This paper presents a framework to address three dimensional (3-D) dynamic walking for a bipedal exoskeleton with underactuated legs. To achieve this goal, the framework is constructed via a trajectory generator and an optimized inverse kinematics algorithm that can cope with underactuation. In order to feasibly attain task velocities with underactuated legs, the inverse kinematics algorithm makes use of a task prioritization method via the exploitation of null space. In doing so, the tasks with lower priority, e.g., swing foot orientation, are attained as much as possible without disrupting the higher priority tasks, such as, CoM trajectory. Meanwhile, the trajectory generator utilizes the ZMP concept analytically and ensures the acceleration continuity throughout the whole walking period, regardless of the contact and phase changes. The proposed method is verified via a lumped human-bipedal exoskeleton model that is developed and simulated in MSC.ADAMS simulation environment. As the result, we obtained feasible and dynamically balanced 3-D walking motion, in which no oblique foot landing or exaggerated torso orientation variations were observed, despite the underactuated nature of the robot legs.

No abstract

In this paper, we present our research study concerning the design and development of an exoskeleton that aims to provide 3D walking support with minimum number of actuators. Following a prior simulation study, the joint configuration was primarily determined. In order for the exoskeleton to possess advanced characteristics, the following design criteria were investigated: i) all the actuators (hip/knee/ankle) were deployed around the waist area to decrease leg weight and improve wearability, ii) custom-built series elastic actuators were used to power system for high fidelity torque-controllability, iii) 3D walking support is potentially enabled with reduced power requirements. As a result, we built the first actual prototype to experimentally verify the aforementioned design specifications. Furthermore, the preliminary torque control experiments indicated the viability of torque control.

This paper presents a framework to optimally tune the stiffness values of spring-supported passive joints that are included in lower body exoskeletons. First, a dynamic model of a combined human-exoskeleton system was created in MSC.ADAMS software. Second, a gradient-descent based algorithm was used to find the optimum value to minimize the ZMP for a range of ankle stiffness values. In order to corroborate the proposed method, simulation experiments were conducted by considering three cases in which different body mass and heights were assigned to the combined humanexoskeleton system. The simulation results indicate that the proposed methodology is effective in order to find the optimum ankle stiffness for the combined human-exoskeleton systems, resulting in reductions in ZMP variations and therefore increasing the balancing ability. As a consequence, it may be possible to reduce the number of active joints in exoskeletons that aim crutch-free 3-D walking motion support.

No abstract

No abstract