Research on estimating smoothed value and differential value by using sliding mode system

Abstract

To be able to recognize an environment, a robot should have as many sensors as possible. When we use sensors, we must consider the characteristics of the sensors, such as range, processing time, error, and so on. In this paper, we focus on the ultrasonic wave sensor that is today the most common sensor employed on indoor mobile robotic systems, and we propose a new technique for estimating the smoothed value and the differential value of the distances measured by the ultrasonic wave sensor. In proposing this system, we take the characteristics of the sensors mentioned above into consideration. In spite of the many methods proposed, it is still very difficult to eliminate the noise of sonar completely. Therefore, we smooth the distance value by assuming the continuity of the signal obtained by the sonar, and taking advantage of this continuity, we compose a robust estimator. The estimator is based on the sliding mode system.