Special Session on Mixed-Criticality Systems (SS01)

Special Session Organizers:

Roman Obermaisser, University of Siegen, Germany, roman.obermaisser@uni-siegen.de
Raimund Kirner, University of Hertfordshire, United Kingdom, r.kirner@herts.ac.uk

Aim and scope:

Modern electronic systems, such as ground-vehicles, aircraft, or space vehicles combine a multitude of distributed applications with varying safety and real-time requirements. Even in high-end consumer cars we already find high criticality control applications as well as low criticality audio/video applications. Another example would be a space vehicle, like a satellite, that for weight reasons aims at using the same physical wiring for control data and video data. A similar weight- and, therefore, cost-reduction argument holds true for novel avionics systems. In general, these systems are called mixed-criticality systems, and it is the traditional federated approach to realize private networks for individual applications to avoid interference between criticality levels. In order to use the on-board resources more efficiently, the trend throughout industrial areas is towards integrated architectures, for example Distributed Integrated Modular Avionics (DIMA) and the Automotive Open System Architecture (AUTOSAR).

Mixed-criticality architectures with support for modular certification make the integration of application subsystems with different safety assurance levels both technically and economically feasible. In particular, time-triggered networks for the segregation of these subsystems are essential to avoid fault propagation and unintended side-effects due to integration. Also, mixed-criticality architectures must deal with the heterogeneity of subsystems that differ not only in their criticality, but also in the underlying computational models and timing models (e.g., sporadic event-based control and periodic control functions). Non safety-critical subsystems often demand event-driven adaptability and support for dynamic system structures, while certification standards impose static configurations for safety-critical subsystems.

Topics within the scope of the Special Session:

The focus of the special session is on event-based and time-based system models, architectures, platforms and applications for systems with mixed-criticality requirements. Suggested topics of interest include (but are not restricted to) the following:

• Event-triggered and time-triggered architectures for mixed-criticality systems
• Communication networks with support for temporal and spatial partitioning
• Integration of heterogeneous mixed-criticality subsystems (e.g., coexistence of event-triggered and time-triggered activities)
• Event-based reconfiguration and resource management
• Validation, verification and certification aspects (e.g., modular certification)
• Applications and case studies (e.g., in avionic and automotive systems)”

Event-Based Vision and Robotics (SS02)

Special Session Organizers:
Thibaut RAHARIJAONA, Aix-Marseille University-CNRS, ISM, Biorobotics
Thibaut.Raharijaona@univ-amu.fr
Sylvain DURAND, Aix-Marseille University-CNRS, ISM, Biorobotics
Sylvain.Durand-Chamontin@univ-amu.fr
Nicolas MARCHAND, University of Grenoble-CNRS, GIPSA-lab, Control Systems Department
Nicolas.Marchand@gipsa-lab.fr
Franck RUFFIER, Aix-Marseille University-CNRS, ISM, Biorobotics
Franck.Ruffier@univ-amu.fr

Aim and scope:

All embedded miniaturized networked systems require novel asynchronous versions of the current time-triggered sensors, actuators and algorithms. At this special session, it is proposed to present the latest experimental studies on methods of optimizing the resources and the efficiency of event-based solutions to actuation, sensing, computation and communication problems.
In the field of robotics, one of the most promising recent alternatives to the “computer vision”  approach consists in mimicking biological visual sensors, using methods based on asynchronous time-stamped events. In order to implement event-based sensors on event-based robots, it will be necessary to develop electronic event-driven computing devices and new event-based control laws and algorithms. Event-driven actuators which can be integrated into the sensorimotor chain can also be an asset. The aim of the overall design is to reduce the computational and communication load without any loss of efficiency.

Topics within the scope of the Special Session:
This special session will focus on event-based visual and robotic applications of the following kinds:

• Integrating event-based visual sensors into robotic/mechatronic applications;
• Implementing event-based strategies for innovative closed-loop control
• Optimizing the computational and communication resources using event-based approaches;
• Applications in the fields of biologically inspired sensing, vision and robotics

Industrial Event-Based Control (SS03)

Special Session Organizers:

Antonio Visioli, University of Brescia
antonio.visioli@unibs.it
Andrzej Pawłowski, UNED
a.pawlowski@dia.uned.es

Aim and scope:

