Final Presentation

1
Final Presentation

• EE 296 Micromouse
• Spring 2007
• Friday, May 4, 2007
• POST 214

2
Team Giga-Cat

• Team Members
• Mark Fujihara
• Christopher Ho
• Katherine Ishida
• David Ota

3
Overview

• Build an autonomous robot mouse that can find the
  center of a 16 x 16 maze and then calculate and
  make multiple runs using the shortest path to the
  center.

4
Our First Design

• Top Down sensors for path/wall detection
• 4 sensors at each corner
• 3 for error correction
• 1 for wall detection
• 2 sensors in the middle for error correction
  during turns
• Two wheels and a ball bearing for mobility
• Stepper motors

5
Drawings and Designs
Sensor Layout
Chassis Layout
6
More Designs

• Top and Side View Base Layout

• Top View with Sensor Blocks

7
Goals For Original Design

• Move forward and back
• Front and back sensors
• Look sleek, not bulky
• Stacked circuit boards
• Put a spoiler and bumper on it
• Use a flood fill algorithm to map and solve the
  maze
• Go to the mainland and win competition
• Build a killer mouse

8
What We Ended Up With

• Top Down sensors for path/wall detection
• 3 sensors at each corner
• 2 for error correction, 1 for wall detection
• 2 sensors in the front and back for front and
  back wall detection
• 1 sensor in front and back for emergency stop
• 2 wheels and 2 slider pads for mobility
• Stepper motors
• Right wall hugger

9
Our Final Design
Here are the early stages of construction
10
Our Final Design
Which eventually became our final beauty!
11
Final Design Features

• Used a stacked circuit board design
• L-braces to hold motors in place
• Had sensors on a metal frame, adjustable
  according to the wall height

12
More Design Features

• Put LEDs on top sensor board instead of directly
  over corresponding sensor
• Incorporated LEDs for each motor step

13
So What Went Wrong?

• Sensor layout was bad
• We re-did the sensor layout
• Number of sensors
• Placement
• Our original chassis was too big
• We built a new chassis
• Our stacked design was terrible
• Hard to solder
• Spacing was bad
• We sucked it up and dealt with it

14
More Problems

• Our mouse did not move the first time
• We found our power and ground rails were
  connected together and fixed it
• One motor did not step correctly
• We found a loose connection in the MOSFET
• Mouse moved, but kept getting stuck
• We found the mouse was too heavy, so we slowed
  the speed down to create more torque
• Battery connections were loose
• We had to put more electrical tape

15
Programming Problems

• Originally pulsed each motor individually (right,
  left, right, left, etc.)
• We changed the program so the motors stepped
  together
• Both wheels were programmed to go clockwise, so
  it went in a circle
• Made one go counter clockwise

16
Whats Still Wrong?

• Mouse can only navigate the maze
• Does not map
• Does not solve the shortest path
• Does not go backward
• Does not win any competition (except the beauty
  contest!)

17
Design Suggestions

• Do not use the stacked design
• Complicated and messy, too many extra wires
• Do not use too many standoffs or extra metal that
  adds weight
• Use a Perforated Circuit Board with copper lining
  in the holes
• It will make soldering neater and easier
• Color-coding wires is very useful

18
Completion Suggestions

• Do not waste time at the beginning of the
  semester. It is not as long as you think.
• Multi-task. Have someone begin code before mouse
  is completely built
• Work hard during spring break

19
Now, lets see our beauty in action!Mini-maze
please
20
Thank You

• Any Questions, Comments, Suggestions?