Thermal Image Sensing Model for Robotic Planning and Search

Area of Science:

• Robotics
• Sensor Technology
• Artificial Intelligence

Background:

• Autonomous robots require effective navigation systems for locating targets in unknown environments.
• Infrared (IR) radiation detection offers a non-invasive method for identifying heat sources.
• Accurate sensor modeling is crucial for precise distance and temperature estimations.

Purpose of the Study:

• To develop and validate a search planning system for a rolling robot to autonomously detect and navigate towards an infrared (IR) heat source.
• To characterize and model the performance of a low-cost, home-made IR passive visual sensor.
• To integrate sensor feedback with autonomous navigation for efficient and safe robot operation.

Main Methods:

• Experimental characterization of IR sensor spectra.
• Development of exponential and polynomial models for distance and temperature estimation from IR intensity.
• Implementation of a nonlinear equation system solved via a Newton-based approach for IR source localization.
• Utilizing trigonometric partial differential equations for robot course control, incorporating attractive and repulsive accelerations.

Main Results:

• Successful experimental characterization of the IR sensor's capabilities.
• Accurate estimation of distance and temperature using combined theoretical models.
• Effective localization of the IR source in global coordinates through the planning system.
• Demonstrated autonomous navigation and obstacle avoidance in complex indoor environments via simulations and real-world experiments.

Conclusions:

• The proposed search planning system enables autonomous navigation towards IR heat sources.
• The integrated sensor modeling and navigation control system proves effective and convenient for robotic applications.
• The approach demonstrates efficacy in complex indoor environments, highlighting its practical utility.