Nowadays development and deployment of industrial control systems introduces into service data networks changing the classical approach of control system design. The closed-loop control systems using field buses, local area networks or even Internet make difficult the adoption of the classic paradigm, where the whole control system structure is driven by constant time events. For such approaches event-based techniques became an alternative for classical time based solutions, since they are not restricted to periodic events occurrence. The presence of the event-driven data flow in control systems requires developing new control techniques that takes advantages of such an approach and compensates drawbacks of the existing solutions.
Considering aforementioned aspects there is a continuous need to improve and adopt the commonly used controllers in industrial areas, focusing mainly on PID and MPC controllers which are widely accepted for industrial applications.In different variants and configurations, those techniques provide the control solutions for almost all industrial processes covering the whole automation pyramid. The PID controllers are the most common controllers for a low and mid-range level industrial applications today. They have an established position due to their feasibility and intelligibility. Moreover, it is likely to be the most important controller also in wireless automation solutions. There are many different architectural options for networked PID controllers to compensate for adaptive and event-based sampling, jitter, packet losses and other network related features. On the other hand, the MPC technology can now be found in a wide variety of application areas covering mid and high level control applications. Because of the MPC’s prediction mechanism, it can deal with numerous process constraints what provides an irreplaceable tool for a large scale process optimization. Furthermore, the same prediction mechanism can be used to extend classical approach into networked or event-based configurations. For all that, there exists a great interest in developing event-based counterparts of the PID and MPC that compensate for previously mentioned issues, making the implementation more flexible and robust enough to deal with many industrial control problems.

Topics within the scope of the Special Session:

The focus of the special session is on event-based and time-based system models, architectures, platforms and applications for systems with mixed-criticality requirements. Suggested topics of interest include (but are not restricted to) the following:

• Applications of event-based control systems
• Case studies involving event-based techniques
• Event-driven measurements & control
• Control strategies within event-based framework
• Industrial communication protocols for event-based process control and automation

Event-Based State Estimation (SS04)

Special Session Organizers
Sebastian Trimpe, Max Planck Institute for Intelligent Systems, Germany
strimpe@tuebingen.mpg.de
Joris Sijs, TNO, the Dutch Institute of Applied Physics, Netherlands
joris.sijs@tno.nl

Aim and scope:

The First IEEE Conference on Event-based Control, Communication, and Signal Processing (EBCCSP 2015) is to bring together researchers and practitioners from a wide spectrum of engineering disciplines including control, communication, signal processing, and electronic instrumentation. State Estimation is a fundamental and important problem that spans all of these disciplines and occurs in various application areas. Due to an increased interest in event-based methods for state estimation recently, and their potential in diverse application areas such as networked systems, autonomous vehicles, under water robotics and target tracking, we organize a Special Session on Event-based State Estimation at EBCCSP 2015. Researchers and practitioners working on this and related topics are kindly invited to submit their recent results to this session. This can include both theoretical and applied work.

Topics within the scope of the Special Session:

Suggested topics for this Special Session on Event-based State Estimation include (but are not limited to):
• Event-triggered and self-triggered architectures for state estimation
• Distributed event-based state estimation
• Triggering/sampling strategies for event-based state estimation
• (Sub-)optimal fusion of sensor data under event-based communication protocols
• Stability and performance analysis of event-based state estimation schemes
• Design guidelines for event-based state estimation systems
• Comparison and fundamental relation of time-based and event-based estimation
• Event-based state estimation for control
• Applications of event-based state estimation

Mathematical Modeling of Event-Based Sampling (SS05)

Special Session Organizer:

Bernhard Moser, Software Competence Center Hagenberg, Austria
bernhard.moser@scch.at

Aim and scope:

This Special Session is about the mathematical foundation of non-linear sampling processes that are triggered by signal based conditions. Examples for such sampling schemes are level-crossing sampling with hysteresis, send-on-delta or integrate-and-fire. In particular, challenges for event-based sampling theory posed by its applications in computational neuroscience and (biomorphic) engineering are addressed.

For example, consider level-crossing sampling with hysteresis, then small variations of the threshold can turn a sub-threshold signal with no events into a sequence with an arbitrary number of alternating +1, -1 events. Is this really instability? Or, rather, does this phenomenon tell us that the underlying metric of our notion of stability is not adequate? This example illustratively demonstrates the necessity to challenge fundamental notions like ‘stability’ and their underlying mathematical concepts in order to clarify the appropriateness of proposed concepts in this field. For this purpose, particularly, discrete mathematical, analytical, stochastic and information theoretical contributions are welcome that address the analysis of shortcomings of state-of-the-art theory of event-based sampling or tackle these shortcomings by novel approaches.

Topics within the scope of the Special Session:

Suggested topics of interest include (but are not restricted to) the following:
• Event-based sampling operators and event sequences
• Stability analysis of event-based sampling
• Metric spaces and metric equivalence
• Stochastic resonance
• Measures of information

Pixel-Level Event-Driven Vision Sensing, Processing and Infrastructures (SS06)

Special Session Organizer:

Tobi Delbrück, Institute for Neuroinformatics
ETHZ (Zurich, Switzerland), tobi@ini.phys.ethz.edu
Bernabé Linares-Barranco, Instituto de Microelectrónica de Sevilla
IMSE-CNM (CSIC and Univ. Sevilla), bernabe@imse-cnm.csic.es

Aim and scope:
Biological and Artificial Vision Systems differ fundamentally in the way visual scenes are sensed and processed: in conventional artificial vision systems a camera sensor captures sequences of still frames, which are processed afterwards frame by frame; in biological eyes and subsequent cortical processing there is no notion about the “frame” concept nor about a “frame rate”. In biology, retina cells send asynchronous spikes to cortex which represent some pixel-level (or pixel-surrounding-level) “scene event”. Event-driven vision sensors, also called Address-Event-Representation (AER) artificial retinas, have been around for about two decades now, but restricted to neuromorphic enthusiasts research laboratories. Since the advent of the first Dynamic Vision Sensor (DVS), which is a special AER sensor sensitive to pixel-level relative light changes, this enthusiasm is slowly and shyly expanding to other disciplines as well as to industry.

