Dirk Ruiken

Postural modes and control for dexterous mobile manipulation: the UMass uBot concept

We present the UMass uBot concept for dexterous mobile manipulation. The uBot concept is built around Bernsteins definition of dexterity-“the ability to solve a motor problem correctly, quickly, rationally, and resourcefully” [1]. We contend that dexterity in robotic platforms cannot arise from control alone and can only be achieved when the entire design of the robot affords resourceful behavior. uBot-6 is the latest robot in the uBot series whose design affords several postural configurations and mobility modes. We discuss these dexterous mobility options in detail and demonstrate the strength of dexterous mobility.

Log-space harmonic function path planning

We propose a log-space solution for robotic path planning with harmonic functions that solves the long-standing numerical precision problem. We prove that this algorithm: (1) performs the correct computations in log-space, (2) returns the true equivalent path using the log-space mapping, and (3) has a strong error bound given its convergence criterion. We evaluate the algorithm on 7 problem domains. A Graphics Processing Unit (GPU) implementation is also shown to greatly improve performance. We also provide an open source library entitled epic with extensive ROS support and demonstrate this method on a real humanoid robot: the uBot-6. Experiments demonstrate that the log-space solution rapidly produces smooth obstacle-avoiding trajectories, and supports planning in exponentially larger real-world robotic applications.

Affordance-based Active Belief: Recognition using visual and manual actions

This paper presents an active, model-based recognition system. It applies information theoretic measures in a belief-driven planning framework to recognize objects using the history of visual and manual interactions and to select the most informative actions. A generalization of the aspect graph is used to construct forward models of objects that account for visual transitions. We use populations of these models to define the belief state of the recognition problem. This paper focuses on the impact of the belief-space and object model representations on recognition efficiency and performance. A benchmarking system is introduced to execute controlled experiments in a challenging mobile manipulation domain. It offers a large population of objects that remain ambiguous from single sensor geometry or from visual or manual actions alone. Results are presented for recognition performance on this dataset using locomotive, pushing, and lifting controllers as the basis for active information gathering on single objects. An information theoretic approach that is greedy over the expected information gain is used to select informative actions, and its performance is compared to a sequence of random actions.

A Compact, Modular Series Elastic Actuator

This paper presents the development of a compact, modular rotary series elastic actuator (SEA) design that can be customized to meet the requirements of a wide range of applications. The concept incorporates flat brushless motors and planetary gearheads instead of expensive harmonic drives and a flat torsional spring design to create a lightweight, lowvolume, easily reconfigurable, and relatively high-performance modular SEA for use in active impedance controlled devices. The key innovations include a Hall effect sensor for direct spring displacement measurements that mitigate the negative impact of backlash on SEA control performance. Both torque and impedance controllers are developed and evaluated using a 1-degree-of-freedom (DoF) prototype of the proposed actuator package. The results demonstrate the performance of a stable first-order impedance controller tested over a range of target impedances. Finally, the flexibility of the modular SEA is demonstrated by configuring it for use in five different actuator specifications designed for use in the uBot-7 mobile manipulator requiring spring stiffnesses from 3 N m/deg to 11.25 N m/deg and peak torque outputs from 12 N m to 45 N m. [DOI: 10.1115/1.4032975]

Choosing informative actions for manipulation tasks

Autonomous robots demand complex behavior to perform tasks in unstructured environments. In order to meet these expectations efficiently, it is necessary to organize knowledge of past interactions with the world in order to facilitate future tasks. With this goal in mind, we present a knowledge representation that makes explicit the invariant spatial relationships between sensorimotor features comprising a rigid body and uses them to reason about other tasks and run-time contexts.

Reconfigurable tasks in belief-space planning

We propose a task representation for use in a belief-space planning framework. The representation is based on specialized object models that enable estimation of an abstract state of a robot with respect to an object. Each manipulation task is represented using a partition over these states defined by the set of known object models. Solutions to such tasks are constructed in a belief-space planner using visual and/or manual interactions with objects that condense belief in a target subset of the task partition. This partition integrates belief over states into a task belief without altering the original belief representation. As a result, sequences of tasks can be addressed that inherit the complete estimate of state over the entire history of observations. Demonstrations of the technique are presented in simulation and on a real robot. Results show that using this task representation and the belief-space planner, the robot is able to recognize objects, find target objects, and manipulate a set of objects to obtain a desired state.