Redundant manipulators provide enhanced flexibility through additional degrees of freedom, enabling operation in environments where conventional manipulators would collide with obstacles. By exploiting null-space projection, these systems can avoid obstacles and joint limits while maintaining end-effector tracking. This thesis investigates and compares joint space control and operational space control methods incorporating null-space projection. Operational space control is expected to offer reduced computational cost by eliminating the need for inverse kinematics. In addition, a novel potential field-guided approach with local minima estimation is proposed to further reduce the computational overhead associated with path planning. Three manipulator models were tested in five simulated environments to evaluate tracking performance, obstacle avoidance, and computational efficiency. The results show that operational space controllers achieve comparable tracking performance to joint space controllers with lower computational demand, though at the expense of a higher failure rate. The potential field-guided control method performs effectively in simple environments, but struggles when the goal is obscured by closely spaced obstacles.