Socially assistive robot with interactive behavioral capability have been improving quality of life for a wide range of users by taking care of elderlies, training individuals with cognitive disabilities or physical rehabilitation, etc. While the interactive behavioral policies of most systems are scripted, we discuss here key features of a new methodology that enables professional caregivers to teach a socially assistive robot (SAR) how to perform the assistive tasks while giving proper instructions, demonstrations and feedbacks. We describe here how socio-communicative gesture controllers-which actually control the speech, the facial displays and hand gestures of our iCub robot-are driven by multimodal events captured on a professional human demonstrator performing a neuropsychological interview. Furthermore, we propose an original online evaluation method for rating the multimodal interactive behaviors of the SAR and show how such a method can help designers to identify the faulty events.

Human interactions are driven by multi-level perception-action loops. Interactive behavioral models are typically built using rule-based methods or statistical approaches such as Hidden Markov Model (HMM), Dynamic Bayesian Network (DBN), etc. In this paper, we present the multimodal interactive data and our behavioral model based on recurrent neural networks, namely Long-Short Term Memory (LSTM) and Bidirectional LSTM (BiLSTM) models. Speech, gaze and gestures of two subjects involved in a collaborative task are here jointly modeled. The results show that the proposed deep neural networks are more effective than the conventional statistical methods in generating appropriate overt actions for both on-line and off-line prediction tasks.

Residual vibration of system base due to high-speed motion of a stage may significantly reduce life span and productivity of the manufacturing equipment. Although a passive reaction force compensation (RFC) mechanism was developed to reduce residual vibration of a linear motor motion stage, the passive RFC mechanism should be redesigned according to variation of motion profiles. In this paper, we develop a fuzzy-P (proportional) controller of an active RFC mechanism to automatically tune the gain according to variation of motion profiles. First, frequency components of motion profiles for a linear motion stage are analyzed and performances of the passive RFC mechanism are approximately evaluated using a motion profile analysis. An active RFC mechanism with an additional control coil is introduced to overcome limitation of the passive RFC mechanism and a fuzzy rule is proposed to automatically tune the P controller of the active RFC mechanism according to variations of motion profiles. Simulations and experiments are performed to show effectiveness of the proposed fuzzy rule for tuning the P control gain of the active RFC mechanism. NOMENCLATURE a M = acceleration of mover c MT = damping of magnet track k MT = stiffness of magnet track k v = viscous friction of mover k P = gain of P control F f = thrust force for mover friction F fc = Column friction of mover F fs = Stribeck friction of mover F m = thrust force for mover motion F t = thrust or reaction force F tran = transmitted force m M = mass of mover m MT = mass of magnet track n = index of frequency component n s = Stribeck friction coefficient for exponential function n max = index of frequency component at peak acceleration of mover n MT = index of natural frequency of magnet track n' MT = output of defuzzification block T = period of motion profile T a = acceleration time T dw = dwell time T r = run time v s = Stribeck friction coefficient for velocity x MT = position of magnet track = velocity of magnet track = acceleration of magnet track ω = frequency ω MT = natural frequency of magnet track ω T = fandamental frequency of motion profile ζ = damping ratio θ n = phase of magnet track motion x · MT x ··

No abstract

This article presents an eddy current damper of a passive reaction force compensation mechanism for a linear motor motion stage. The reaction force compensation mechanism with a movable magnet track and eddy current damper resolves problems with the existing spring-based reaction force compensation mechanism such as resonance, design freedom, and difficulty of assembly and manufacturing. A simplified mathematical model of the eddy current damper is derived considering the sinusoidal magnetic flux density distribution and effective width of the eddy current damper, which shows important design factors of the eddy current damper–based reaction force compensation mechanism. Then, force of the eddy current damper according to the constant speed motion of the magnet track is investigated using multi-physical finite element analysis and is verified by experiments. Finally, the passive reaction force compensation with movable magnet track and eddy current damper is identified by experiments, and the finite element analysis of the eddy current damper is verified with free and forced vibration responses.