Our Team
|
Maxwel Cichon |
Kaila Day |
Daniel Hill |
Joshua Love |
Spencer Norwick |
|||||
|
|
|
|
|
|
|||||
Project Overview
For this project, we were tasked with creating a
robot that navigates to 5 positions as shown on the map below, and also locates
and “exterminates” pink and blue weeds placed randomly on the course. While
doing this, the robot must avoid all obstacles on the course to maintain the
best score possible. Once all of the points are reached, the robot returns the
number of weeds found to their respective home circle.
Strategy & Methods
Path Planning
Our robot uses the A* method to help the robot navigate
through the course as efficiently and as fast as possible as it calculates the
optimal route with respect to the obstacles. Our version of the A* method was
slightly altered, as to interrupt for detected weeds and avoid obstacles.
Through integration of a LabView Program, we were
able to watch the robot in real time move through the course.
Obstacle Avoidance
The robot uses LADAR data collection in order to
avoid the obstacles placed randomly on the course. By using wall following, we
were able to ensure that given the event that our robot did encounter a wall or
dead end, it would be able to navigate away from it in order to continue safely
on its journey to navigate to the waypoints.
Weed Detection
Additionally, the robot sends the locations of
the weeds and waypoints as well as its current location to a LABView program depicting the course. The instantaneous
location of the robot is detected using the OptiTrack
Motion Capture system and sent to the robot for localization purposes. By
tuning our color detection algorithm in MATLab to
only stray from the most optimal route in the event of
a weed, we ensured that we would find all of the weeds during our
demo. In the LabView program we created, we marked
each weed found with its respective color, and each waypoint with a picture of
Professor Block’s face for clarity purposes.
Competition & Demo
Our results from the competition proved that
although our robot was extremely thorough, it was not very fast and thus created
inefficiencies. With a final time of 193 seconds, the robot successfully
located all of the weeds and did not hit any walls leaving us with a score of
-67. This put our team in 8th out of 9 teams, so if we were to do this project
again, we would definitely want to speed things up. A video of the demo is
shown below.
Source Code & Files
Source Code & Files
All source code can be
found using the link below:
https://github-dev.cs.illinois.edu/kaday2/FinalProject