Initial Steps Toward Ubiquitous Biomedical Implantable System-on-Chip – An Architectural Perspective

Tuesday October 30, 2007
Hamerschlag Hall D-210
4:30 pm

Allen Cheng
University of Pittsbugh

The idea of implanting micro/nano-scale computer chips inside a human body for the purpose of collecting, storing, processing, and communicating biological data ubiquitously for clinical and biomedical applications is giving the society and its people a revolutionary break-through perspective on how next-generation computing can evolve to advance the state and quality of healthcare in the twenty-first century. These biomedical implantable chips, once realized, can fundamentally change the way we implement monitoring, diagnosing, and combating diseases. To date, successful deployment is limited only to implanting mainly plain RFID tags for storing brief personal identification, contact information, and some personal medical history. Most of these RFID tags are passive, read-only devices; they do not act until being acted upon by some other external devices. One of the main reasons that prohibit us from obtaining more sophisticated implantable devices is due to the problem of "energy wall" – today, even the most advanced state-of-the-art fuel cell technology can not sustain any existing CPU long enough to avoid frequent battery replacement or recharging; the power available through existing energy scavengers and RF-based reader fields is orders of magnitude less than what is required to support even the simplest general-purpose low-power processor. Because of this huge energy demand-supply gap, only simple, unsophisticated IC chip with very little processing power, if any, can be implanted, which has significantly limited how these chip-based implants can be used. In this talk, I will be discussing some of the issues and challenges in designing processors for biomedical implantable systems. Then, I will describe a promising architecture technique capable of significantly improving energy-efficiency of a processor so they can be more adequately powered and correctly functioning in a constrained implant environment by current and emerging energy-supply technologies in a foreseeable future.

Professor Allen C. Cheng is an Assistant Professor in the Department of Electrical and Computer Engineering, Department of Neurological Surgery, and Department of Computer Science at University of Pittsburgh. He received his Ph.D. and M.S. in Computer Science and Engineering from the University of Michigan at Ann Arbor and his B.S. in Computer Engineering from the North Carolina State University. Prior to join the University of Pittsburgh, he was a faculty in the College of Engineering at the University of Michigan at Ann Arbor. His research interest is in the area of biomedical computing and implantable system-on-chip (SoC) with emphasis on energy-efficient, reliable, secure, and high-performance computer system architecture, processor microarchitecture, and compiler optimizations. He is a member of several international scholar and honor societies including Eta Kappa Nu, Phi Eta Sigma, and Golden Key. He is also a member of IEEE, ACM, and AAAS.