Naadir M. Kirlew

Knight Foundation School of Computing and Information Sciences

Lecture Information:
  • March 26, 2021
  • 5:00 PM
  • Zoom

Speaker Bio

Naadir Kirlew is a Master’s student at Florida International University (FIU), Knight Foundation School of Computing and Information Sciences (KFSCIS). He obtained his Bachelor’s degree in Mechanical Engineering at FIU with a focus in robotics. His subject of interest is Artificial Intelligence, Machine Learning, and Robotics.


Multi-Robot collaboration has become a topic of great interest as problems have become more complex and technology has improved to allow for a single controller of multiple agents. Many are exploring the field of swarm robotics alongside homogeneous and heterogeneous robot collaboration. This report explores the concept of utilizing multiple robots to explore and map an unknown environment as thoroughly as possible within a limited time constraint.

This report employs a distinct robot model known as a differential drive rover, as its simplicity allows this experiment to be reproducible without require complex components. The rover is tasked with an exploration and mapping mission in which it must successfully navigate the environment while avoiding obstacles and other robots. Additionally, in the multi-robot configuration each robot performs their own intelligent decision making and are not taking commands from each other.

A visibility algorithm was developed which allowed the robots to explore the environment prioritizing areas that would reveal more of the environment. This was done by implementing a visibility heuristic based on the robot’s lidar data as well as onboard odometry. This vision planner was then tested in a single robot test case and a multi-robot test case. In the single robot test case the visibility algorithm performed better than the standard wall following algorithm that it was tested against, achieving a total map coverage of 82.04%. In the multi-robot test case, the visibility algorithm was able to achieve a total map coverage of 84.94%.

Through experimentation it was determined that the visibility planner proposed in this experiment is a viable solution to the exploration of unknown environment problem.