A novel approach to integrate artificial potential field and fuzzy logic into a common framework for robots autonomous navigation

Abstract

Artificial potential field and fuzzy logic are efficient approaches for mobile robots autonomous navigation. However, both have advantages and drawbacks. Their integration into a common control scheme can significantly improve the performances of the resulting hybrid controller. In this article, we propose a novel hybrid approach in order to better exploit their advantages. The present work contributes in three aspects: first, the proposed control scheme integrates interval type-2 fuzzy logic concepts with artificial potential field concepts into a common framework in order to better exploit their advantages. Second, the proposed control scheme is a simple and realizable design for real-time implementation because only 15 fuzzy rules are sufficient to control the mobile robot. Third, the proposed control scheme is a synthesized design which utilizes both heuristic knowledge and the sampled input–output data pairs. An implementation in real-time on an omnidirectional mobile robot validates the effectiveness of the proposed control scheme.

Target tracking for a correlated multi-sensor system with constant time delay

“…The majority of multi-sensor target tracking algorithms assume that multi-sensor noises and estimations are uncorrelated. [5][6][7] However, in practice, depending on the common estimator of each sensor, the common interference environment of measurements and the transformation of spatial coordinates, the measurements and estimations are correlated. [8][9][10] Therefore, there is a practical reason to research the multi-sensor target tracking of estimated correlation.…”

Global navigation approach for assistant robot

“…Other proposals can be discussed, such as neural networks approach which can find paths developing networks (Andrakhanov, 2013) and contribute to solve paths on uncertainty environments, but at a high computation cost. Classical methods such as potential field are another choice at plane surface (Arslan & Koditschek, 2016;Bermudez, Castellar, Montiel, & Ceballos, 2004) and 3D Cartesian space (Chen, Luo, Mei, Yu & Su, 2016;Mac, Copot, Hernandez & De Keyser, 2016;Melingui, Merzouki, Mbede & Chettibi, 2014;Orozco-Rosas, Montiel & Sepulveda, 2015;Vechet, Chen & Krejsa, 2014), with low cost computation time and feasible to real-time applications. The idea of this research is to try innovative path-planning techniques for real-time applications.…”

Collision avoidance for mobile robots based on artificial potential field and obstacle envelope modelling

“…According to [8], the path planning problem in the high level is a NP-hard optimization one which is better to be solved by probabilistic heuristic algorithms such as GA rather than deterministic equational algorithms such as APF. Assuming that a preliminary path has been developed by some kind of high level algorithm, such as probabilistic roadmaps algorithm [6] [9], genetic algorithm [8] [10] [11] [12] [13], fuzzy logic control [14] [15] and so on, this paper mainly focuses on solving the collision avoidance problem in the low level to eliminate local minima as well as trajectory oscillations.…”