Representation and control of coordinated-motion tasks for human-robot systems
It is challenging for robots to perform various tasks in a human environment. This is because many human-centered tasks require coordination in both hands and may often involve cooperation with another human. Although human-centered tasks require different types of coordinated movements, most of the existing methodologies have focused only on specific types of coordination. This thesis aims at the description and control of coordinated-motion tasks for human-robot systems; i.e., humanoid robots as well as multi-robot and human-robot systems. First, for bimanually coordinated-motion tasks in dual-manipulator systems, we propose the Extended-Cooperative-Task-Space (ECTS) representation, which extends the existing Cooperative-Task-Space (CTS) representation based on the kinematic models for human bimanual movements in Biomechanics. The proposed ECTS representation can represent the whole spectrum of dual-arm motion/force coordination using two sets of ECTS motion/force variables in a unified manner. The type of coordination can be easily chosen by two meaningful coefficients, and during coordinated-motion tasks, each set of variables directly describes two different aspects of coordinated motion and force behaviors. Thus, the operator can specify coordinated-motion/force tasks more intuitively in high-level descriptions, and the specified tasks can be easily reused in other situations with greater flexibility. Moreover, we present consistent procedures of using the ECTS representation for task specifications in the upper-body and lower-body subsystems of humanoid robots in order to perform manipulation and locomotion tasks, respectively. Besides, we propose and discuss performance indices derived based on the ECTS representation, which can be used to evaluate and optimize the performance of any type of dual-arm manipulation tasks. We show that using the ECTS representation for specifying both dual-arm manipulation and biped locomotion tasks can greatly simplify the motion planning process, allowing the operator to focus on high-level descriptions of those tasks. Both upper-body and lower-body task specifications are demonstrated by specifying whole-body task examples on a Hubo II+ robot carrying out dual-arm manipulation as well as biped locomotion tasks in a simulation environment. We also present the results from experiments on a dual-arm robot (Baxter) for teleoperating various types of coordinated-motion tasks using a single 6D mouse interface. The specified upper- and lower-body tasks can be considered as coordinated motions with constraints. In order to express various constraints imposed across the whole-body, we discuss the modeling of whole-body structure and the computations for robotic systems having multiple kinematic chains. Then we present a whole-body controller formulated as a quadratic programming, which can take different types of constraints into account in a prioritized manner. We validate the whole-body controller based on the simulation results on a Hubo II+ robot performing specified whole-body task examples with a number of motion and force constraints as well as actuation limits. Lastly, we discuss an extension of the ECTS representation, called Hierarchical Extended-Cooperative-Task Space (H-ECTS) framework, which uses tree-structured graphical representations for coordinated-motion tasks of multi-robot and human-robot systems. The H-ECTS framework is validated by experimental results on two Baxter robots cooperating with each other as well as with an additional human partner.
Lee, Purdue University.
Off-Campus Purdue Users:
To access this dissertation, please log in to our