Sunday, November 30, 2014

SIMONA Research Flight Simulator

Last week we had a short visit to the SIMONA Research Simulator of Delft University of Technology with Mike Jones. He used to work with SIMONA team for ARISTOTEL project while he was in University of Liverpool. While he was visiting his old colleagues, I attended him to see the SIMONA facility and meet the SIMONA researchers. It was an exciting visit for me as a researcher working for possibly the newest research simulator facility AVES, to see SIMONA where a huge amount flight simulator research has been carried out in the last 15 years.  
Research simulators are used to various purposes but mainly to conduct flight sciences and flight systems research including new aircraft configurations, new flight systems and HMI concepts, and flight simulator research.  Following the publications, one can figure out that SIMONA has been constructed starting from late 90's to early 2000's by TU Delft researchers, by themselves. This includes all the structure to motion system, cockpit to all software components. To my understanding, this construction experience provided SIMONA researchers with a unique opportunity to identify and research into the problems of flight simulators. Thus SIMONA researchers contributed various aspect in flight simulator research including real time simulation, motion cueing and simulator validation. 
I had the chance to meet Deniz Yilmaz, Olaf Stroosma and Marilena D. Bos-Pavel. We had a very nice conversation about various aspects of SIMONA. One of the interesting topics was SIMONA real time simulation framework DUECA (Delft University Environment for Communication and Activation). In AVES, we have a similar infrastructure that we call 2Simulate. We discussed our practices how we support current model based simulation engineering practices with our real time frameworks. Another exciting discussion was about tuning and validating simulation cues (basically motion) with subjective pilot assessments. The objective criteria for most of the cueing channels in training simulators are still not enough to identify the fidelity level to guarantee transfer of training while subjective assessments, for sure, possesses a great variation depending on the evaluator. So the training simulators still ask for beyond state of the art approaches to validation.
It was a great occasion for me to see SIMONA  and meet very nice people and have invaluable discussions over our research interests. With this blog post, I would like to thank Deniz, Olaf and Marilena hosting us.    

Monday, September 8, 2014

Recent Issues and Trends in V&V of Simulation Models

Simultech is one of the leading modeling and simulation conferences in Europe. Simultech 2014 was carried out end of August in Wien. This year, the key notes and senior attendants of the conference were more than well know by simulation community. Prof.Dr.Tuncer Oren, Prof.Dr.Bernard Zeigler, Prof.Dr.Paul Fishwick, Prof.Dr.Helena Szczerbicka and Prof.Dr.François E. Cellier performed inspiring talks and discussions. One of them was the opening panel of the event, namely "Recent Issues and Trends in Validation and Verification of Simulation Models".

The panel started with the over viewing introduction of Prof.Dr.Mohammad S. Obaidat. Then Zeigler stressed the importance of formalizing "intended use" in his talk . He proposed intended use as the key fundamental for any verification and validation activity. Then he visited his experimental frames in this context and informed audience that there is a standardization effort in US for specifying intended use of simulations. Cellier started his talk with emphasizing the emerging complexity of the systems. Further he highlighted that the current system complexity almost prevents a full coverage testing. Then he concluded that advance tools and automation is now a must in validation and verification. In the second part of his talk, he underlined practical concerns. He introduced how model interface become sources of error. He visited assertions and model based specifications as examples of key validation and verification practices. Oren performed an inspiring talk that proposes a perspective change in validation and verification. He proposed "failure avoidance" as the his avant-garde interpretation of motivation for verification and validation. He explained how this perspective change brings new practices and outcomes. Fishwick, on the other hand, in his talk introduced governmental practices for model verification and validation. He presented government regulated and sponsored verification and validation process of risk models in insurance industry.
Verification and validation in modeling and simulation, although have been discussed for a long time now, still provides various challenges to researchers. Fast evolution of systems as well as the engineering tools and practices that develop these systems brought modeling and simulation to the center of any systems development or analysis effort. So as once again visited in this conference, we will be challenging ourselves with developing beyond the horizon verification and validation tools and methodologies.     

Wednesday, April 9, 2014

RACoN - Yet Another RTI Abstraction Component

RACoN
Developing an abstraction layer (wrapper) over High Level Architecture (HLA) Runtime Infrastructure (RTI) is a popular approach in many HLA-based distributed simulation development projects, since it offers more maintainable, robust, and portable methods. In this regard, RACoN (Rti Abstraction COmponent for .Net ) is a new wrapper targeted for Microsoft .NET environments, where it encapsulates the entire RTI low-level class interface in order to simplify the repetitive, error-prone, and low-level implementation details of the HLA federate interface. With the help of RACoN, any .NET language can be selected to implement an HLA federate (i.e. a simulation component).
RACoN is developed specifically for a graduate level distributed simulation course as a lab tool. Therefore, its intended use is for education, where the students easily develop HLA-based simulation projects during the course. Due to the target audience is the students, RACoN package comes with a Programmer's Guide written for beginners and with a sample application.
I don't want to give boring technical details, but RACoN architecture is specifically suitable for the layered simulation architecture style. You can check technical details in this paper. Another feature is its support for code generation. Currently, SimGe (introduced in my previous post) can generate all the RACoN compatible classes of a SOM using the federation execution configuration files (i.e. FDD/FED). A preliminary preview of the tool can be obtained from http://www.ceng.metu.edu.tr/~otopcu/simge/.
On the down side, RACoN does not support all the HLA services, but the major ones thought in the course. Currently, the main limitations of RACoN are as follows. 
  • RACoN currently supports HLA13 standard and DMSO RTI 1.3NGv6 API. This imposes the installation of DMSO 1.3 NG libraries prior to RACoN.
  •  RACoN does not support the full HLA13 federate interface services
The RACoN and its documentation are available at www.ceng.metu.edu.tr/~otopcu/racon.