In this Special Session the aim is to present aspects of event-driven sensing, event-driven processing for both low level as well as higher level more cognitive tasks, but also ping on infrastructure aspects necessary for this new but growing technology.

Topics within the scope of the Special Session:
The following is a list of non-exclusive topics that can be covered within this Special Session:
• event-driven computation techniques for vision
• event-driven stereo vision techniques
• event-driven vision filtering techniques
• event-driven sensor applications
• event-driven infrastructure techniques

Event-based Technologiesin Building Automation Systems (SS07)

Special Session Organizers

Andrzej Ożadowicz, AGH University of Science and Technology, Poland, ozadow@agh.edu.pl
Jakub Grela, AGH University of Science and Technology, Poland, jgrela@agh.edu.pl

Aim and scope:

The Building Automation Systems (BASs) are well established field of distributed control implemented by the use of heterogeneous networking technologies, but there are still new challenges in it. Most of these technologies are based on event-triggered architectures because in general building automation systems belong to the class of non-real-time or soft real-time applications. Therefore, the networking technologies adopted in BASs provide usually a best-effort instead of a guaranteed strategy of service delivery. Various networking technologies in BASs are widely used for data communication between devices – network nodes. The representative examples of commercial network platforms based inherently on event-driven paradigm are LonWorks, KNX, BACnet, or CAN. Event-based architectures in BASs incorporate event-triggered task scheduling, data reporting strategy (send-on-delta), event-driven application programming and event-triggered protocol stack. An emerging trend in BASs is the adoption of the Internet of Things and its fusion with existing networking technologies within a vision of Smart Cities and Smart Grids

Topics within the scope of the Special Session:

The special session is focused on event-triggered and time-triggered architectures applied to the building automation and control. Topics of interest are included but not limited to:
•    Event-triggered and time-triggered architectures in BASs
•    Integration of event-triggered and time-triggered architectures in BASs
•    Event-driven application programming
•    Event-triggered protocol stacks (e.g., various versions of CSMA)
•    Wireless communication technologies in BASs
•    Energy-efficient control over event-triggered control systems
•    Event-triggered technologies in component-based automation systems
•    Applications and case studies in building and home automation
•    The best practices in BASs implementations
•    Energy management systems (EMS) based on event-triggered architectures
•    Event-triggered networks in energy efficient buildings
•    BASs in Smart Grid and Smart City
•    Emerging trends in BAS based on event-triggered architectures
•    Internet of Things (IoT) in BASs
•    Building Internet of Things (BIoT) – web based management of event-triggered control systems
•    Distributed control networks for residential buildings

Event-based Energy Metering and Control (SS08)

Special Session Organizers

Mikhail Simonov, Istituto Superiore Mario Boella (ISMB), Italy, simonov@ismb.it
Gianluca Zanetto, Teamware Srl, Italy, gianluca.zanetto@teamware.it
Gianfranco Chicco, Politecnico di Torino, Italy, gianfranco.chicco@polito.it

Aim and scope:

The special session of the EBCCSP 2015 is to bring together researchers and practitioners from a wide spectrum of engineering disciplines including control, communication, signal processing, and electronic instrumentation dealing with the event-driven sensing in general, and event-driven procedures of measurements of energy/power and its quality, and the control of smart grid. In the recent past, smart metering panorama was dominated by timer-driven meters acquiring metering data at regular intervals of time. Averaged pseudo-measurements are useful to support billing procedures, but an important step beyond the state of the art is given by the event-driven approach to metering.

Next generation event-driven meters are hybrid devices that enable both the billing and the decentralized control of smart grid segments. Due to an increased interest in event-based methods for the control of distributed generation from poorly predictable renewable sources of energy, we organize a Special Session on Event-based energy metering and control at EBCCSP-2015. Researchers, domain experts, and practitioners working on such topics are kindly invited to attend this session by submitting their recent work.

Topics within the scope of the Special Session:

We will consider any kind of contribution related to event-driven energy measurements and control without restrictions on topics. We invite theoretical and applied work. Examples of topics are: event-triggered procedures of measurement, service-oriented and event-driven architectures, use of the Future Internet (Internet of Things, Services, and Energy), stateless and statefull triggers, publish-subscribe, decentralized multi-agent systems, self-triggered architectures and applications, proactive control, knowledge elicitation from metering data, fusion of sensor data under event-based communication, event-based communication protocols, and so on. We welcome use cases, examples of real life applications exploring the use of event-driven measurements/metering, and studies making comparison between time-based and event-based operations.