Page 25

EETE SEP 2015

of haptic feedback (maybe Peltier elements), the amputee could also feel temperature projected to his/her missing limb. Prior to this feat, Egger headed the research project “mind controlled prosthesis arm” at prosthetics manufacturer Otto Bock in cooperation with American research partners, the prosthesis later got the FDA approval (USA) in 2014. Although in this particular research, only the feedback channel was implemented (thus the patient cannot control the foot prosthesis by his mind like in other studies), theoretically, a two-way-connection, i.e., a motor-path (targeted muscle reinnervation) and a sensory-path (targeted sensory reinnervation) from the human body to the prosthesis is possible, explained Egger. Now, could future prostheses require a dedicated integrated circuit to figure out the best signal mapping, from the nerve terminations to the prosthesis and vice-versa? Or should the patient figure out how to make use of all the signals he generates (even through a form of sensory substitution) if the actual nerve terminations were not properly corresponding to the intended prosthesis sensing points? We asked Egger. “As sensory nerves are redirected to a small skin area (during sensory reinnervation), the different types of tactile sensory nerve fibers grow to their corresponding tactile skin receptors for vibration, temperature, force, or pain” detailed the researcher. “Then motor nerves can be redirected to residual muscles of the stump. EMG-Signals of the residual muscles (connected with the redirected residual motor nerves) produce EMG-Signals corresponding to the phantom limb movement, which only take place in the patient’s mind (thus mind controlled). EMG-signals can be picked up from the surface of the skin by sensitive differential amplifier (with a typical gain of 70 dB - 100 dB). But reinnervation surgery cannot be exactly planned. That means it is not possible to exactly predict which way the different types of nerve fibers use to get to their residual muscles and skin receptors, respectively. At the end the result for each reinnervation surgery is individual. But this is not a big problem from a technical point of view”. “For the control signals, pattern recognition methods (Support Vector Machines) are implemented to recognise the different movements of the phantom limb (thought movements, EMG-patterns). During an initial training procedure taking only a few minutes, the patient performs phantom movements (in his mind) according to the moves shown on a video. The corresponding EMG-patterns are picked up from the reinnervation area and stored into the prostesis’ embedded memory. After that training procedure, phantom movements can be recognised and artificial joints are thus controlled simultaneously, following the phantom limb’s movement”. “As for the feedback-signals, they are first applied by an array of actuators according to a test routine (actuators are switched on/off sequentially). During this initial test routine the patient has to decide which actuators he can feel best and then select them. This procedure takes a few minutes too, but once completed, the prosthesis is ready for use” concluded Egger. gri_EET EU+9.15.qxp_gri_eetEU_9.15 8/17/15 2:12 PM Page 1 Tough...But Oh, So Sensitive. Greenray TCXOs are tough as nails – but they’ve got a sensitive side, too. As in 7 x 10-11/g g-Sensitivity, for example. Our TCXOs are achieving new performance standards for phase noise, g-Sensitivity, and temperature stability, providing Aerospace, Defense and Commercial customers precision signal sources that work – on the ground, in the air, and most definitely, in motion. Designing for demanding Defense or Industrial applications? Let’s put Greenray innovation to work for you. Here are a few examples from our latest catalog: Best Stability ±0.5 ppm Output CMOS, Clipped Frequency 50 - 100 MHz Attributes Low Phase Noise Best Stability ±1.0 ppm Phase Noise-135 dBc/Hz @ 1kHz Output CMOS Size 17.27 x 17.27 x 5.0mm 0.68 x 0.68 x 0.20 in., SMD For Industry, for Defense. New Specification! Frequency 10 - 50 MHz Attributes Tight Stability Dual Compensation Best Stability ±0.03 ppm Output CMOS Squarewave Size 12.70 x 20.32x10.92 mm 0.50 x 0.80 x 0.43 in., SMD Frequency 10 - 50 MHz Attributes g-Sensitivity to 7 x 10-11/g Sine Size 9.0 x 7.0 x 3.7 mm 0.35 x 0.28 x 0.15 in., SMD www.electronics-eetimes.com Electronic Engineering Times Europe September 2015 23


EETE SEP 2015
To see the actual publication please follow the link above