Özet:

Abstract In the recent years, the probabilistic path planning is an emerging area in the field of navigation. The navigation applications increases day by day which helps the society to solve the real world problems. The major challenge in this path planning is to deal with the dynamics in the environment. The dynamicity refers with the ability of an obstacle to move around the working environment and able to change their possessions frequently. There are many solutions were available to deal with the challenges in the dynamic environment and support the robot to navigate over the environment to move from source to a destination. Sampling based algorithms are the one which are most suitable for the path planning where the dynamic obstacles are present in the working environment. Rapidly Exploring Random Tree (RRT) and their variants like RRT*, F-RRT*, PQ-RRT*, etc. are the algorithms, have been shown significant improvement over predicted path and the time to navigate over predicted path. These variants of RRT* algorithms are analyzing over the asymptotic behaviour and the cost of the generated path is also analyzed. In this work we have proposed a dynamic RRT* algorithm which is critically analyzed in terms of optimality and the asymptotic behaviour is cortically analyzed to present a dynamic RRT* planner, which can be effectively used for the path planning over dynamic environment. However these algorithms were not analyzed in terms of dynamicity of the environment and make the approach dynamic which can adapt the environment features to be work dynamically to generate a path. The presented simulation results shows that the presented dynamic RRT* algorithm work significantly better in terms of path length and the navigation time during the actual run from source to a destination. The study of path planners developed over the years by the research community has been discussed and presented here in this work. Moreover the proposed dynamic RRT* approach shows that the computational cost of the algorithm makes it to a probabilistically complete solution to work with the dynamic environment.

Anahtar Kelimeler: