Team Eggplant

Intel Cornell Cup 2016

Thanks

• Professor Braunstein
• Intel
• Cornell
• Judges

Agenda

• Introduction and Pitch
• Concepts/Motivations
• Metrics, Solution Selection
• Accomplishments/Progress
• Execution Overview
• Moving Forward

The Team

Our Advisor

The Pitch

• "Dynamic" displays systems are costly to purchase and install
• Having dynamic display systems everywhere allows for crowd control via distributed sensor systems.
• Better directions can save lives and get you places quicker
• Can also replace disposable posters, recurring announcement boards etc.

Existing Display Technologies Not Cutting It

• Costly to purchase (especially as size increases)
• Complex and costly mounting and cable routing
• Wasteful when throwing out
• Not feasible as a safety device

Importance

• Flow control can be everywhere
• Many places capable of helpful hints or direction don't have affordable ways to mount a dynamic display system
• Flow control increases efficiency and has the potential to save lives

Problem Statement

Soft Objectives

• Image and message projection for a lower cost than TVs/Monitors
• Capable of indicating readable directions for a person to follow
• Portable and easily mounted

Key Functionality

Evaluation Metrics

Brainstorming

Conceptual Test (1)

Conceptual Test (2)

Investigation of Mechanical Solutions

Dual Rotating Mirror

Dual Rotating Mirror (cont)

Dual Fast Servo

Dual Fast Servo (cont)

2-Axis Oscillation

2-Axis Oscillation (cont)

Decision Matrix

Investigation of Microcontroller Solutions

ATmega328, ESP8266/Wifi Shield

• Familiar technology
• Low Cost
• Slow wifi communication
• Extremely resolution limited

PSOC 5LP, Redpine Signals WiFi

• Low Power Device
• Low Cost
• Custom fabrication required
• Intensive firmware development
• Resolution Limited

Intel Atom, Intel Galileo

• VGA Resolution
• Realtime Image manipulation via FPGA
• Capable wifi via mPCIe
• Expensive (for development)

Chosen Solution

Subsystem/Interface Overview

• Mechanical System
• Laser Circuit
• Software/Web API

Mechanical System Control

• Motor control via 5V 16 bit PWM
• individual horizontal and vertical adjustment

Laser Circuitry

• Single wire data line
• Frequency adjustment should correspond with motor speed

Software/Web API

• USB/UART serial communication to microcontroller
• Simple protocol for sending image bits
• Scripts for converting ascii art to binary blob
• Computer vision "dot" position identification
• Web server with POST endpoints

Prototype Progress

Initial Prototype

Initial Protoype (cont)

Initial Prototype Projection Example

Refined Prototype

Refined Prototype (cont)

Refined Prototype (cont)

Refined Prototype (cont)

Example Projection (cont)

Live Projection

Example Projection (cont)

Computer Vision Frequency Calibration

Evaluation of Prototype via Performance Metrics

Metric: Cost

• <$50 spent on final prototype
• Bill of materials worth approximately $90

Metric: Portability

• Recalibration necessary
• Relatively Fragile
• Mountable

Metric: Low Energy

• ~0.3A at 24V = 7.2 Watts
• Not ideal for battery-powered applications

Metric: Projection Quality

• Low Resolution (roughly 64x64 should be possible)
• Alignment issues due to ATmega328 limitations
• Display brightness very low

Metric: Ease of Implementation

• Resources on-hand
• Few undocumented/uncommon components

Results Discussion

• Rastering
  • Works!
  • Reasonable power consumption
  • Low visibility
  • Low resolution (due to microcontroller)

Execution Summary

• CAD and several system designs
• Functional Raster-based Prototype
• Software for image rastering
• Beginnings of computer vision frequency calibration

Issues / Risks

Iteration and Rapid Development

Budget Justification

Moving Forward

Department Support

• Increased budget to explore more prototypes

Integration w/ DE2i-150 (Intel Atom Dev Board)

• Increase resolution of rastered image
• Computer vision frequency alignment on-board
• Video support!

Development of online platform

• Interfaces for video support
• Web-based interface/GUI

Development of vector image prototype

Exploration of Photopolymers

Development of Safety Features

• Distance detection via integrated camera
• Incorrect distance or "unflat" shape turns off laser

Development of miniature raster image prototype(s)

In Review

• Laser projection systems allow for efficient building evacuation and flow control
• Built laser projection system capable of low-resolution images
• Designed several other laser projection systems
• Future work includes building unique new systems, safety features, photopolymers and Intel Atom integration.

Thanks!

• Professor Braunstein
• Intel
• Cornell
• Judges

Questions?