University of Texas at Austin
Data Analytics for Converting Neuromuscuoskeletal Biometric Data into Valuable Information
This talk will give an overview of a novel, system-based methodology for monitoring fatiguing and resting in a human neuromusculoskeletal (NMS) system. It is based on dynamic models that link surface electromyographic signals and joint kinematic or force variables in order to track changes in system dynamics, as well as to assess joint level and muscle level contributions to those changes. The methodology is demonstrated on several data sets. One data set consists of EMG and limb movement signals recorded from 12 human subjects completing a repetitive sawing motion until voluntary exhaustion. Statistically significant changes were noted with for all subjects, with fittest subjects displaying the least amount of degradation. Results seem to clearly indicate that the model based approach successfully tracks subjects’ fatiguing. Another dataset consists of one subject maintaining leg force until voluntary exhaustion, after which this subject’s performance was tracked during the resting period, illustrating tremendously consistent recovery after the subject rested. Finally, relevant muscle EMGs and jaw velocity were concurrently collected from one subject as the subject kept its mouth fully open until voluntary exhaustion. The same data were collected again after the subject rested, once again showing remarkable resemblance of the recovered model to the model initially built on the initial, fresh data set. The talk will end with a suggested vision how this analytics paradigm could evolve and become an integral part of future business infrastructures enabled by pervasive sensing, computing and communications in the Internet of Things environment.
Dragan Djurdjanovic obtained his B.S. in Mechanical Eng. and in Applied Mathematics in 1997 from the Univ. of Nis, Serbia, his M.S. in Mechanical Eng. from the Nanyang Technological Univ., Singapore in 1999, and his M.S. in Electrical Eng. (Systems) and Ph.D. in Mechanical Eng. in 2002 from the Univ. of Michigan, Ann Arbor. His research interests include advanced quality control in multistage manufacturing systems, intelligent proactive maintenance techniques and applications of advanced signal processing in biomedical engineering. He co-authored 54 published or accepted journal publications, 3 book chapters and more than 50 conference publications. He is a Fellow of the International Society for Asset Management and the Director of the University of Texas Industry-University Cooperative Research Center on Intelligent Maintenance Systems. Dr. Djurdjanovic is a Fellow of the International Society for Engineering Asset Management (ISEAM) and was a Research Affiliate of the International Academy for Production Research (CIRP) from 2008-2014. He is the recipient of several prizes and awards, including the 2006 Outstanding Young Manufacturing Engineer Award from the Society of Manufacturing Engineers (SME).
Data Scientist, Office of the Chief Information Officer
Jet Propulsion Laboratory (JPL), NASA
Running the Gamut of IoT: From Spacecraft Telemetry to Facilities Maintenance
JPL’s robotic spacecraft contain some of the most complex and robust sensor networks in existence, and the ability to reason about relationships in these networks is essential in ensuring smooth operations for missions costing hundreds of millions to billions of dollars. As a result, JPL is developing robust cloud-based data processing and machine learning pipelines to provide insight into anomalies and patterns exhibited by various spacecraft. This level of IoT maturity strongly contrasts with JPL’s use of IoT to improve the way we work and maintain our facilities, where the value of information is less obvious and practical concerns like network security have a stronger influence on investment decisions.
Kyle Hundman is a data scientist within the Office of the Chief Information Officer at NASA’s Jet Propulsion Laboratory (JPL) and prior to joining JPL he graduated from Northwestern University’s Masters of Science in Analytics (MSiA) program in 2014. In addition to developing more nuanced spacecraft anomaly detection systems, he is working on adapting and extending natural language processing and machine learning research from the DARPA Memex program for use by the Earth Science community.
Distinguished Research Statistician, Advanced Analytics Division
A Novel SVDD Based Control Chart for Monitoring Multivariate Sensor Data
Support Vector Data Description (SVDD) is a machine learning technique used for single class classification and outlier detection. We propose a new SVDD based control chart which addresses the challenges faced when using previously proposed control charts on big-data generated at a high frequency and thus, is applicable to many Internet of Things (IoT) applications. Our proposed control chart does not require normality as an assumption and it can simultaneously monitor both, the process variation and process mean shifts. The method utilizes a new sampling based approach of training SVDD models making large sample size a non-issue and also does not require scoring observation by observation. Because SVDD is a single class classifier our approach should have many applications including monitoring the health of complex equipment and predicting impending failures, predicting fraudulent behavior, and discovering abnormal health states.
Anya McGuirk is a Distinguished Research Statistician in the Advanced Analytics Division at SAS where she currently focuses on analytics related to the Internet of Things (IoT). She joined SAS in September, 2004. Prior to joining SAS, Anya was a Full Professor at Virginia Tech with a joint appointment in the Departments of Agricultural and Applied Economics, and Statistics. At Virginia Tech, McGuirk taught econometrics and micro-economic theory at the graduate and undergraduate levels.
Chief IoT Technologist, Managing Director
IoT: Shining a Light on the Darkside of the Factory Line
Deloitte Digital and a Heavy Equipment Manufacturer are implementing IoT solutions that leapfrog past the dark days of clipboards and paper lists and thinking you know, into the bright shiny future of sensors, data collection and actually knowing. IoT “lights the dark factory” by connecting currently disconnected, ‘stepchild like’ processes to transform the last parts of the factory into a “digital twin” that enhance analytics, guidance, and automation. The resulting “Digital Factory” is capable of recording and tracking movement, location, station time, and production quality. And, turning to the “light side,” manufacturing operations can realize substantial returns on their IoT investment.
Robert Schmid is a pragmatically audacious executive with creative thinking, grounded in fundamentals, pushing honed tech innovation for business impact. He has a deep passion and experience in all things digital – IoT, wearables, social, mobile, and cloud computing – in order to deliver business value.
Robert is currently the Chief IoT Technologist, Managing Director for Deloitte Digital. Some accomplishments include: creating and leading Deloitte IoT practice, grew it to over 100 people and an extended network of over 1,000 members; Built strategy, vision and plan for new business unit; Pioneered social media and video approach within Deloitte, and Evangelized IoT vision, strategy and approach in keynotes and online videos.
CTO & Director, Systems Architecture
IOTG Industrial & Energy Solutions Division
Enabling Autonomous Cyber-Physical Systems in Manufacturing
The Internet of Things (IoT) are rapidly gaining adoption throughout the information and operational technology industries. IoT has been projected to extend computing to over 50 billion devices, 200 billion sensors and greater than 40 zettabytes of potentially storable data. Industry 4.0 promises to bring IoT and cloud computing technologies to Cyber-Physical Systems (CPS) to enable a new era in manufacturing automation and control systems. In this talk, we examine the challenges faced by the current industrial automation and control industry and discuss the disruptive opportunities that are envisioned in this transition. Key to the success of these trends are the advancements in computing technologies, data management systems and the critical role of artificial intelligence that may enable CPS and accelerate autonomous production and manufacturing systems.
John has over 32 years of computing industry experience spanning information technology research, architecture and engineering. John has been with Intel Corporation for 23 years and is currently the CTO of the Industrial and Energy Solutions Division in the Internet of Thing Group (IOTG). John has published numerous technical papers and patents in the area of internetworking and distributed systems. Prior to Intel, John worked at DEC, MA for eight (8) years and his career began at IBM, NY. John received his EE PhD from Columbia University, New York in 2011, an MSEE from the University of Southern California, Los Angeles, CA in 1991 and a BSEE from Northeastern University, Boston, MA in 1986.