Muscular activity when walking in a non-anthropomorphic wearable robot.

Wearable robots should be designed not to alter human physiological motion. Perturbations introduced by a robot can be quantified by measuring EMG activity. This paper presents tests on the LENAR, an intrinsically back-drivable non-anthropomorphic lower limb wearable robot designed to provide hip and knee flexion/extension assistance. In previous works the robot was demonstrated to exhibit low mechanical impedance and to introduce minor alterations to human kinematic patterns during walking. In this paper muscular activity is assessed, demonstrating small alterations in the EMG patterns during the interaction with the robot, in both unpowered and assistive mode.

Optimization of kinetic energy harvesters design for fully implantable Cochlear Implants.

Fully implantable Cochlear Implants (CIs) would represent a tremendous advancement in terms of quality of life, comfort and cosmetics, for patients with profound sensorineural deafness. One of the main challenges involved in the development of such implants consists of finding a power supply means which does not require recharging. To this aim an inertial Energy Harvester (EH), exploiting the kinetic energy produced by vertical movements of the head during walking, has been investigated. Compared to existing devices, the EH needs to exploit very low frequency vibrations (<2.5 Hz) with small amplitude (<9 m/s(2)). In order to maximize the power transduced, an optimization method has been developed, which is the objective of this paper. The method consists in calculating the dynamical behavior of the EH using discrete transforms of experimentally measured acceleration profiles. It is shown that the quick integration of the second order dynamical equation allows the use of computationally intensive optimization techniques, such as Genetic Algorithms (GAs). The robustness of the solution is also evaluated.