Band 4, Heft 1

This paper presents a two layered control architecture - Superior hand control (SHC) followed by Local hand control (LHC) for an extreme upper limb prosthesis. The control architecture is for executing grasping operations involved in 70% of daily living activities. Forearm electromyogram actuated SHC is for recognition of user’s intended grasp. LHC control the fingers to be actuated for the recognized grasp. The finger actuation is controlled through a proportionalintegral- derivative controller customized with fingertip force sensor. LHC controls joint angles and velocities of the fingers in the prosthetic hand. Fingers in the prosthetic hand emulate the dynamic constraints of human hand fingers. The joint angle trajectories and velocity profiles of the prosthetic hand finger are in close approximation to those of the human finger

This article describes the design and evaluation of AIBOStory - a novel, remote interactive story telling system that allows users to create and share common stories through an integrated, autonomous robot companion acting as a social mediator between two remotely located people. The behaviour of the robot was inspired by dog behaviour, including a simple computational memory model. AIBOStory has been designed to work alongside online video communication software and aims to enrich remote communication experiences over the internet. An initial pilot study evaluated the proposed system’s use and acceptance by the users. Five pairs of participants were exposed to the system, with the robot acting as a social mediator, and the results suggested an overall positive acceptance response. The main study involved long-term interactions of 20 participants using AIBOStory in order to study their preferences between two modes: using the game enhanced with an autonomous robot and a non-robot mode which did not use the robot. Instruments used in this study include multiple questionnaires from different communication sessions, demographic forms and logged data from the robots and the system. The data was analysed using quantitative and qualitative techniques to measure user preference and human-robot interaction. The statistical analysis suggests user preferences towards the robot mode.

In this paper, we propose an effective planning method for whole-body motions of humanoid robots under various conditions for achieving the task. In motion planning, various constraints such as range of motion have to be considered. Specifically, it is important to maintain balance in whole-body motion. In order to be useful in an unpredictable environment, rapid planning is an essential problem. In this research, via-point representation is used for assigning sufficient conditions to deal with various constraints in the movement. The position, posture and velocity of the robot are constrained as a state of a via-point. In our algorithm, the feasible motions are planned by modifying via-points. Furthermore, we formulate the motion planning problem as a simple iterative method with a Linear Programming (LP) problem for efficiency of the motion planning. We have applied the method to generate the kicking motion of a HOAP-3 humanoid robot. We confirmed that the robot can successfully score a goal with various courses corresponding to changing conditions of the location of an obstacle. The computation time was less than two seconds. These results indicate that the proposed algorithm can achieve efficient motion planning.

The field of mobile robotic telepresence for social communication is in rapid expansion and it is of interest to understand what promotes good interaction. In this paper, we present the results of an experiment where novice users working in health care were given a guided tour while maneuvering a mobile robotic telepresence system for the first time. In a previous study, it was found that subjective presence questionnaires and observations of spatial configurations based on Kendon’s F-formations were useful to evaluate quality of interaction in mobile robotic telepresence. In an effort to find more automated methods to assess the quality of interaction, the study in this paper used the same measures, with the addition of objective sociometric measures. Experimental results show that the quantitative analysis of the sociometric data correlates with a number of parameters gathered via qualitative analysis, e.g. different dimensions of presence and observed problems in maneuvering the robot.

Policy improvement methods seek to optimize the parameters of a policy with respect to a utility function. Owing to current trends involving searching in parameter space (rather than action space) and using reward-weighted averaging (rather than gradient estimation), reinforcement learning algorithms for policy improvement, e.g. PoWER and PI 2 , are now able to learn sophisticated high-dimensional robot skills. A side-effect of these trends has been that, over the last 15 years, reinforcement learning (RL) algorithms have become more and more similar to evolution strategies such as (μW , λ)-ES and CMA-ES. Evolution strategies treat policy improvement as a black-box optimization problem, and thus do not leverage the problem structure, whereas RL algorithms do. In this paper, we demonstrate how two straightforward simplifications to the state-of-the-art RL algorithm PI 2 suffice to convert it into the black-box optimization algorithm (μW, λ)-ES. Furthermore, we show that (μW , λ)-ES empirically outperforms PI 2 on the tasks in [36]. It is striking that PI 2 and (μW , λ)-ES share a common core, and that the simpler algorithm converges faster and leads to similar or lower final costs. We argue that this difference is due to a third trend in robot skill learning: the predominant use of dynamic movement primitives (DMPs). We show how DMPs dramatically simplify the learning problem, and discuss the implications of this for past and future work on policy improvement for robot skill learning