emreay- / snoopy Goto Github PK

##Snoopy

• snoopy_launch: General launch files
• snoopy_setup: Includes nodes for setting up the snoopy including YAML parameter files, brain, tf broadcasting and listener nodes, exploration node etc...
• snoopy_control: PID for motor control.
• snoopy_odom: For dead reckoning
• snoopy_teleop: Teleoperating using keyboard
• snoopy_vision: Object detection, shape classification and color detection from camera.
• snoopy_pathfind: Path finding using A*
• snoopy_localize: Particle filter localization
• snoopy_navigate: Action lib for navigation
• snoopy_map: Map server for occupancy grid map
• snoopy_uarm: Using uarm

ros_control pkg:

  sudo apt-get install ros-indigo-ros-control ros-indigo-ros-controllers

#Demos

The PF localization system was tested in the maze environment for both global localization and pose tracking. The dead reck- oning system was also kept tracked to be compared. In the experiment, 2000 particles and every 8th range measurement were used. With this setup, the mean estimation rate was a lit- tle more than 3 Hz.

• Detailed report on localization can be found here.

To navigate the robot to a desired location A* is implemented. In the demo the code is running and giving a path from an origin to the clicked points on the map.

The explore service in snoopy_setup node was tracking the tiles in the map that were explored, and when called it was sampling a target point in the map from a uniform distribution that was lying between a minimum and maximum distance from the current position of the robot which was neither occupied nor explored. Exploration state was switching in brain node if there is an object detected from the vision or a collision is detected.