LoboVault Home

Symbolic planning for heterogeneous robots through composition of their motion description languages

LoboVault

Please use this identifier to cite or link to this item: http://hdl.handle.net/1928/9349

Symbolic planning for heterogeneous robots through composition of their motion description languages

Show full item record

Title: Symbolic planning for heterogeneous robots through composition of their motion description languages
Author: Zhang, Wenqi
Advisor(s): Tanner, Herbert
Committee Member(s): Abdallah, Chaouki T.
Fierro, Rafael
Lumia, Ron
Russell, John
Department: University of New Mexico. Dept. of Mechanical Engineering.
Subject: motion planning
Heterogeneous Robots
LC Subject(s): Incremental motion control.
Autonomous robots--Control systems.
Robots--Programming.
Degree Level: Doctoral
Abstract: This dissertation introduces a new formalism to define compositions of interacting heterogeneous systems, described by extended motion description languages (MDLes). The properties of the composition system are analyzed and an automatic process to generate sequential atom plan is introduced. The novelty of the formalism is in producing a composed system with a behavior that could be a superset of the union of the behaviors of its generators. As robotic systems perform increasingly complex tasks, people resort increasingly to switching or hybrid control algorithms. A need arises for a formalism to compose different robotic behaviors and meet a final target. The significant work produced to date on various aspects of robotics arguably has not yet effectively captured the interaction between systems. Another problem in motion control is automating the process of planning and it has been recognized that there is a gap between high level planning algorithms and low level motion control implementation. This dissertation is an attempt to address these problems. A new composition system is given and the properties are checked. We allow systems to have additional cooperative transitions and become active only when the systems are composed with other systems appropriately. We distinguish between events associated with transitions a push-down automaton representing an MDLe can take autonomously, and events that cannot initiate transitions. Among the latter, there can be events that when synchronized with some of another push-down automaton, become active and do initiate transitions. We identify MDLes as recursive systems in some basic process algebra (BPA) written in Greibach Normal Form. By identifying MDLes as a subclass of BPAs, we are able to borrow the syntax and semantics of the BPAs merge operator (instead of defining a new MDLe operator), and thus establish closeness and decidability properties for MDLe compositions. We introduce an instance of the sliding block puzzle as a multi-robot hybrid system. We automate the process of planning and dictate how the behaviors are sequentially synthesized into plans that drive the system into a desired state. The decidability result gives us hope to abstract the system to the point that some of the available model checkers can be used to construct motion plans. The new notion of system composition allows us to capture the interaction between systems and we realize that the whole system can do more than the sum of its parts. The framework can be used on groups of heterogeneous robotic systems to communicate and allocate tasks among themselves, and sort through possible solutions to find a plan of action without human intervention or guidance.
Graduation Date: May 2009
URI: http://hdl.handle.net/1928/9349


Files in this item

Files Size Format View
WenqiZhangDissertation.pdf 650.9Kb PDF View/Open

This item appears in the following Collection(s)

Show full item record

UNM Libraries

Search LoboVault


Browse

My Account