Bojan Jakimovski

Artificial Immune System Based Robot Anomaly Detection Engine for Fault Tolerant Robots

Robot anomaly detection method described in this paper uses an approach inspired by an immune system for detecting failures within autonomous robot system. The concept is based on self-nonself discrimination and clonal selection principles found within the natural immune system. The approach applies principles of fuzzy logic for representing and processing the information within the artificial immune system. Throughout the paper we explain the working principle of RADE (Robot Anomaly Detection Engine) approach and we show its practical effectiveness through several experimental test cases.

In Situ Self-Reconfiguration of Hexapod Robot OSCAR Using Biologically Inspired Approaches

A highly desirable feature for next generation robots operating on terrestrial or extraterrestrial environment (Chien et al., 2006) is that they possess the property of sustainable autonomic systems enveloping the self-management and self-x characteristics such: self-reconfiguration, self-optimization, and self-healing. Responses taken automatically by a system without real-time human intervention are called autonomic responses (Sterritt et al., 2006; Lewandowski et al., 2006). The self-x properties will enable the robot to continue with its mission tasks even in the cases when the robot has some faults within the system. The robot shall be able to reconfigure itself and continue with its mission tasks. The autonomic concept was introduced with the IBM Manifesto for Autonomic Computing (IBM, 2001). This proposed several key elements important for the autonomic systems: selfconfiguration, self-healing, self-optimization, and self-protection which were inspired by the human body’s autonomic nervous system. Complementary to the IBM’s initiative, the Organic Computing initiative (DFG, 2004) on the other hand, proposes the means of achieving such self-x properties of the next generation of self-organizing embedded systems, inspired by information processing seen within the biological systems. Transferring such biologically inspired paradigms (Hinchey & Sterritt, 2007) into computing systems and robots will enable the systems to perform in a more robust, safe, and flexible way. In that context we have been researching towards practically applying biologically inspired methodologies and developing novel procedures for next generations of selfreconfiguring and joint leg walking robots. The rest of the paper is organized as follows: In the second chapter we give an overview of our hexapod robot prototype OSCAR on which we have conducted the experiments. We also introduce our innovative and patent pending mechanism for in-situ walking robot leg amputation and for robot reconfiguration. In the third chapter we explain the biologically inspired fault detection method used for fault/anomaly detection. In the fourth chapter we describe the swarm intelligence concept for robot reconfiguration, which is used to perform a stable spatial reconfiguration of the hexapod walking robot. We also present the results from real experiments on self19

Swarm intelligence for self-reconfiguring walking robot

A robust swarm intelligence based approach for the self-reconfiguration of a fault-tolerant multi-legged walking robot is elaborated in this paper. It is used to reconfigure the posture of the legs of the robot after some failure has occurred within the robotpsilas legs and it is based on the intrinsic properties seen within swarms and boids in nature. The reconfiguration method presented does not consider any conventional inverse kinematics modeling. Instead it is solely based on swarm intelligence concepts. Throughout the paper we describe the idea behind this approach, the principle of its operation, and we demonstrate its practical usefulness in several test-cases demonstrated on our hexapod robot - OSCAR (Organic Self Configuring and Adapting Robot).

Application of the Organic Robot Control Architecture ORCA to the Six-Legged Walking Robot OSCAR

Walking robots are complex machines, which are challenging to engineer and to program. In order to master this complexity, in this article Organic Computing (OC) principles in terms of self-organisation, self-reconfiguration and self-healing are applied to a six-legged walking robot named OSCAR (Organic Self-Configuring and Adapting Robot). The Organic Robot Control Architecture ORCA, developed in the same project, provides the architectural framework. OC principles are employed on all layers of the hierarchical robot control system starting at the reflexive layer with gait generation and reflexes over the reactive behavioural layer up to the deliberative planning layer. Many experimental evaluations with OSCAR have shown that the robot is able to flexibly adapt to internal faults as well as to unforeseen environmental situations and thus continues its mission in the best still possible way.

ORCA: An Organic Robot Control Architecture

Mastering complexity is one of the greatest challenges for future dependable information processing systems. Traditional fault tolerance techniques relying on explicit fault models seem to be not sufficient to meet this challenge. During their evolution living organisms have, however, developed very effective and efficient mechanisms like the autonomic nervous system or the immune system to make them adaptive and self-organising. Thus, they are able to cope with anomalies, faults or new unforeseen situations in a safe way. Inspired by these organic principles the control architecture ORCA (Organic Robot Control Architecture) was developed. Its aim is to transfer self-x properties from organic to robotic systems. It is described in this article with a specific focus on the way ORCA deals with dynamically changing uncertainties and anomalies.

SelSta - A Biologically Inspired Approach for Self-Stabilizing Humanoid Robot Walking

In this paper we elaborate a study on self-stabilizing humanoid robot that achieves run-time self-stabilization and energy optimized walking gait pattern parameters on different kinds of flat surfaces. The algorithmic approach named SelSta uses biologically inspired notions that introduce robustness into the self-stabilizing functionality of the humanoid robot. The approach has been practically tested on our S2-HuRo humanoid robot and the results from the tests demonstrate that it can be successfully used on humanoid robots to achieve autonomic optimized stabilization of their walking on different kinds of flat surfaces.

Ein Organic Computing Ansatz zur Steuerung einer sechsbeinigen Laufmaschine

Obwohl die Rechengeschwindigkeit von Computern und die Komplexitat unserer Systeme standig zunimmt, sind die heutigen Laufmaschinen nicht in der Lage, sich mit den Fahigkeiten von Landtieren wie zum Beispiel Insekten zu messen. Das Verstandnis biologischer Konzepte und das Lernen von der Natur konnten zur Verbesserung der heutigen Maschinen beitragen und sie ein wenig “lebensahnlicher“ machen. Dieser Artikel stellt einen Kontrollarchitekturansatz basierend auf “Organic Computing“-Prinzipien vor, der die Nutzung von Dezentralisierung und Selbstorganisation an einer sechsbeinigen Laufmaschine demonstriert. Die vorliegende Arbeit erklart die elementaren Mechanismen fur das gerade Laufen, das Kurvenlaufen sowie das Drehen auf der Stelle und den Umgang mit strukturellen korperlichen Anderungen wie einer Beinamputation und stellt die Ergebnisse experimenteller Versuche vor.