文档下载 文档下载导航
设为首页 | 加入收藏
搜索 请输入内容:  
 导航当前位置: 文档下载 > 所有分类 > A Computational Framework for the Simulation, Verification, and Synthesis of Force-Guided R
侵权投诉

A Computational Framework for the Simulation, Verification, and Synthesis of Force-Guided R

Abstract. Robotic assemblies are inherently hybrid systems. This paper pursues a class of multi-tiered peg-inhole assemblies that we call “peg-in-maze ” assemblies. These assemblies require a force-responsive, low-level controller governing physical cont

A Computational Framework for the Simulation, Verification, and Synthesis

of Force-Guided Robotic Assembly Strategies

Michael S. Branicky1 and Siddharth R. Chhatpar2

Case Western Reserve University

10900 Euclid Ave, Cleveland, OH 44106-7221, U.S.A

{src2, msb11}@po.cwru.edu

Abstract. Robotic assemblies are inherently hybridsystems. This paper pursues a class of multi-tiered peg-in-hole assemblies that we call “peg-in-maze” assemblies.These assemblies require a force-responsive, low-levelcontroller governing physical contacts plus a decision-making, strategic-level supervisor monitoring the overallprogress. To capture this dichotomy we formulate hybridautomata, where each state represents a different force-controlled “behavior” and transitions between statesencode the high-level strategy of the assembly. Our over-arching goal is to produce a computational framework forthe simulation, verification, and synthesis of such force-guided robotic assembly strategies. We investigate the useof three general hybrid-systems software tools (Hybrid cc,HyTech, and CEtool) for the simulation and verification ofthese strategies. We describe the computationalenvironment we developed at Case to synthesize andimplement real-world assembly strategies.

1. Introduction

Robotic assemblies are inherently hybrid systems [1]because they involve the making and breaking of contacts[2,3]. This paper pursues a class of such assemblies thatwe call “peg-in-maze” assemblies, which includes real-world examples such as clutch mating and gear meshing.These are multi-tiered versions of traditional “peg-in-hole”assembly problems. There are further complications dueto the fact that the layers may move with respect to eachother (allowing position uncertainties of the componentsthat exceed the corresponding assembly clearances). Thus,successful assembly requires altering component locationsthrough physical contact between the mating parts in asearch for the next insertion hole. Once the hole is found,insertion may proceed. This already induces a layer-by-layer switching between search and insertion “behaviors.”Further complexity may arise if the assembly reaches adead-end or “jam” state, requiring the invocation of someother behavior (backing up, re-planning, or aborting).

We have successfully automated example “peg-in-maze” assemblies using a two-level control in previouswork [4,5,6]. To govern physical contacts, we use aparameterized family of low-level force-responsivealgorithms known as natural admittance controllers [7].

1

To switch between different low-level controllers, invokenew behaviors, and monitor the overall progress of theassembly, we use a higher-level discrete-event [8]supervisor. Together, these two levels of controlimplement an assembly strategy that guides the assemblyprocess from start to finish.

Both the strategy and the assembly process itself arehybrid systems. Any “run” of the system consists ofcontinuous movements (such as searching for a hole ormoving down until contact) that are governed bycontinuous control laws and discrete events (such asfinding a hole or contacting a layer) that are processed by alogical decision maker. Thus, to formally model peg-in-maze assemblies and their associated assembly strategies,we must use a hybrid systems approach [9]. In this paper,we formulate hybrid automata [10] models ofassemblies/strategies, where each state represents adifferent force-controlled behavior and transitions betweenstates encode the high-level strategy of the assembly.

Our over-arching goal is to produce a computationalframework for the simulation, verification, and synthesisof such force-guided robotic assembly strategies. Weinvestigate the use of three general hybrid systemssoftware tools for the simulation and verification of thesestrategies. Specifically, we use Hybrid cc for simulation,HyTech for verification, and CEtool for both simulationand verification. Herein, we briefly describe the resultsobtained with these tools and our own C++ analysisalgorithms specifically tailored to peg-in-maze problems.

Finally, we describe a computational environmentdeveloped at Case to help automate the synthesis andimplementation of force-guided robotic assemblystrategies. The tool includes a state machine editor forentering overall strategy and a mode editor for creating theunderlying force-controlled behaviors. There is also acode generator and a real-time monitor.

The paper is organized as follows. Section 2 describespeg-in-maze assemblies further and formulates the hybridautomaton representation of the strategies to automatethem. Section 3 presents a simplified two-dimensionalpeg-in-maze assembly, which is used to fix ideasthroughout. Section 4 describes the simulation andverification of assembly strategies using the general hybrid

Corresponding Author. Asst. Prof., Electrical Engineering and Computer Science Dept., Glennan 515B.2

Ph.D. Student, Mechanical and Aerospace Engineering Dept., Glennan 109.

第1页

你可能喜欢

热门文档

相关文档

站点地图 | 文档上传 | 侵权投诉 | 手机版
本站所有资源均来自互联网,本站只负责收集和整理,均不承担任何法律责任,如有侵权等其它行为请联系我们.
文档下载 Copyright 2013 doc.guandang.net All Rights Reserved.  -  email