INVERSE KINEMATICS, CONTROL AND PATH PLANNING OF MOBILE MANIPULATOR

Abstract

Mobile manipulators offer better maneuverability over conventional fixed base manipulators. Since the robotic manipulator is mounted over a mobile base, the mobile manipulator is not only versatile in its application but also poses many challenges. Generally, the mobile manipulator is a redundant system, therefore, inverse kinematics and path planning become important issues. The dissimilar dynamics of mobile base and manipulator arm make the system as an interesting topic of research in robotic community. Also, dynamic interactions between mobile robot base and manipulator arm require extra attention for its control. To this end, this thesis attempts to study and address these issues of mobile manipulator. In this thesis a three wheeled mobile robot with Omni-directional wheels has been considered as mobile robot base. The inverse kinematics of mobile manipulator is a crucial task due to redundancy in the system. This thesis introduces a fast converging inverse kinematic solution using particle swarm optimization technique. A bidirectional search method based on manipulator decoupling technique is employed in obtain the inverse kinematics solution. A collision avoidance algorithm for line type obstacles is also proposed for finding inverse kinematic solution. Case studies of four degree of freedom planar manipulator mounted over a mobile base are discussed for various combinations of decoupling. Further simulations have been performed to demonstrate the efficacy of the proposed inverse kinematic algorithm both for cluttered environment and for obstacle free environment. The results obtained from these simulations suggest that the proposed inverse kinematic solution through manipulator decoupling gives faster convergence. This thesis discusses development of dynamic model of mobile manipulator using bond graph. The dynamic model helps to see observe the dynamic interaction effect of manipulator arm movement on mobile base. Tip trajectory control schemes of mobile manipulator are proposed. The developed control is a combination of conventional PID and a novel amnesia recovery control. Jacobian transpose method is used to convert the control force on the tip xv to torque on the joints. Two control strategies viz. only manipulator arm control and simultaneous manipulator and mobile base control are discussed. The primary objective of these control schemes is tip trajectory control while nullifying the errors occurred in tip trajectory due to dynamic interaction between mobile robot base and manipulator. Simulations have been performed to verify the effectiveness of the proposed control strategies. The simulation results suggest that simultaneous control of both mobile base and manipulator gives better trajectory control performance. Another key issue is development of system which can tolerate the impending faults in the system. The ability of the system to tolerate unexpected faults, increases its autonomous capabilities. The fault tolerant control system is designed to keep the system performance intact even in case of locked joint failure. This thesis discusses the fault tolerant control of mobile manipulator using the redundancy in the system. The closed loop inverse kinematics method for finding the inverse kinematics of redundant robots has been extended for mobile manipulator. Performance of the fault tolerant system is evaluated using the redundancy in the manipulator and mobile base in case of joint lock failure of manipulator. The concept of Model based control system is developed based on the inverse dynamics. Model based control gives better performance than traditional controllers like PID for control of nonlinear system. This thesis makes an attempt to devise the reduced model based control of mobile manipulator. The reduced model based control is developed by simplification of original model. Model order reduction algorithm is employed to identify and remove the low active elements of the model. Finally, this thesis talks about path planning of mobile manipulator. Particle swarm optimization algorithm is employed to obtain energy optimal path for the tip of mobile manipulator.