Teach a robot to guide a person to a predefined destination.
-Use a Turtlebot2 as the robot.
-The robot must use a vision algorithm(s) as the main guidance.
-Sensors such as bump,infa-red and ultra-sound should provide direction and safety for the robot.
-The solution should be easily modified for alternative routes.
Since the TurtleBot relies so heavily on its Kinect camera, I needed to see what the TurtleBot actually sees. Here is a video of me demonstrating the TurtleBot's ability to see depth.
Here is an image of the TurtleBot looking at a portion of a room in my house.
Here is what the TurtleBot sees with its Kinect. From this, you can see that it does not only see a 2D picture, but does see in 3D.
A key element of autonomous navigation is the creation of a map of the area, I've made a short presentation of what the map making process looks like. A large problem that I was facing with this project was that the TurtleBot did not seem to be calibrated properly and I could not figure out how to calibrate it.
Here is an another image of the TurtleBot looking at a portion of a room in my house.
Here is what the TurtleBot sees when pointed in this direction. As you can see, the map the TurtleBot has drawn is nowhere near accurate. The dark blue/green areas are what the TurtleBot detects as walls and the light white cone is what the TurtleBot is seeing.
After the TurtleBot makes a map, it should be able to autonomously navigate it. However, I was not able to get it to do so but I made a presentation to show what I could get.
Thanks to Kurt Ahonen, I was able to have my presentation recorded. Please watch it if you think it will help you out.
Design from this website taken from CSCI322 Client/Server Project.
