Project “Easter Egg Hunt”

Final project with Edward Macdonalds, Daniel Pickem and Kunal Muchhala in Ayanna Howard’s class ‘Autonomous Control of Robotic Systems’ at Georgia Tech. In a simulated environment (Player & Gazebo), we enabled a pioneer robot to autonomously explore its environment and collect and retrieve colorful easter eggs to a base location. We deployed a SLAM system, A* for path planning, a motion controller and computer vision algorithms.

The simulated pioneer robots mission is to collect and retrieve easter eggs that are randomly distributed in a unknown maze-like environment. To accomplish this task the robot utilizes a laser scanner to perform Simultaneous Localization and Map building (SLAM). A path planner (A*) works on the grid based map to find paths to unexplored areas, where eggs could be hidden. A motion controller for this kind of mechanical setup controls the wheels such that the robot follows the desired trajectory. During the whole time a color-blob extractor works on the frames grabbed by a camera attached to the robot gripper. Once a green and a pink blob are detected in close proximity, a sequence of actions is executed to bring the robot in pick-up position, grab the Egg at its ribbon and lift it up into a save position over the robot. Then the path planner finds the shortest trajectory back to base of the robot, where the Egg is dropped, before the hunt for more Eggs goes on.

Results

The original run of the robot was 2h and 12 minutes long, so we speeded it up by a factor of 25. In this run the robot was able to map almost the whole environment while retrieving around 16 Easter Eggs!