Control and navigation system for autonomous vehicles and robots using fuzzy logic and Kalman filtering

This thesis presents autonomous vehicle navigation in 3-D or 2-D environment. The vehicle has three different kinds of sensors to navigate in the obstacles populated environment. The obstacles may be static or dynamic. The vehicle's main sensor systems are sonar, global positioning system (GPS) and inertial navigation system (INS). The first sensor is used for obstacle avoidance and object recognition. The second and third sensor is used to determine the position and velocity. The signals from the Global Positioning System (GPS) and Inertial Navigation System (INS) are fused together using Adaptive Fuzzy Logic Kalman filter and the fused signal is fed to the vehicle control system. The control system is based on fuzzy logic controller (FLC). The FLC consists of two geometric modes and three dynamic loops. First group, geometric modes, the controller is making the decision how to follow from the starting point to its final goal and trace the edge of obstacles. Second group, dynamic loops, the controller is changing its velocity, attitude and acceleration dynamically. The results of simulations show that the fully autonomous vehicle can navigate in sparsely as well as densely populated environment. It has been demonstrated that the Fuzzy Adaptive Kalman Filter gives more accurate results than the Extended Kalman Filter does when INS fails or is modeled improperly.