NeTS: Small: Advancing Time Synchronization for Sustainable Wireless Networks

Home

Projects

Publications

Testbeds

Software

People

Pictures

Supported by NSF CNS #1423379

Description

Time synchronization in sustainable wireless networks is expected to be the first class citizen due to the ever-increasing timeliness demands from resource-constrained systems. This is a significant challenge due to four major reasons: (1) Time synchronization services should adapt to their environment and system energy availability, (2) Duty-cycle operation at the MAC layer disrupts synchronization services, adversely affecting synchronization errors, (3) Multi-hop nature of time dissemination in sustainable wireless networks calls for scalable time synchronization solutions, and (4) Wide variety of applications result in diverse timing accuracy requirements. To address these challenges, time synchronization services should be an integral part of the networking stack. This project aims to fuse existing understanding of cross-layer networking with novel synchronization solutions.

 
The project has the potential to transform clockwork in several application domains
for sustainable wireless networks supports graduate students in an interdisciplinary environment, where commercialization is an integral part of the research activities. The
results of the project are disseminated to attract female students and minorities as a part of the Building Recruiting and Inclusion for Diversity (BRAID) initiative. Resources from Bright Lights and 4-H programs in Nebraska are leveraged to reach out to K-12 students
through easy-to-learn and fun activities. Finally, results from the proposed work are published in premier conference and journals; and developed software packages are disseminated through our web pages.
 

Goals

The project explores the following research goals:

  1. Energy-aware reactive time synchronization: As sensors are mainly developed for environmental monitoring tasks and are deployed in open environments, ambient factors significantly impact the operation of sensor nodes. To preserve the limited energy of a node, reactive synchronization activities should be adopted to reduce extra energy consumption. 
  2. Efficient synchronization under duty-cycle radio operation: Duty-cycling operations result in intermittent network connections and dynamic network topology, which could significantly impact the timely delivery of synchronization packets. Efficient and robust synchronization scheme should be designed to eliminate the impacts of the duty-cycled operations.
  3. Scalable time synchronization in large scale sustainable wireless sensor network: Traditional synchronization protocols do not scale in large scale sensor networks as synchronization are performed pair-wisely. Several concerns should be addressed for large scale network: 1) The time dissemination path should be reduced to avoid cumulative errors; 2) Compounding errors should be prevented; 3) Collective synchronization should be enabled through network time connectivity to efficiently achieve network consensus on time.
  4. On-demand synchronization with configurable QoS: Most of the existing synchronization techniques periodically disseminate time information. To increase energy efficiency by reducing communication overhead, synchronization can be optimally scheduled to adapt to the QoS demand.
  5. Cross-layer integration and optimization: Time information is needed in multiple network layers for specific tasks existed in different layers. Time sharing and transferring can be improved through cross-layer integration and optimization to improve energy efficiency, reduce latency and increase reliability.

Personnel

PI: Mehmet C. Vuran

Students: Fujuan Guo,

                Mohammad Lunar,

                Baofeng Zhou

Outcomes

Intellectual Merit Outcomes

  1. Timestamp-free syntonization: To our best knowledge, we designed the first work that leverages CFO in clock skew estimation for WSN clock synchronization. While removing timing uncertainty due to communication randomness, CFOSynt is impacted only from frequency fluctuations and phase-locked loop (PLL) phase noise, which is smaller in nature. We presented a theoretical analysis to capture these error sources. Through a novel experimental system, we analyze relationships between CFO and clock skew and validate the theoretical analysis using common o the shelf (COTS) wireless sensor platforms. The error analysis in this particular scenario has been presented for the first time based on empirical experiment results. Accordingly, we evaluate the clock skew estimation performance of CFOSynt.
  2. Correlation-based synchronization: We designed the first correlation-based clock synchronization protocol. Through extensive experiments, we showed with empirical evidence of the impacts of spatial correlation in temperature on clock skew. accordingly designed the first correlation based time synchronization solution that takes link-level synchronization solutions to a network-level application domain in a systematic way.
  3. On-demand optimal synchronization scheduling: We developed the first theoretical framework for optimizing synchronization scheduling based on stochastic network optimization theory. 
  4.  

Broader Impact Outcomes

  1. Concepts in distributed network synchronization have been integrated into an advanced embedded systems course with a course project. In collaboration UNL School of Music, interdisciplinary course modules are developed to educate students about details of percussion, timing, and their relations to distributed control and timing in embedded systems. As a part of the course project, students built a drum set that is automated through a wireless servo network.
  2. In collaboration UNL School of Music, interdisciplinary course modules were developed to educate students about details of percussion, timing, and their relations to distributed control and timing in embedded systems. In this investigation period, the PI has started teaching the second offering of this course this Fall. The course includes a course project, which uses the automated drum set that was developed in the previous offering of the course.
 

Resources

See the resource page.

Attachment Size
syntplatform.jpg 1.66 MB
cfosyntplatform2.jpg 888.06 KB