Volume 57, Issue 2 - Special Issue: Robot Control Architectures / Karsten Berns, Rüdiger Dillmann, Erik Maehle

Robotic software frameworks have major impact on development effort and quality of robot control systems. This paper provides a condensed overview on the complex topic of robotic framework design. Important areas of design are discussed – together with design principles applied in state-of-the-art solutions. They are related to software quality attributes with a brief discussion on their impact. Based on this analysis, the approaches taken in the framework Finroc are briefly presented.

Robotic systems are complex, software intensive and heterogeneous composite systems. Software systems engineering and system integration is still a major challenge in robotics. We describe how component based software engineering (CBSE), model-driven software development (MDSD) and domain-specific languages (DSLs) for variability management complement each other in addressing the robotics software challenge. We outline how these approaches pave the way towards a software business ecosystem in robotics. We put a focus onto challenges still being considered as open and worth being addressed next.

With ArmarX we introduce a robot programming environment that has been developed in order to ease the realization of higher level capabilities needed by complex robotic systems such as humanoid robots. ArmarX is built upon the idea that consistent disclosure of the system state strongly facilitates the development process of distributed robot applications. We show the applicability of ArmarX by introducing a robot architecture for a humanoid system and discuss essential aspects based on an exemplary pick and place task. With several tools that are provided by the ArmarX framework, such as graphical user interfaces (GUI) or statechart editors, the programmer is enabled to efficiently build and inspect component based robotics software systems.

Embodied artificial agents operating in dynamic, real-world environments need architectures that support the special requirements that exist for them. Architectures are not always designed from scratch and the system then implemented all at once, but rather, a step-wise integration of components is often made to increase functionality. In order to increase flexibility and robustness, a task planner was integrated into an existing architecture and the planning process was coupled with the pre-existing execution and the basic monitoring processes. This involved the conversion of monolithic SMACH scenario scripts (state-machine execution scripts) into modular states that can be called dynamically based on the plan that was generated by the planning process. The procedural knowledge encoded in such state machines was used to model the planning domain for two RoboCup@Home scenarios on a Care-O-Bot 3 robot. This was done for the JSHOP2 hierarchical task network (HTN) planner. A component which iterates through a generated plan and calls the appropriate SMACH states was implemented, thus enabling the scenarios. Crucially, individual monitoring actions which enable the robot to monitor the execution of the actions were designed and included, thus providing additional robustness.

Advanced mobile robot systems need to accomplish increasingly complex task sets. However, to solve demanding problems, they are typically optimized to a very restricted set of tasks and environments. This work will therefore propose a self-reconfigurable software and hardware architecture to allow the dynamic optimization of a robot system depending on the current situation, i. e. the current task, the robot inner state, and the environment. The proposed framework is based on organic computing principles and unsupervised machine learning techniques. It further uses dynamically reconfigurable Field Programmable Gate Arrays (FPGA) as hardware accelerators. Preliminary results will be presented, which demonstrate the feasibility of the self-reconfiguration approach.

My research focuses on methods to analyze and mine large datasets as well as their practical realizations and applications. The key question of interest to me is: How can we effectively and efficiently distill useful information from large, complex, and potentially noisy datasets? To approach this question, we are developing systems for scalable data analysis and data mining, for working with incomplete and noisy data, for data-intensive optimization, as well as for extracting structured information from natural-language text. This article highlights some of my work in these areas.

I approach privacy issues on the Internet from two ends. On the one hand, I design and evaluate defensive measures, so-called privacy enhancing technologies (PETs), which can be used by individuals to protect themselves against surveillance on the Internet. On the other hand, I study the efficacy of offensive techniques. I am especially interested in passive surveillance techniques that cannot be detected. For instance, I have shown how machine learning techniques can be used to infer the contents of encrypted traffic and how to track users solely based on characteristic behavioral patterns.

In this paper, we present a series of scheduling approaches targeted for massively parallel architectures, which in combination allow a wider range of algorithms to be executed on modern graphics processors. At first, we describe a new processing model which enables the efficient execution of dynamic, irregular workloads. Then, we present the currently fastest queuing algorithm for graphics processors, the most efficient dynamic memory allocator for massively parallel architectures, and the only autonomous scheduler for graphics processing units that can dynamically support different granularities of parallelism. Finally, we show how these scheduling approaches help to advance the state-of-the-art in the rendering, visualization and procedural modeling.