Cyber-physical systems research to date has focused on the development of synergy and tight coupling of the physical and computational processes, vis-a-vis, the control of the system. However, this tight coupling is also enabling the accumulation of large amounts of data, which can be analyzed, interpreted, and appropriately leveraged. When multiple systems are interacting with each other, and closed-loop control is implemented, real-time analysis of the large amount of cross-device data becomes a critical requirement. As we evolve towards the Internet of Things, we see the deployment of multitude of wireless sensors and agents spanning many application domains including: environmental, healthcare, avionics components in the latest commercial airplane, smart interconnected automobiles and trucks, and smart buildings. These produce massive amounts of multisystem data that need to be sifted through to facilitate reasonably accurate decision-making and control.  

We convened the community for a workshop to explore challenges and opportunities in moving from IoT to real-time control and CPS.  This workshop focused on the current state of big data real time analytics in CPS (in medical, transportation (automotive, aerospace, rail), energy, and other fields), evaluated the promises and shortcomings, and evaluated what needs to be done to enable big data real-time analytics in closed-loop cyber-physical systems.

Topics of discussion included:

• As we move from IoT sensing to real-time applications, the need for dependability and security emerges, what technologies and research are essential?  What are the challenges?  How do approaches scale?
• How do you integrate IoT and big data into cloud for CPS and get real-time control?
• Machine learning and other approaches for real time data analytics in closed-loop CPS
• Physical/virtual testbeds for real-time closed-loop control
• Real-time meanings in various application spaces: aeronautics, medical, transportation, energy
• Opportunities for closing the loop at (or near) real-time in “smart city” cyber-physical systems
• Multisystem data analytics: how do we ensure data from multiple systems are input correctly and at right times?
• Integrating provenance into IoT and CPS.
• Real-time sense making and decision making with big data including (how does the system work when some components are providing data at different rates, and / or are off the grid)?
• Human-in-the-loop (behavioral aspects of data analytics)
• Moving from IoT (sensing and agent end of CPS) to CPS (real-time control through big data analytics)

We invite broad participation from the academic research community, industry groups, national labs, and government organizations.  Interested participants are asked to submit an extended abstract (1-2 pages, pdf format) addressing one or more of the discussion topics outlined above. Submitted abstracts are peer-reviewed by at least three reviewers. Accepted abstracts are invited for full paper submission (IES conference manuscript format).  The full papers will be submitted to IEEE Xplore, subject to final approval from the conference committee.  Partial participant support may be available.

Paper submissions will be handled through the CPS Virtual Organization. Please go here to upload your pdf format extended abstract: http://cps-vo.org/group/BigDataAnalyticsinCPS2015/submit-position-paper .

Program Committee:

• John Baras (UMD)
• David Corman (NSF)
• Brad Martin (NSA)
• Tho Nguyen (NSF/AAAS)
• Vinay Pai (NIH)
• Janos Sztipanovits (Vanderbilt)

Program Committee wil be updated with additional members.

Important Dates:

• Deadline for extended abstract submission is February 1, 2015 extended to February 8, 2015

• Paper acceptance notification is February 21, 2015  extended to March 10, 2015

• Full paper submission deadline is March 27, 2015  - TBD

• Deadline for submission of presentation slides: April 10, 2015

• Date of workshop is April 13, 2015

Inquiries about this workshop can be directed to Tho Nguyen at thnguyen@nsf.gov