THE BEAM-CURVATURE METHOD: A NEW APPROACH FOR IMPROVING LOCAL REALTIME OBSTACLE AVOIDANCE

摘要

This paper describes a new local avoidance method for indoor mobile robots. The method uses a directional method named the Beam Method to improve the performance of a local obstacle avoidance approach called Curvature Velocity Method (CVM). The proposed Beam Method employs radial distances provided by the robot sensors to calculate the best one-step headings. The CVM uses this information to calculate the optimal translational and rotational velocities. Our approach incorporates the local information in an intuitive way, similar to human vision, by taking advantage of the natural disposition of the mobile robot sensor systems. Compared with other methods it shows more efficient performances, being able to navigate through openings very efficiently way. Several experiments in populated and dynamic environments with our mobile robot RATO have proved to be very successful.