Selection and control of limb posture for stability.

Impedance control can be used to stabilize the limb against both instability and unpredictable perturbations. Limb posture influences motor noise, energy usage and limb impedance as well as their interaction. Here we examine whether subjects use limb posture as part of a mechanism to regulate limb stability. Subjects performed stabilization tasks while attached to a two dimensional robotic manipulandum which generated a virtual environment. Subjects were instructed that they could perform the stabilization task anywhere in the workspace, while the chosen postures were tracked as subjects repeated the task. In order to investigate the mechanisms behind the chosen limb postures, simulations of the neuro-mechanical system were performed. The results indicate that posture selection is performed to provide energy efficiency in the presence of force variability.

Experimental investigation of antenna selection and transmit beamforming for capacity enhancement in 3×3 MIMO-enabled radio-over-fiber DAS

This paper examines the impact of two simple precoding schemes on the capacity of 3 × 3 MIMO-enabled radio-over-fiber (RoF) distributed antenna systems (DAS) with excess transmit antennas. Specifically, phase-shift-only transmit beamforming and antenna selection are compared. It is found that for two typical indoor propagation scenarios, both strategies offer double the capacity gain that non-precoding MIMO DAS offers over traditional MIMO collocated antenna systems (CAS), with capacity improvements of 3.2-4.2 bit/s/Hz. Further, antenna selection shows similar performance to phase-only beamforming, differing by <0.5% and offering median capacities of 94 bit/s/Hz and 82 bit/s/Hz in the two propagation scenarios respectively. Because optical DASs enable precise, centralized control of remote antennas, they are well suited for implementing these beamforming schemes. Antenna selection, in particular, is a simple and effective means of increasing MIMO DAS capacity. © 2013 IEEE.

Decision tree-based context clustering based on cross validation and hierarchical priors

The standard, ad-hoc stopping criteria used in decision tree-based context clustering are known to be sub-optimal and require parameters to be tuned. This paper proposes a new approach for decision tree-based context clustering based on cross validation and hierarchical priors. Combination of cross validation and hierarchical priors within decision tree-based context clustering offers better model selection and more robust parameter estimation than conventional approaches, with no tuning parameters. Experimental results on HMM-based speech synthesis show that the proposed approach achieved significant improvements in naturalness of synthesized speech over the conventional approaches. © 2011 IEEE.