The autonomous navigation of a mobile robot in an unknown environment depends on the robot’s ability to perceive and adapt to that environment. The dynamic window approach (DWA) is a local planning algorithm for obstacle avoidance that considers the motion dynamics of the robot while steering to avoid collisions. The DWA generates a window with all the velocities that are reachable in a time interval and then evaluates all the velocities with an objective function to select the best velocity. In this paper, a novel hybrid algorithm is proposed based on DWA, fuzzy inference system (FIS) and Kalman filter. The proposed work discusses the implementation of DWA algorithm along with FIS to tune the weights of the DWA objective function. Also, the algorithm is extended to incorporate a Kalman filter for state estimation of dynamic obstacles in the environment.