MODIFIED ALGORITHM FOR PATH POINTS SMOOTHING

Abstract

The article deals with the issue of linearization of a broken curve consisting of path points. Restrictions of classical Ramer-Douglas-Peucker algorithm are shown in the paper. These restrictions are important for a robot for resources consumption and keeping the speed in the field. They include the use of such base curve points that have maximum values of coordinate deviation between the first and the last points. In this case robot’s motion path is smoothed, but the rotary angles are still wide. The robot has to stop at base points and turn according to the path. This action increases the path time. Due to this it is impossible to use the algorithm directly to smooth the mobile robot motion. The phase of pre-processing has been introduced into Ramer-Douglas-Peucker algorithm, which includes collection of data about numerical data of via points of the initial curve. Consideration of these data allows objective identification of points of the curve which are understood as noise. In contrast to the initial algorithm, base points with large coordinate values are not considered. To do this, deviations of all points between the first and the last ones are analyzed. Statistic data for the curve are acquired. Based on user-set limit deviations, points which will be understood as noise are defined. As the result, the obtained broken curve has a smaller span between points. The number of points in the curve decreases. These changes provide for better performance of a mobile robot.