Application of Neuro-Fuzzy Controller for Sumo Robot control

This paper proposes the application of Neuro-Fuzzy (NF) hybrid system for Sumo Robot (SR) control. This robot is frequently designed by engineering students for robotic competition. As the relation between sensors output signals and motors control pulses is highly nonlinear in SR, soft computing techniques can be used to define this nonlinear relation and control of the robot in a competition ring. Application of intelligent methods for SR control not only simplifies robot control and improves robot responses during competition, but also encourages engineering students to use intelligent methods for solving real world’s problems. Regarding above rationale, a NF controller for SR control is proposed and implemented. Firstly, a Fuzzy Inference System (FIS) for detecting and tracking of the opponent in the competition ring is developed, which relates sensor output signals to motor control pulses. Secondly, Artificial Neural Networks (ANN) based learning algorithm is used for rule extraction and tuning the FIS parameters. The design approach of the proposed controller is presented in detail, and effectiveness of the controller is demonstrated by hardware implementation and experimental results. The results show that the intelligent control methods can be easily applied in various robot competitions by engineering students.

Research highlights
► This paper proposes the application of a Neuro-Fuzzy system for Sumo Robot (SR) control. A Fuzzy Inference System (FIS) for detecting and tracking of the opponent is developed. Artificial Neural Networks is used for tuning the FIS parameters. Intelligent methods can be applied in robot competitions by engineering students. This study encourages students to use intelligent methods in real world.