Abstract:
Mobile robots are becoming increasingly popular in modern-day robotics. Mobile
robots are being used in many fields like exploration, defence, agriculture, cleaning, lawn
mowing, and warehouse management to name a few. The operation of a mobile robot requires
accurate data for mapping, path planning, and navigation. Each subsequent operation is
dependent on the previous operation so the margin for error is very little. In this work, we
develop a complete framework for mapping, complete coverage path planning (CCPP), and
navigation along with local planning for handling the complexities of dynamic environments.
Some improvements in already existing literature have been proposed like the most popular
Boustrophedon Cellular Decomposition (BCD) to decrease cost and increase efficiency in
terms of time and energy of the system. This work also proposes a memory-efficient approach
for solving local planning problems through the velocity obstacle (VO) method.
Experimental work was carried out in simulated environment of Robot Operating
System (ROS). Global path planning results showed significant improvement in terms of
minimizing overlapping issues when compared with the original BCD algorithm. Navigation
results showed that all waypoints generated by global planner were visited with 94% accuracy.
The velocity obstacle approach was implemented as local path planner to handle moving
obstacles. To further improve the performance, a two-level approach was used to handle
moving obstacles
