shage: a framework for self-managed robot software · 1 may 21~22, 2006 seams, shanghai, china...

1

May 21~22, 2006 SEAMS, Shanghai, China

SHAGE: A Framework for SHAGE: A Framework for SelfSelf--managed Robot managed Robot

SoftwareSoftware

Sooyong Park<[email protected]>

Software Systems Engineering Lab.Sogang University

May. 21~22, 2006

ICSE 2006 Workshop on Software Engineering for Adaptive and

Self-Managing Systems(SEAMS), Shanghai, China

2May 21~22, 2006 SEAMS, Shanghai, China

OutlineOutline

• Introduction

• An Approach

• SHAGE Framework

• Demonstration

• Research Issues

• Conclusions

2


Home Service Robots for the ElderlyHome Service Robots for the Elderly

• Home service robots(HSR) provide useful services such as – Face recognition

– Navigation

– Bringing an object

– Conversation

– …

• CIR(the Center for Intelligent Robotics) at KIST(theKorea Institute of Science and Technology) has been developed various home service robots.– T-Rot(High-end), Infotainment robot, H-Robot(Healthcare robot)


Overall S/W ArchitectureOverall S/W Architecture1) 1) of HSRof HSR

Deliberative Layer

Reactive Layer

Task Manager

Action ComponentsAction Components Action Components

Action Components

Behavior ControlComponent



…

Functional UnitComponent






Sequencing Layer

SHAGE Framework

1) E. Gat, “On three-layer architectures,” in Artificial Intelligence and Mobile Robots (D. Kortenkamp, R. P.Bonnasso, and R. Murphy, eds.), MIT/AAAI, 1997.

3


Research MotivationResearch Motivation

Malfunction

newEnvironment

Self-Healing, Adaptive, and Growing Technologies for Intelligent RobotsSelf-Healing, Adaptive, and Growing Technologies for Intelligent Robots

User’s new needs

LimitedResources


Example Example –– (1)(1)

• Need for safety: an old man says “move carefully!”– During a party: a lot of unrecorded moving objects

PathPlanner

MapBuilder Localizer

MontionController

AutoMove

LaserProxy VisionProxySonarProxySkinProxy

Quality requirement:Don’t care execution time, but avoidall possible collisions

Quality requirement:Don’t care execution time, but avoidall possible collisions

Reconfigured Architecture:ready to use various sensors and tomake a map by using the sensors.

Reconfigured Architecture:ready to use various sensors and tomake a map by using the sensors.

4


Example Example –– (2)(2)

• Need for agility: an old man says “move quickly!”– The old man is home alone: no unrecorded object

PathPlanner

MapBuilder Localizer

MontionController

AutoMove

LaserProxy

Quality requirement:Short response time, Shortexecution time, optimize use ofbatteries.

Quality requirement:Short response time, Shortexecution time, optimize use ofbatteries.

Reconfigured Architecture:ready to use one sensor(laser).It cannot avoid (unrecorded) moving objects

Reconfigured Architecture:ready to use one sensor(laser).It cannot avoid (unrecorded) moving objects


Our ApproachOur Approach

MonitoringMonitoring

DecisionDecisionMakingMaking

ReconfiguReconfigu--rationration

BrokeringBrokering LearningLearning

5


SHAGE FrameworkSHAGE Framework

• SHAGE(Self-Healing, Adaptive, and Growing SoftwarE) Framework integrates following technologies– Monitoring

– Brokering: Ontology(authoring relations between environmental information and architectural information)

– Decision & Learning: Case-Based Decision Theory

– Reconfiguration: Slot-based architectural style

• This framework provides a test bed for self-managed software.


SHAGE Overall ArchitectureSHAGE Overall Architecture

6


Architecture/Component BrokerArchitecture/Component Broker

<i1, s1>

…

A s

et o

f Inclin

atio

n-Situ

atio

n P

airs

<i1, s2>

<in, sm>

Current Situation(given by the Task Manager)

…

A set of architecture configurations(Abstract Level Architectures)

Candidate Set(reduced by theArchitecture Broker)


Architecture/Component BrokerArchitecture/Component Broker

• Role– Searching abstract-level architecture configurations related to the

current situation.

• Technology– Ontological descriptions.

• Current Status– It can only search in a small set of configurations related to the

navigation subsystem.

– Rule-based search: it cannot relax rules.

7


Decision Maker & LearnerDecision Maker & Learner

…A set of architecture configurations

(Abstract Level Architectures)

Candidate Set(reduced by theArchitecture Broker)

Selected ConcreteComponent(Selected by theDecision Maker)

Selected ArchitectureConfiguration(Selected by theDecision Maker)


Decision Maker & LearnerDecision Maker & Learner

• Role– Selecting exactly one configuration and one component for each slot

from the candidate set retrieved by the architecture broker.

• Technology– Case-Based Decision Theory

• Current Status– It only carries out in limited scope.

– Limited search space: only in the navigation subsystem.

– Limited learning time: few scenarios.

8


ReconfiguratorReconfigurator

Concrete Level

Abstract Level

Slot: MotionControl Slot: Coordinator Slot: Localizer

Slot: PathPlanner

Slot: MapBuilder

Slot: MotionControl Slot: Coordinator Slot: SLAM

Slot: PathPlanner

Slot: MotionControl Slot: Coordinator Slot: Localizer

Slot: PathPlanner

Slot: MapBuilder

A B C

D E

Slot: MotionControl Slot: Coordinator Slot: SLAM

Slot: PathPlanner

A.1 B F

G

C1 C1

C1 C1

C1 C2

C2

After ReconfigurationAfter ReconfigurationBefore ReconfigurationBefore Reconfiguration


ReconfiguratorReconfigurator

• Role– Reconfiguring the current software architecture dynamically.

• Technology– Slot-based two-level software architectural style.

• Current Status– It has reconfigured only the navigation subsystem.

• All configurations for the subsystem were verified in the demonstration.

– It can manage components distributed in SBCs(Single Board Computers) by RMI.

– It supports components implemented in Java and C++(through JNI).

9


DemonstrationDemonstration


Research IssuesResearch Issues

• Internal monitoring

• Ontology construction

• Learning speed

• Run-time measurement and validation

• Componentization

• Domain Knowledge

10


ConclusionsConclusions

• SHAGE Framework has been developed to provide ‘self-managing capabilities’ to robot software.

• The framework integrated ontology, decision theory, and dynamic architecture and comprises

– Monitor

– Architecture/Component Broker

– Decision Maker & Learner

– Reconfigurator

• In the experiment, we showed a simple scenario.– The robot managed its architecture to adapt its changing requirements.

• The second phase(three years) of our research just launched to improve key technologies.

shage: a framework for self-managed robot software · 1 may 21~22, 2006 seams, shanghai, china...

Documents