Multisensor controlled obstacle avoidance and navigation of intelligent security robot

Abstract

In the real world, development of robotics has replaced our manpower gradually, and assisted human to perform rather difficult tasks. One of the most important factors for the ISR(Intelligent Security Robot) to reach the goal safely is how to avoid the unknown obstacles in the unknown environment.The unknown obstacles include static and dynamic obstacles which are detected by ultrasonic sensors and IR sensors respectively. If we use only one of the two kinds of sensors to detect obstacles, it is not useful for ISR to complete tasks in the real world, because the surrounding environment in the real world is dynamic, unknown and time-varying. Therefore, we use sixteen pieces of ultrasonic sensors and eight IR sensors to detect obstacle, and measure distance from the center of ISR to obstacles. Hence, the multi-sensor integration technique is applied in the robotics. This technique acquires sensory information to construct environment information. On the other hand, shapes and areas of obstacles will both change the data that IR sensor and Ultrasonic sensor registered. As a result, they will change the decision that ISR made during the process of obstacle avoidance. So the time ISR spent and the length of path will have some differences when ISR navigate to the given position from the start position. The system has been successfully demonstrated experimentally.

Vision based obstacle detection using 3D HSV histograms

“…However, none of these sensors are perfect, and they all have certain limitations. Infrared sensors, for example, are highly material dependent, and usually have a low range of detection [1]. Ultrasonic sensors can detect obstacles as far as 6-10 meters, but suffer from specular reflections and poor angle of resolution [2].…”