Title

Adaptive Learning for Raspberry Pi Controlled Smart Streetlights

Presenter Information

Patrick Toplikar

Department

Electrical and Computer Engineering

Major

Electrical Engineering

Research Advisor

Obafemi-Ajayi, Tayo

Advisor's Department

Electrical and Computer Engineering

Funding Source

Missouri State Cooperative Engineering Department Self

Abstract

For the project, a raspberry pi, and an Arduino control house lights in response to detecting cars with ultrasonic distance sensors. The project aims to demonstrate that street lights could save energy by adapting to traffic via neural networks in the raspberry pi, which would make the lights turn off when no cars are driving by at night, and flickering could be reduced. The project also investigates whether lights consume more energy when flickering, as compared to being powered constantly. In theory, neural networks reduce the energy consumed by smart-street lights, as well as flickering; ultrasonic distance sensors detect cars and send feedback to the neural networks, and the networks adapt the lights’ behavior based off of the learned traffic patterns. The project has successfully detected a car and turned a house lamp on and off with an ultrasonic distance sensor.

Biography

Patrick Toplikar is enrolled in the cooperative engineering program in Springfield, Missouri. He is pursuing an Electrical Engineering undergraduate degree from Missouri University of Science and Technology, as well as an Applied Physics degree from Missouri State University.

Research Category

Engineering

Presentation Type

Oral Presentation

Document Type

Presentation

Location

Carver Room

Presentation Date

17 Apr 2018, 11:30 am - 12:00 pm

This document is currently not available here.

Share

COinS
 
Apr 17th, 11:30 AM Apr 17th, 12:00 PM

Adaptive Learning for Raspberry Pi Controlled Smart Streetlights

Carver Room

For the project, a raspberry pi, and an Arduino control house lights in response to detecting cars with ultrasonic distance sensors. The project aims to demonstrate that street lights could save energy by adapting to traffic via neural networks in the raspberry pi, which would make the lights turn off when no cars are driving by at night, and flickering could be reduced. The project also investigates whether lights consume more energy when flickering, as compared to being powered constantly. In theory, neural networks reduce the energy consumed by smart-street lights, as well as flickering; ultrasonic distance sensors detect cars and send feedback to the neural networks, and the networks adapt the lights’ behavior based off of the learned traffic patterns. The project has successfully detected a car and turned a house lamp on and off with an ultrasonic distance sensor.