Abstract
In this paper, a robot navigation model is constructed in MATLAB-Simulink. This robot navigation model make the robot capable for the obstacles avoidance in unknown environment. The navigation model uses two types of controllers: pure pursuit controller and fuzzy logic controller. The role of the pure pursuit controller is to generate linear and angular velocities to drive the robot from its current position to the given goal position. The obstacle avoidance is achieved through the fuzzy logic controller. For the fuzzy controller, two novel fuzzy inference systems (FISs) are developed. Initially, a Mamdani-type fuzzy inference system (FIS) is generated. Using this Mamdani-type FIS in the fuzzy controller, the training data of input and output mapping, is collected. This training data is supplied to the adaptive neuro-fuzzy inference system (ANFIS) to obtain the second FIS as of Sugeno-type. The navigation model, using the proposed FISs, is implemented on the simulated as well as real robots.
Highlights
Mobile robots have a wide range of applications like space missions, household and office work, receiving-delivering orders of clients in restaurants, transportation of logistics in inventories, inspection and maintenance, agriculture, security and defence, operations in radioactive areas etc
The training data obtained from the Mamdani-type fuzzy inference systems (FISs) is quit suitable for adaptive neuro-fuzzy inference system (ANFIS) model
It is observed that the ranges and types of input membership functions of Sugeno-type FIS can successfully obtained using ANFIS model
Summary
Mobile robots have a wide range of applications like space missions, household and office work, receiving-delivering orders of clients in restaurants, transportation of logistics in inventories, inspection and maintenance, agriculture, security and defence, operations in radioactive areas etc. The global navigation method is used to find globally optimal path on the basis of priory information like map of the environment. Instead of the priory information, the local navigation is based on the on-line sensory data received from the sensors mounted on the robot. Neural network can be employed to map the relationships between inputs and outputs for interpreting the sensory data, obstacle avoidance and path planning. To improve automatic learning and adaptation, ANFIS combines fuzzy logic and neural network. A fuzzy controller for robot navigation in unknown static and dynamic environment is implemented in [10]. The remaining part of this paper is organized as follows: Section 2 presents the theoretical background of controllers used in our robot navigation model.
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
