IWO-based adaptive neuro-fuzzy controller for mobile robot navigation in cluttered environments

In this paper, a new hybrid intelligent motion planning approach to mobile robot navigation is presented. In this new hybrid methodology, the invasive weed optimization (IWO) algorithm is used for training the premise parameters, and the least square estimation (LSE) method is used for training the consequent part of the adaptive neuro-fuzzy inference system (ANFIS). In this proposed navigational model, different kinds of sensor-extracted information, such as front obstacle distance (FOD), right obstacle distance (ROD), left obstacle distance (LOD), heading angle (HA), left wheel velocity (LWV), and right wheel velocity (RWV), are given input to the hybrid controller, in order to calculate the suitable steering angle (SA) for the robot. Using the IWO algorithm, the obtained root mean of squared error (RMSE) for the training data set in the ANFIS is 0.0013. The simulation results are verified by the real-time experimental results, using the Khepera III mobile robot to show the versatility and effectiveness of the proposed hybrid navigational algorithm. The results obtained using the proposed hybrid algorithm are validated by comparison with the results from other intelligent algorithms. Finally, it is proved that the proposed hybrid navigational controller can be implemented in the robot for navigation in any complex environments.