Energy is the most critical resource in wireless sensor networks. Its coupling to the availability of precise timing is often converse -- higher time precision lowers the communication energy due to smaller guard-bands, but increases the energy of the timing subsystem. For example, the finer the time granularity, the faster the clock necessary to get to that resolution, and thus the higher the frequency of the digital circuit whose power drain increases linearly with the increase in frequency. Thus, it seems that high-resolution, low-power time synchronization is an oxymoron.
In this talk I will outline three technical challenges to balance time and energy -- frequency stability compensation, time information dissemination, and clock scheduling and duty-cycling -- and present Temperature Compensated Time Synchronization (TCTS), Time Information Routing Protocol (TIRP), and Virtual High-Resolution Time (VHT).
TCTS exploits the on-board temperature sensor existing in many sensor network platforms. It uses this temperature sensor to autonomously calibrate the local oscillator and removes effects of environmental temperature changes. This allows a time synchronization protocol to increase its resynchronization period, without losing synchronization accuracy, and thus saves energy and communication overhead.
TIRP addresses the problem of how time information gets disseminated in a multi-hop network of sensor nodes. While flooding or routing integrated time synchronization initially seems like a good idea, we show that one unstable node can introduce large time error ripples.
TIRP addresses this problem by realizing that time information is not correlated to the communication link quality. By means of a time quality metric, similar to ETX in routing, TIRP creates a time dissemination tree that minimizes synchronization error propagation.
VHT is a power-proportional time-keeping service that offers a baseline power draw of a low-speed clock (e.g. 32 kHz crystal), but provides the time resolution that only a higher frequency clock could offer (e.g. 8 MHz crystal), and scales essentially linearly with access (i.e. the “reading” and “writing” of the clock). We achieve this performance by revisiting a basic assumption in the design of time-keeping systems -– that to achieve a given time-stamping resolution, a free-running timebase of equivalent frequency is needed.
The key technical challenge lies in duty cycling the fast clock and synchronizing the fast and slow clocks. Our results show power-proportional operation with a 10× improvement in average power and a synchronization accuracy exceeding 1 μs at duty cycles below 0.1%
The combination of TCTS, TIRP, and VHT allow low-power time synchronization at an unprecedented accuracy and precision for wireless networks. These services will spur new system architectures allowing server farms, buildings, or transportation to be more energy efficient.