TY - JOUR

T1 - Dynamic distribution of robot control components under hard realtime constraints

T2 - Modeling, experimental results and practical considerations

AU - Dietrich, Franz

AU - Maaß, Jochen

AU - Hagner, Matthias

AU - Steiner, Jens

AU - Goltz, Ursula

AU - Raatz, Annika

N1 - Funding information: This work has been supported by the German Research Foundation (DFG), Collaborative Research Center 562 (SFB562). The authors gratefully acknowledge the work of Frank Sowinski and Ana Amado, who contributed to the implementation and experiments. The present article has evolved in the context of the Collaborative Research Center 562 “Robot Systems for Handling and Assembly” ( sfb 562 ), a major research project funded by the German Research Foundation (DFG) over a period of 12 years involving nine institutes at TU Braunschweig and German Aerospace Center (DLR, Braunschweig). This scheme comprises numerous sub-projects that investigate fundamentals related to methods and components for robot systems based on closed kinematic chains [5] . Fig. 1 shows two of the robots developed in the SFB 562, Hexa II and Triglide . 4 4 These so-called Parallel Kinematic Machines (PKMs) are designed to have their actuators located near their non-moving base frame, a paradigm that minimizes inertia of the masses moved. Low inertia results in high velocity and acceleration, which yield short cycle time of production tasks to be accomplished. PKMs can be designed to feature properties superior to serial kinematics, but this specialization inherently suffers from loss of flexibility compared to more generalized robots. For this reason, a large variety of kinematic structures has emerged to suit particular requirements, where each of them is designed and optimized for a special purpose. In order to cope with all the individualities, genericity is the key to efficient software development of robot control software. Triggered by this line of argument, the generic software architecture prosa-x 5 5 has been proposed by the SFB 562. It has been designed to carry the control architecture rca 562 , 6 6 and has been implemented on top of qnx and the middleware mirpa-x to control the Hexa II and the Triglide . What is to be noted is that, although these robots are very different, pattern oriented design made it possible to build robot control applications for both robots on the same code base. This statement will be explained and detailed in the remainder of the article.

PY - 2013/1/26

Y1 - 2013/1/26

N2 - It can be seen in numerous applications that embedded systems take advantage of distributed execution of tasks. Such distribution is studied in the present article, which investigates the deployment of robot control architectures across multiple computers. Besides the patterns for deployment across multiple hosts, this article proposes to introduce aspects of self-management into robot control architectures. It is proposed to use graph partitioning algorithms to determine the distribution pattern (mapping of control tasks to CPU resources while minimizing bus communication load). The underlying model and the respective analysis guarantee that, after adaption of the distribution pattern, real-time properties are preserved and load is balanced. In this way, poor a priori assumptions about worst-case execution times are detected and corrected continuously during runtime. This is a considerable improvement in comparison to using only offline analysis of worst-case execution times.

AB - It can be seen in numerous applications that embedded systems take advantage of distributed execution of tasks. Such distribution is studied in the present article, which investigates the deployment of robot control architectures across multiple computers. Besides the patterns for deployment across multiple hosts, this article proposes to introduce aspects of self-management into robot control architectures. It is proposed to use graph partitioning algorithms to determine the distribution pattern (mapping of control tasks to CPU resources while minimizing bus communication load). The underlying model and the respective analysis guarantee that, after adaption of the distribution pattern, real-time properties are preserved and load is balanced. In this way, poor a priori assumptions about worst-case execution times are detected and corrected continuously during runtime. This is a considerable improvement in comparison to using only offline analysis of worst-case execution times.

KW - Distributed computing

KW - Robot control architecture

KW - Self-management

KW - Worst case execution time analysis

UR - http://www.scopus.com/inward/record.url?scp=85032131446&partnerID=8YFLogxK

U2 - 10.1016/j.sysarc.2012.12.001

DO - 10.1016/j.sysarc.2012.12.001

M3 - Article

AN - SCOPUS:85032131446

VL - 59

SP - 1047

EP - 1066

JO - Journal of Systems Architecture

JF - Journal of Systems Architecture

SN - 1383-7621

IS - 10 PART C

ER -