chrisbove / rbe-3002-rosbert

Team ROSbert repo for lab code

RBE-3002-ROSbert

Team ROSbert repo for RBE 3002 labs at Worcester Polytechnic Institute. The final project involved having a Turtlebot fully explore and map a maze.

This code is completely unsupported in any way. :)

Setup

cd ~/catkin_ws/src/
git clone https://github.com/ChrisBove/RBE-3002-ROSbert.git
cd ..
catkin_make
source devel/setup.bash

Full setup

mkdir -p ~/catkin_ws/src
cd ~/catkin_ws/src
catkin_init_workspace
cd ~/catkin_ws/
catkin_make
source devel/setup.bash
#bashrc stuff
echo source ~/catkin_ws/devel/setup.bash >> ~/.bashrc
echo alias rbhome="'cd ~/catkin_ws/src/RBE-3002-ROSbert/'" >> ~/.bashrc
echo alias rbmake="'catkin_make install -DCMAKE_INSTALL_PREFIX:PATH=~/catkin_ws/install -C ~/catkin_ws -DCMAKE_BUILD_TYPE=Release'" >> ~/.bashrc
echo alias rbeclipse="'catkin_make --force-cmake -G\"Eclipse CDT4 - Unix Makefiles\" -DCMAKE_BUILD_TYPE=Debug -DCMAKE_ECLIPSE_MAKE_ARGUMENTS=-j8 -C ~/catkin_ws'" >> ~/.bashrc
#workspace setup
cd ~/catkin_ws/src
wstool init
wstool set RBE-3002-ROSbert https://github.com/ChrisBove/RBE-3002-ROSbert --git
. ~/.bashrc
wstool update
. ~/.bashrc
cd ~/catkin_ws/
catkin_make
. ~/.bashrc

##Aliases

echo source ~/catkin_ws/devel/setup.bash >> ~/.bashrc
echo alias rbhome="'cd ~/catkin_ws/src/RBE-3002-ROSbert/'" >> ~/.bashrc
echo alias rbmake="'catkin_make install -DCMAKE_INSTALL_PREFIX:PATH=~/catkin_ws/install -C ~/catkin_ws -DCMAKE_BUILD_TYPE=Release'" >> ~/.bashrc
echo alias rbeclipse="'catkin_make --force-cmake -G\"Eclipse CDT4 - Unix Makefiles\" -DCMAKE_BUILD_TYPE=Debug -DCMAKE_ECLIPSE_MAKE_ARGUMENTS=-j8 -C ~/catkin_ws'" >> ~/.bashrc

rbe-3002-rosbert's Issues

Need way to interrupt A* and give it a new goal

Waypoints aren't ordered properly

Reversing them doesn't seem to fix it

Random path point on the corner of map

There is an erroneous path point on the corner of the map. It seems to be the last thing in the list.

doing a new waypoint:
x: 9.97500014864
y: -0.0250000003725
z: 0

Nav Attempt/Failed topic doesn't go through when we try publishing True

So our code just hangs, waiting for a status update.

A star seems to like walls

It uses the expanded obstacles sometimes as part of the path.

That, or we have a big offset problem.

final: find edges always searches the initial map

It does not seem to update and search the updated map. It does the same edges everytime, so the robot makes one move, then remains at the same spot since it was already there.

A* Fails to navigate around corner traps quickly

It can take a very, very long time for it to figure out how to get around a wall.

It seems like the heuristic isn't weighted properly.

Something is wrong with map visualization code

The given code has some errors. Says Eric:

As a side note about the lab, the code given for lab 3 that goes through the map data and colors any cell that is occupied I would strongly recommend you to rewrite it as if you ran it you should of found there were issues with it. The map data is a 1D array that you need to search through given the width and height and you can use two for loops to easily turn it into a 2D array. For placing the points, you want to add + 0.5 * resolution to both the x and y point values instead of the offset it currently has in the code so that you put a point into the center of the cell you desire so it will light it up in RViz.