Stefan Gries

A decentralized architecture and simple consensus algorithm for autonomous agents

Autonomous agents not only need to make decisions, but also might need to negotiate situations with other autonomous systems and arrange how to proceed. Architecting such systems requires developers to weigh several options and decide, whether a central authority is beneficial or if consensus can be delegated to the agents themselves. We present a decentralized architecture accompanied by a simple consensus algorithm for autonomous agents and demonstrate its application within a simulated traffic system with model cars. We furthermore compare this concept with an older, centralized incarnation, and elaborate advantages and disadvantages.

Tracking Information Flow in Cyber-Physical Systems

Cyber-Physical Systems are distributed, heterogeneous, decentralized and loosely coupled networks in which individual systems measure physical processes, exchange information, and influence processes. Sensors measure these physical processes, while aggregators process them and actuators perform resulting actions. Decisions are often based on sensor data collected by other systems. Furthermore, the aggregators also interchange information and use them to derive own decisions. Decisions must be comprehensible. However, this is only the case if all data dependencies are known. Due to the size of these networks, their loose coupling and their dynamic behavior, decisions made by a system are not always easy to understand. If an error occurs in the system, the error source must be identified. It must be known on which data a decision was based. However, since the decision can be based on information from other nodes, the search for the error source is not a trivial task. Keep in mind, that dependent nodes can have dependencies themselves as well. We present the Information Flow Monitor (IFM) that collects information about semantic data dependencies in dynamic networks. The collected dependency information is provided at a central network location. Subsequently, semantic dependencies between information can be visualized.

Property-based routing in clustered message brokers for CPS: Doctoral Symposium

Cyber-Physical Systems (CPS) are interconnected systems that can measure, manipulate, and adapt their environment via sensors and actors. The high number of measured data means that a reliable and scalable communication infrastructure is indispensable, especially if data is processed in real time. Data can be available in different measurement qualities, which usability depends on the particular application. As a result, data is regularly discarded, resulting in network inefficiencies when they are previously transmitted. This effect becomes more important as the number of heterogeneous sensors increases. In this paper, we discuss the implications and show our first approach to solve the problem based on MQTT [1], one of the most widely used public-subscribe protocols in the area of IoT and CPS.

Cascading Data Corruption: About Dependencies in Cyber-Physical Systems: Poster

CPS are interconnected systems that observe and manipulate real objects and processes. They allow dynamic extension and show emergent behavior which leads to dynamic decision-making processes that can change at runtime. They cannot always be easily understood because of the high number of components involved. If an error occurs in such a process, it is difficult to comprehend which component involved in the decision process is responsible for that error. The decision therefore has a high degree of dependency on the nodes involved in the process. Therefore, errors are not easily traceable to their original source. In this paper, we present the idea of dependency trees, which should help to identify error sources in the event of a fault.

Engineering a Cyber-Physical Intersection Management – An Experience Report

The engineering of cyber-physical systems (CPS) imposes a huge challenge for today’s software engineering processes. Not only are CPS very closely related to real objects and processes, also their internal structures are more heterogeneous than classical information systems. In this experience report, we account on a prototypical implementation for an intersection management system on the basis of physical models in the form of robotic cars. The steps to implement the working physical prototype are described. Lessons learned during the implementation are presented and observations compared against known software processes. The insights gained are consolidated into the novel Double Twin Peaks model. The latter extends the current software engineering viewpoints, specifically taking CPS considerations into account.