An integrated computer vision and infrared sensor based approach to autonomous robot navigation in an indoor environment
Document Type
Conference Proceeding
Publication Date
1-1-2009
Abstract
We present a novel approach to robot navigation in an indoor environment based on analysis of single lens camera images and infrared sensor readings. Our software design is based on the state diagram methodology where the robot transitions between states when a change in the environment, known as trigger, is detected. These environmental triggers are based on computer vision techniques, namely, (1) the Hough Transform for lines along with K-Means clustering used to compute orientation points for the robot and (2) histogram difference measurement for classification of objects (wall or door) the robot sees through its camera. We further use the infrared sensor to trigger obstacle avoidance behavior. By putting all these elements together within a state diagram framework we have created a platform for indoor robot navigation that is robust and easy to adapt to different topological layouts of indoor environments.
Recommended Citation
Bhattacharya, Sambit; Czejdo, Bogdan; and Mobley, Sean, "An integrated computer vision and infrared sensor based approach to autonomous robot navigation in an indoor environment" (2009). College of Health, Science, and Technology. 161.
https://digitalcommons.uncfsu.edu/college_health_science_technology/161