connectivity in cerebral signals. Herein, we propose a comprehensive review on concepts and methods previously presented for assessing cerebral synchrony, with focus on phase synchronization, as a tool for brain connectivity evaluation.

Research findings show that several diseases can be detected by quantitative analysis of sleep si... more Research findings show that several diseases can be detected by quantitative analysis of sleep signals.

Drafts by Esmaeil Seraj

arXiv, 2020

The purpose of this tutorial is to help individuals use the \underline{FireCommander} game enviro... more The purpose of this tutorial is to help individuals use the \underline{FireCommander} game environment for research applications. The FireCommander is an interactive, probabilistic joint perception-action reconnaissance environment in which a composite team of agents (e.g., robots) cooperate to fight dynamic, propagating firespots (e.g., targets). In FireCommander game, a team of agents must be tasked to optimally deal with a wildfire situation in an environment with propagating fire areas and some facilities such as houses, hospitals, power stations, etc. The team of agents can accomplish their mission by first sensing (e.g., estimating fire states), communicating the sensed fire-information among each other and then taking action to put the firespots out based on the sensed information (e.g., dropping water on estimated fire locations). The FireCommander environment can be useful for research topics spanning a wide range of applications from Reinforcement Learning (RL) and Learning from Demonstration (LfD), to Coordination, Psychology, Human-Robot Interaction (HRI) and Teaming. There are four important facets of the FireCommander environment that overall, create a non-trivial game: (1) Complex Objectives: Multi-objective Stochastic Environment, (2)Probabilistic Environment: Agents' actions result in probabilistic performance, (3) Hidden Targets: Partially Observable Environment and, (4) Uni-task Robots: Perception-only and Action-only agents. The FireCommander environment is first-of-its-kind in terms of including Perception-only and Action-only agents for coordination. It is a general multi-purpose game that can be useful in a variety of combinatorial optimization problems and stochastic games, such as applications of Reinforcement Learning (RL), Learning from Demonstration (LfD) and Inverse RL (iRL).

arXiv, 2020

American Control Conference, 2020

2020 American Control Conference, 2020

Fighting wildfires is a precarious task, imperiling the lives of engaging firefighters and those ... more Fighting wildfires is a precarious task, imperiling the lives of engaging firefighters and those who reside in the fire's path. Firefighters need online and dynamic observation of the firefront to anticipate a wildfire's unknown characteristics, such as size, scale, and propagation velocity, and to plan accordingly. In this paper, we propose a distributed control framework to coordinate a team of unmanned aerial vehicles (UAVs) for a human-centered active sensing of wildfires. We develop a dual-criterion objective function based on Kalman uncertainty residual propagation and weighted multi-agent consensus protocol, which enables the UAVs to actively infer the wildfire dynamics and parameters, track and monitor the fire transition, and safely manage human firefighters on the ground using acquired information. We evaluate our approach relative to prior work, showing significant improvements by reducing the environments cumulative uncertainty residual by more than 10 2 and 10 5 times in firefront coverage performance to support human-robot teaming for firefighting. We also demonstrate our method on physical robots in a mock firefighting exercise.

or from the public repository on GitHub, at address below:
https://github.com/EsiSeraj/ERP Connectivity EMG Analysis

arXiv preprint arXiv:1903.06847, 2019

arXiv preprint arXiv:1610.02249, 2018

arXiv preprint arXiv:1612.04295, 2018

arXiv, 2020