Soda Can
Retrieval Robot
Team
Members:
Paul “Catfish” Arendt Leslie DiMare Danielle Lasater Bethany Jackson
Overview:
The purpose of
this bot is to retrieve a certain soda can among
other cans and bring it to an assigned person, or
location by use of voice recognition.
The bot starts in the middle of the course,
receives a command to find a specific type of soda, drives out the cans and
spins around the find the correct can.
Once the correct can is found, a gripper is used to grab the can and
bring it back to the start of the course.
Next a command is sent to the bot to bring the
can to one of four people (locations).
The bot drives to that location, drops the
can, and waits for the next command.
This process was done with 3 different types of soda cans and four
different dropoff destinations.
Voice
Recognition:
The first
attribute of this bot was to recognize a voice
command to know which soda to pick up and to know where to drop the can. This
was done by using the Dragon Naturally Speaking program. Features of this
program allow the user to dictate, edit and control virtually all Windows-based
applications. No additional coding was
required for this feature, but each word used was specifically trained in the
voice program. In VB, this dictation
controls all command buttons. Buttons
were inserted that were connected to each soda can (Coke, Sprite, and Orange
Crush) and each person (Bethany, Danielle, Leslie, and Catfish). The VB application was connected to the C
code via Wireless Send, which allowed the recognition of each button to connect
to the appropriate can color or location in C.
Color
Sensing:
A color camera was
used to identify the different colors of the three cans. Each can was tested to obtain an appropriate
color threshold which was then added to the C code and used to identify each
can. After sensing the color, the bot was coded to follow the correct can and grip it once
the can was reached.
Obstacle
Avoidance:
The bot was also coded to avoid obstacles that were placed in
its way while delivering the can to the correct person. However, the video on this website does not
show any obstacles placed in the course.
Since the bot was coded to go to a specific
person/location, it would correct its path if any obstacles get in the way.
A special shoutout to Dan the Man & Doctor Juan for a great semester!