Behavior Integration via MS-ANFIS for Realistic Navigation of Multisensor Autonomous Vehicle

In the recent works either the obstacle avoidance of robot in the 2-D unconstrained environment or the wall following behavior of mobile robot in constrained motion has been discussed. In some of them target alignment behavior has been discussed. Also the integration of some but not sufficient behaviors have been discussed in some research works. But all of the behaviors have not yet been integrated. So the problem of designing the optimal controller to perform on all conditions has been a major attraction for the designers and scientists. In this paper the effort has been made to take into account various types of factors like the limitations of suspension system and the physics laws like friction, motion, momentum, centrifugal and centripetal force, Tilting avoidance due to turning Traffic rules, Traffic congestion, signals, overtaking, passing high speed vehicle, speed breakers, single lane road and varying width road, final target alignment in no-road or broad area, limited time to reach the target etc. during the design of the controller. Designing the optimal controller with all the behaviors integrated for proper motion planning in all types of environments needs much treatment. This work mainly deals with designing of the unified controller for optimal navigation of a multisensor mobile robot in real-environment by applying fuzzy rule simplification technique (FRST) via multi-stage adaptive neuro-fuzzy inference system (MSANFIS).The Proportional and Derivative (PD) control has also been used in connection with some ANFIS stages for this work.