Obstacle Avoidance Using Markov Chains and Fuzzy Control for an Underwater Robot

This paper proposes an obstacle avoidance strategy for an underwater robot that integrates Hidden Markov Chains (HMC) with a fuzzy controller. The HMC predicts a vector of future states, leveraging the maximum state probabilities to guide the behavior of the fuzzy controller. The proposed fuzzy controller utilizes five inputs: the robot’s state predictions (crisp sets), instantaneous linear velocity, underwater depth, and yaw and pitch velocities, modeled with Gaussian and sigmoid membership functions. It generates three outputs, also using Gaussian and sigmoid functions, corresponding to the robot’s actuators: the propulsion motor, the steering motor, and the ballasting motor, which operates through an integrated hydraulic piston system. The paper also develops physics-based models for the propulsion, steering, and ballasting systems, alongside sensor fusion models that provide real-time control feedback. Additionally, it presents the robot’s platform and system architecture, designed with multi-threaded, real-time control capabilities. Experimental and simulation results validate the effectiveness of the proposed strategy, demonstrating robust obstacle avoidance in underwater navigation.