Odor Source Localization in Obstacle Regions Using Switching Planning Algorithms with a Switching Framework.

Area of Science:

• Robotics
• Environmental Science
• Chemical Engineering

Background:

• Odor source localization (OSL) robots are critical for safety and rescue operations involving hazardous chemical plumes.
• Complex environments hinder accurate odor plume dispersion modeling, limiting traditional OSL algorithms.
• Time-sensitive OSL tasks require adaptive strategies balancing exploration and exploitation.

Purpose of the Study:

• To develop an effective OSL strategy for robots in complex environments without precise plume dispersion models.
• To enhance the speed and success rate of OSL by dynamically adjusting exploration and exploitation.
• To address the limitations of existing probabilistic search algorithms in practical scenarios.

Main Methods:

• Environment simplification by dividing complex regions into multiple sub-environments of varying resolutions.
• A framework integrating Infotaxis and Dijkstra algorithms for adaptive navigation.
• Switching between algorithms to guide plume clue searching and inter-sub-environment movement.

Main Results:

• The proposed framework significantly improved OSL success rates and reduced search times.
• Experimental validation on an autonomous mobile robot demonstrated a 3.5-fold increase in success rate.
• Average robot movement steps were reduced by approximately 35%.

Conclusions:

• The developed framework effectively overcomes challenges in OSL within complex environments.
• The adaptive algorithm switching enhances robot navigation efficiency and task performance.
• This approach offers a practical solution for real-world hazardous plume detection and localization.