Small-Scale Systems

                                 Team and Organization
Motivation: While the "Large-Scale Systems" theme address the energy-smart systems spanning large areas, this theme concentrates on the opposite corner of the distributed system spectrum. It explores the opportunity of developing sensing systems that zoom in to the microscopic level. It is only in the last couple of years that the semiconductor and IT industries (taking the lead from the defense community) started to realize the business potential of so-called "cyber-physical" systems, which provide the interface between the physical, human and the computational worlds. Potentially consisting of trillions of wireless sensor nodes distributed in the environment, these systems address large-scale problems such as security, defense, environment, energy, disaster management, health, and generally quality of life. While major progress was made over the past decade, a broad range of true opportunities is still out of reach. Reducing energy dissipation and size of the wireless sensory systems by even one or two orders of magnitude would open the door for some large-impact applications such as intelligent materials or surfaces, smart objects that are aware and adjust to their surrounding conditions (think about intelligent car tires, for instance), and enriched human sensory and monitoring interface systems (enabled by either body-area or implanted networks). The prospect of ubiquitous sensing at all scales on the battlefield could also revolutionize military capabilities and systems. For instance, personal sensor networks may extend and improve the sensory field of individual soldiers. Microscopic sensing systems present by far the greatest opportunity for new markets and growth in the semiconductor industry, as they open potential markets for literally trillions of circuits - but only if orders of magnitude reduction in energy dissipation and size beyond current systems is achieved.

Energy is indeed THE overriding concern in the design and deployment for these microscopic systems. In contrast to the large-scale systems, they must be "energy-frugal". While some nodes may carry batteries, others must be totally energy self-contained (a single battery charge must suffice for the life-time of the project, or the node must replenish its energy reservoir through scavenging). For example, current "personal" surveillance and environment-awareness systems (as carried by soldiers in the field) come with a prohibitive battery load which now approaches 20 lbs or more for typical missions; increasing the capability and performance of the electronic systems available to the individual soldier now depends first and foremost on radical improvements in energy efficiency of small-scale systems.

Vision: Simple solutions to the energy and related challenges are not in sight. The International Technology Roadmap for Semiconductors states that for SOC consumer portable chips "simply extrapolating from current state of the art technology … the resulting power consumption substantially exceeds the requirements". Moore's Law or business-as-usual will not come to the rescue and enable game-changing microscopic systems. Fundamental new approaches are needed. They require the adoption of truly innovative technologies in terms of sensing, data processing, communication and energy harvesting, all integrated using aggressive heterogeneous packaging strategies. Yet as in the large systems space, the true opportunities lie in a holistic system vision. A distributed energy-driven management strategy, adaptively selecting the right information, data resolution and rate, signal representation, communication bandwidth, and data latency over a broad dynamic range can help to provide the desired functionality at an absolute minimal energy cost. Plenty of inspiration for such "attention-optimized" multi-scale systems can be found in the biological world. The unified multi-scale system design strategies, advocated in this proposal, are hence an instrumental part of the solution; there is simply no other place to turn for the needed improvements.

The small-scale system theme of MuSyC is hence built on a dual foundation. Overlaying it is a distributed multi-scale dynamic run-time environment, operating over many layered abstractions and merging sensing, computing, control and actuation (as advocated in the SCS Theme), inspired by meaningful systems metrics (utility). The actual realization of this environment requires an implementation platform that fully supports the concept of attention-optimization over a broad dynamic range, while being extremely energy-frugal at the same time. The multi-scale aspect of the small-scale system theme primarily is in the combination of multi-layered abstractions, combined with the heterogeneous integration of electronic, mechanical, chemical and biological components into a single entity. Its innovative agenda is to explore the absolute bounds of energy-efficiency and miniaturization by simultaneous and dynamic exploration of the system, architecture, platform and technology layers. It is most plausible that ideas emerging from this arena will ultimately impact computation and communication at the larger scales as well. The exciting area of biological interfaces to support personal-area sensory enhancement (including brain-machine interfaces or BMI) has been chosen as the application driver for this theme (in synergy with activities at the Platform, Module, and Connectivity centers).