|9:00 – 10:00||Invited Keynote by Prakash Sarathy|
|10:00 – 10:30||Networking Break|
|10:30 – 11:00||Invited Talk by Kirstie Bellmann |
|11:00 - 12:00||
Presentations from Participants (Elevator Pitch Style)
|12:00-13:30||Lunch Break & Poster Viewing|
|13:30 -14:00||Invited Talk by Richard Voyles|
|14:00-15:10||Break-out Group Work|
|15:10 – 15:40||Networking Break & Poster Viewing|
|15:40 – 16:10||Invited Talk by Javier Cámara M.|
|16:10 – 17:20||Panel Discussion|
Unmanned vehicles when viewed as cyber-physical systems can span from simple to very tightly coupled, vertically integrated and highly complex. Much of this complexity is driven by the twin threads of a demanding mission environment and a need for assuring safe behavior of such systems. Cross-layer resiliency and self-awareness can provide potent tools that could significantly reduce the component coupling and need for tight vertical integration, while providing higher levels of safety assurance organic to the system design paradigm. This talk will focus on key requirements within unmanned aerospace domain that could be primary drivers for higher levels of resiliency and self-awareness for targeted platforms.
Dr. Prakash Sarathy is the mission management IPT lead for Northrop Grumman N-UCAS programs. He earned his doctoral degree in Aerospace Engineering from Syracuse University. He has held positions as post doctoral fellow, research engineer and tenured engineering faculty in his prior career. He has over 30 years of experience in the area of software applied to aerospace engineering, providing technical and project/program management and oversight for advanced technology programs requiring accelerated risk burn down and rapid maturation. His technical expertise areas include cooperative distributed architectures for multi-agent systems, cooperative decision frameworks, agent architectures for mission management, aggregate control of distributed assets, user interface design for aggregate level situational awareness. Insertion of neural networks, evolutionary computing and emergent behavior to decision making paradigms. This software engineering expertise coupled with his in-depth experience in linear and nonlinear dynamics of vehicle systems, applied to guidance, navigation and control of aircraft, spacecraft and robots as well as of real-time and embedded simulations, high fidelity modeling, implementation, VV&A and testing, provide an excellent framework for the challenges of next generation autonomous aircraft such as the N-UCAS systems.
There are now a number of good examples showing the potential benefits of having self-awareness in both traditional large complex systems, as well as in autonomous systems. What has become increasingly obvious to the self-awareness community is that reflection and reasoning about data, an external situation, and goals within a traditional computational system is changed when that reflection and reasoning takes place within a situated, embodied system. A cyber-physical system (CPS) is such a system. From classical neuro-ethology, we know something about how different the goals, constraints, and decisions are in different animal systems because they have different bodies and different capabilities both cognitively and physically (sensors and effectors). We also know from neuro-biology that for any given animal, there are many diverse self-awareness capabilities with a wide range of capabilities and requirements. Drawing on lessons learned from classical neuro-ethology, from work in adaptive CPSs and from recent work on what one needs in a system that not only uses self-models, but builds them, we discuss several challenges in developing appropriate self-awareness approaches and methods for CPSs. We will especially focus on these challenges: 1) pushing self-awareness down into sensors, effectors, and devices, 2) dealing with the changing boundaries (and hence dynamic self-models and adaptive behavior) in semi-open distributed CPSs (e.g., Internet of things, Service Oriented Architectures, System of Systems); new approaches to conflict resolution and integration among loosely coupled CPSs. We close with a discussion on what is known, knowable, and unknown in the self-models of CPSs, and this knowledge may change over time.
Robotic systems are complex, as well as difficult to build and maintain. One of the main reasons is that the intent behind the design of such systems is often times embedded in the code, hampering traceability of run-time decisions, and making it difficult to answer questions such as "Why did the robot do X?", or "What would the robot do, given a set of assumptions Y?". To improve on the current situation and enable formal reasoning about a rich space of run-time decisions, there is a need of making design intent explicit, and link it to the implementation (and ultimately, to the run-time behavior) of the system in a principled way. This talk overviews some work that enables run-time adaptation in mobile robotic systems based on explicit models of intent. The approach is able to integrate information from models that capture different facets of the problem domain (e.g., physical space, energy consumption, software configuration) into formal specifications that can be used to synthesize run-time decisions that can be traced back to explicit models of intent.
|Rafael Macieira and Edna Barros||A Contract-based Mechanism for Monitoring The Communication Between Device Drivers and Devices|
|Robert A. Nawrocki and Richard M. Voyles||Structured Computational Polymers: Self-Aware Materials for Co-Design of Cyber-Physical Systems|
|Lukas Esterle, Peter R. Lewis, Xin Yao and Richie McBride||Self-aware Camera Networks: The Challenge of Going Mobile|
|Lukas Esterle||Computational Self-awareness: Challenges in Cyber-physical Systems|
|Kalle Tammemäe||Hierarchical attention network to manage processing resources of CPSs|
|E.Engeler||An Operator Theory of Structured Distributed Control|
|Dávid Juhász, Axel Jantsch, Maximilian Götzinger, and Nima TaheriNejad||Modeling Self-Awareness|
|Maximilian Götzinger, Nima TaheriNejad, Amir M. Rahmani,
Pasi Liljeberg, Axel Jantsch, and Hannu Tenhunen
|Self-Awareness in Remote Health Monitoring Systems using Wearable Electronics|
|Yanzhe Cui, Richard M. Voyles, and Shreyas Sundaram||A Self-Awareness Infrastructure for Resource Constrained Multi-Robot Systems in the Presence of Faults and Uncertainty|
© 2016 & 2017 TU Vienna - Institute of Computer Technology, Inc.