Arthur Gomes
Pranjal Mishra
Sahas Agarwal
Submission for Vichesta-Takshak 2021, a pan-india robotics competition organised by IIT-Dhanbad. This solution was awarded the first position.
View the final presentation here
cd ~/catkin_ws/src
git clone https://github.com/Pranjalmishra30/Vichesta-Takshak-2021.git
cd ..
catkin_make
source devel/setup.bash
sudo apt install ros-[version]-[package_name]
sudo apt install ros-melodic-move-base
sudo apt install ros-melodic-gmapping
sudo apt install ros-melodic-dwa-local-planner
sudo apt install ros-melodic-explore-lite
sudo apt install ros-melodic-ros-numpy
sudo apt install ros-melodic-map-server
roslaunch takshak launch_all.launch
A robot has to navigate an indoor environment of corridors and rooms while avoiding obstacles and performing ancillary tasks with no human intervention. The robot to be used is the Husky-a 4-wheel differential drive robot base with an Intel realsenseD435i depth camera, an RP lidar and an inertial measurement unit.
- The aruco markers have to be located and read. Each marker represents a number from 0 to 4. The number represented by the marker has to be matched with the colour of the object above it.
- The spheres in the arena should be counted. Any number of spheres could be present in any location.
- The number of spheres is divided by 5. The remainder of the operation corresponds to the marker with the same integer and by extension the associated colour of the marker. The gate with said colour has to be navigated through.
- Currently the hough circles algorithm is unreliable. 3D Point cloud data could be used for more accurate detection.
- A 3D map of the environment would result in better obstacle avoidance.
- Map precision could be increased at the price of added computation.
- Competition Github
- External packages used