The topic of my thesis work as part of my M.S. degree in Electrical Engineering under the advice of Professor Darrin Young was the design of a wireless electromyogram sensor for
control of lower extremity prosthesis. My project was in collaboration
with Dr. Ronald Triolo of the Cleveland APT Center.
I attended the 2008 IEEE Sensors Conference in Lecce, Italy, where
I won 1st prize in the Best Student Paper Competition for my paper, Wireless Implantable EMG Sensing Microsystem. I also won the grad prize at the Case
Research Showcase 2008 for a poster about the same work.
Wireless EMG Sensing Microsystem for Prosthetic Control
From my IEEE Sensonrs Conference Abstract:
This paper presents a wireless, subfascially implantable electromyogram (EMG)-sensing microsystem design for intelligent myoelectric control of powered prostheses. The implantable system consists of two Pt-Ir epimysial EMG electrodes, a custom-designed ASIC, and an RF telemetry coil and is capable of wirelessly transmitting digitized EMG data to an external telemeter mounted in a prosthetic socket. The prototype microsystem is powered by a near-field inductive link operating at 8 MHz with 10% DC power transfer efficiency. On-chip rectification and regulation produce stable 2.1 V and 2.7 V supplies with a DC current driving capability up to 100 uA. The EMG electrodes are interfaced with a differential capacitively-coupled amplifier with 38 dB closed-loop gain, 1 kHz bandwidth, and 78 nV/rtHz input-referred noise floor. The amplified EMG signal is then digitized on chip using an 11-bit algorithmic ADC. The digital EMG data can be Manchester-coded and transmitted to the external telemeter using phase shift keying (PSK) modulation scheme on the same wireless link as the inductive powering system.
Myoelectric signals are critical natural control sources for enhanced biomimetic performance of powered prosthetic lower limbs. Previous implementations of myoelectrically controlled prosthetic lower limbs have relied on surface EMG recorded using either electrodes in the prosthetic socket or by using skin patch electrodes attached to the residual limb. The former option precludes the use of a prosthetic sock, which causes uncomfortable friction between the socket and the residual limb. The latter option requires the user to prepare the skin surface and attach the electrodes and lead wires daily. Both options are prone to motion artifacts and poor connections due to perspiration. A fully implantable EMG sensor obviates these issues by measuring muscle activity directly from the muscle surface, allowing for the use of a prosthetic sock, and eliminating any additional steps when donning the prosthesis.
Integrated EMG Sensing Electronics