Silvestri, Simone

Description: 

The goal of this project is to create a system for identifying electrical loads on a system given their power consumption in real-time without prior training and when relying on an unreliable, non-expert oracle for labeling samples. This can help with fine-grained energy management in households and load balancing.

Computational need:

We are developing a comparison approach to our KAN method utilizing Convolutional Neural Networks which requires significant processing power to train.

Collaborators:

Dr. Atieh Khamesi (PostDoc at UKY, and Missouri S&T)

Students:

Jackson B Codispoti

Software Used:

Python/Tensorflow

Grants:

Title: CAREER: Energy Management for Smart Residential Environments through Human-in-the-loop Algorithm Design
Total value: $524,989
Sponsor: National Science Foundation (NSF)
Duration: 5 years
Role: PI
Status: Funded

Title: Crosslayer Optimization of Energy and Cost through Unified Modeling of User Behavior and Storage in Multiple Buildings
Total value: $349,340 - including $16,000 REU supplement
Sponsor: National Science Foundation (NSF)
Duration: 3 years, Sept 2019-Aug 2022
Percent involvement: 1% Academic Salary, 0.5 Summer salary per year
PI: Simone Silvestri
Co-PI: Dan Ionel

Title: Integration of Social Behavioral Modeling for Smart Environments to Improve the Energy Efficiency of Smart Cities
Total value: $802,981
Sponsor: National Institute of Food and Agriculture (NIFA)
Duration: 3 years, Feb 2017-Feb 2021
Percent involvement: 1% Academic Salary, 1 Summer salary per year
PI: Simone Silvestri
Co-PI: Denise Baker, Jhi-Young Joo

Publications:

TBA

Next-generation challenged networks

Description: The goal of this project is to design easy-to-deploy, low-cost and sustainable network architectures and protocols for improved connectivity in challenging transient environments where the primary communication infrastructures are limited or non-existent, such as rural/remote area.

Computational need:

We are developing a network simulation tool in python programming language. The objective is to analyze the proposed network architectures and protocols on real and synthetic datasets through large-scale simulation experiments.

Collaborators:

Sajal K. Das (Missouri S&T, Rolla)

Students:

Vijay K. Shah (UK CS PhD student)

Jackson B Codispoti (Postdoc)

Software used:

Python

Funding:

Publications:

None

Disaster Response Networks

Description: In the event of a large-scale disaster, natural (e.g., earthquake) or man-made (e.g., terrorist attacks), a communication infrastructure is essential for prompt rescue and relief operations. However, the partial (or complete) impairment of primary communication and power infrastructures makes it difficult for the responders to assess the required logistics for recovery efforts, identify and extend rescue/relief to the survivors. In this project, the goal is to design novel topology and routing protocols for designing robust and energy-efficient communication networks, termed Disaster Response Networks; to enable information exchange between survivors and responders (until the primary infrastructures are restored).

Collaborators:

Sajal K. Das (Missouri S&T, Rolla)

Students:

Vijay K. Shah (UK CS PhD student)

Software used:

Python

Java

Opportunistic Network Simulator (ONE)

Funding: NATO SPS G4936 - Hybrid sensor networks for emergency critical scenarios

Publications:

Vijay K. Shah, Satyaki Roy, Simone Silvestri, and Sajal K. Das, CTR: A Cluster based Topological Routing for Disaster Response Network" , IEEE International Conference on Communications (ICC), 2017.