The advancement of robotics technology has enabled automation in various
tasks, including cleaning in public facilities such as libraries. However, the
complex library environment with numerous fixed obstacles and dynamic
user activities demands an adaptive navigation system. This study
proposes an adaptive navigation system for library cleaning robots based
on the integration of the Best First Search (BFS) algorithm, fuzzy
inference, and Multilayer Perceptron (MLP). BFS is used to determine the
optimal path based on priorities, while MLP predicts floor dirt levels based
on coordinates and floor types. The MLP predictions are then converted into linguistic priority categories (high, medium, low) using fuzzy logic.
Simulations were conducted on a 5×5 grid map representing a library area
with varying start and goal points. The results demonstrate that the system
can avoid obstacles, select paths based on dirt levels, and allocate cleaning
time efficiently. The integration of these three methods significantly
enhances the adaptability and efficiency of robot navigation in dynamic
environments. This research contributes significantly to the development
of intelligent robotic systems for automated cleaning in public spaces.
You may also start an advanced similarity search for this article.