DDDAS.org is a scientific community web site dedicated since 2001 to promoting Dynamic Data-Driven Application Systems (DDDAS) techniques and research. It accepts suggestions and contributions for the web site from all members of the DDDAS community. The site has hyperlinks to workshop reports, virtual proceedings, DDDAS projects, a bibliography (in BibTeX), and announcements or news.
A dynamic data-driven application system is the integration of a simulation with dynamically assimilated data, multiscale modeling, computation, and a two way interaction between the model execution and the data acquisition methods.
Observations can come from many sources, but sensors can provide an almost infinite amount of data. Using software agents to mine very large, growing datasets or data streams can also provide observations. Both are potentially Big Data [Wikipedia] sources that frequently are involved in data-driven applications using Big Computing environments. Researchers in the DDDAS field have been discussing recently how to naturally incorporate Big Data techniques and communicate DDDAS concepts to Big Data theoreticians and practitioners.
As an intellectual challenge, DDDAS research is about how to make the best use of the union of data acquisition, computation, and simulation methods in building systems. Creating a DDDAS requires understanding sensors, computing at levels ranging from an embedded processor to a supercomputer, networks, data security, mathematical and computational modeling, inverse problems, control theory, and visualization.
DDDAS creates a rich set of new challenges for applications, algorithms, systems software, and measurement methods. DDDAS research typically requires strong, systematic collaborations between applications domain researchers and mathematics, statistics, and computer sciences researchers, as well as researchers involved in the design and implementation of measurement methods and instruments. Consequently, most DDDAS projects involve multidisciplinary teams.
Suggestions to online customers, financial management, manufacturing process controls, resource management, weather and climate prediction, disaster management, traffic management, systems engineering, civil engineering, geological exploration, social and behavioral modeling, cognitive measurement, medical devices and surgical procedures, and bio-sensing are all examples where DDDAS has been used successfully.
Consider an autopilot for a commercial airplane. The airplane is told by a controller to go to one or more locations in a sequence. Using a variety of data, the vehicle goes there unless new locations are provided en route. Data acquisition includes airplane velocity, wind velocity, a three dimensional location with respect to the Earth, fuel loads, weather conditions en route, ice conditions on the airplane surface, fire conditions, engine data, and mechanical and sensor failure data. The autopilot maintains the flight route at altitudes that both dynamically change en route based on factors such as turbulence, flight lane rules, destination airport congestion, and changing weather.
Depending on flight conditions, nonlinear partial differential equations are solved using different scales to maintain the proper speed to get to the final destination at a specified time and with adequate remaining fuel. Data is regularly and securely transmitted to airline and airplane manufacturing computers and feedback data and subsequent queries are returned to the airplane.
Autopilots in modern airplanes have extensive capabilities well beyond anything possible in early ones, which only used data assimilation. Early airplanes used hydraulic systems to move flaps and rudders and the pilots flew the plane with the help of the autopilot. Modern airplanes use digital commands to move parts (known as fly by wire) so that pilots do not actually fly the airplane anymore. Instead, pilots give an airplane computer system commands that are executed only if they will work safely. The control systems of all modern airplanes are a very complex DDDAS in action.
Understanding a DDDAS and how it interacts with people can be challenging. In the case of Air France 447 on June 1, 2009, the autopilot disconnected during a tropical storm when the airplane's speed sensors froze, producing inconsistent data. The pilots unexpectedly had to fly the airplane themselves. By misinterpreting the instrument data they flew the plane incorrectly, causing it to stall and crash into the Atlantic Ocean [BEA].
There are other examples of DDDAS from before the term was created, including Mars rovers (developed since the 1990's or earlier) which avoid hazards that are not visible on Earth from preliminary video feeds prior to rover movement and oil and gas pipeline control and monitoring systems (that have evolved since the 1970's).
Data assimilation is the scientific term used to describe “the process of incorporating observations into a computer model of a real system.” [Wikipedia] The weather forecasting community developed this paradigm first in the 1960's. Data assimilation has been an active research field since then and there remain many open questions independent of DDDAS (which places additional requirements on data assimilation). While data assimilation is concerned with better forecasts by integrating outcomes from “black box” models and new data, DDDAS redesigns the whole system by opening up the black box and allowing modification of the entire model and data collection processes. There are very interesting and different challenges in applied mathematics and computing for DDDAS over data assimilation as a result.
CyberPhysical Systems (CPS) resembles the DDDAS paradigm, but is different. Researchers in the CPS field define it by, “CPS is about the intersection, not the union, of the physical and the cyber.” [http://cyberphysicalsystems.org] CPS is primarily an engineering field and they trace their roots back to Norbert Weiner's work on automatic aiming and firing of antiaircraft guns in World War II. DDDAS differs since it is interested in the union, not the intersection, of the above. An example CPS is an embedded system in a shoe that transmits the wearer's activity data to an analysis application running on a nearby mobile phone. Probably the most visible CPS is Apple's Health app and HealthKit library that was introduced with iOS 8 for iPhones and iPads.
The Internet of Things (IoT) is the network of physical objects accessed through the Internet [WhatIs]. Each object has embedded processing so it can represent itself digitally. Hence, it can be controlled from anywhere. IoT means more data is gathered from more places with more ways to improve efficiency, safety, and security. Connecting all of these objects will lead to new automation capabilities in many fields that are not obviously related. The IoT represents subsets of either DDDAS or CPS. An IoT example is a thermostat for a heating or cooling system that is connected to the Internet and can be remotely controlled.
Data Intensive Scientific Discovery (DISD) is sometimes promoted as the fourth paradigm in science [Microsoft Research]. It represents how Big Data methods can be used to make scientific discoveries based on hidden data that has already been collected in databases. The dynamic aspect is through new data added to databases over time and data mining making discoveries. DISD represents only a small subset of either DDDAS or CPS. An example of DISD is searching through all available medical databases, which grows over time, and during a search finding the origin of a disease.
The DDDAS.org web site is self supported without aid from government grants. The host was located from 2001-2008 at the University of Kentucky and since 2008 at the University of Wyoming. Over the years many people from many institutions have contributed information and time to this web site. Their contributions are gratefully acknowledged.
The DDDAS field itself would not exist in its present state without the long term, generous support provided by the National Science Foundation, the Air Force Office of Scientific Research, and funding sources in Europe, Brasil, the Kingdom of Saudi Arabia, and possibly other countries. The DDDAS research community gratefully acknowledges these funding sources and their staffs. Specific acknowledgments can be found on some web pages in this site.
This website has been established and is maintained by Prof.
Craig C. Douglas.
©2014-2015 by www.dddas.org and Craig C. Douglas