+ All Categories
Home > Documents > Software Testing for Intelligent Robots

Software Testing for Intelligent Robots

Date post: 06-Jan-2016
Category:
Upload: yosef
View: 38 times
Download: 0 times
Share this document with a friend
Description:
Software Testing for Intelligent Robots. Justin Peckner Maria Velasquez November 13, 2012. Paper Reviewed. Yun Koo Chung and Sun-Myung Hwang. "Software Testing for Intelligent Robots", International Conference on Control, Automation, and Systems, Seoul, 2007. - PowerPoint PPT Presentation
Popular Tags:
16
Software Testing for Intelligent Robots Justin Peckner Maria Velasquez November 13, 2012
Transcript
Page 1: Software Testing for Intelligent Robots

Software Testing for Intelligent

RobotsJustin Peckner

Maria VelasquezNovember 13, 2012

Page 2: Software Testing for Intelligent Robots

•Yun Koo Chung and Sun-Myung Hwang. "Software Testing for Intelligent Robots", International Conference on Control, Automation, and Systems, Seoul, 2007.

•Relevance: shows how software testing can be adapted to a particular software type, using ISO standards.

Paper Reviewed

Page 3: Software Testing for Intelligent Robots

•Robots are becoming increasingly used in businesses and homes.

•Software components such as voice recognition, facial recognition, movement control, and more, each need thorough testing.

•Paper describes researchers' spearheading of an initiative for testing robotics software using ISO standards.

Introduction

Page 4: Software Testing for Intelligent Robots

•Ubiquitous Robot Companion (URC): Korean project aimed at developing server/client based robotso "Robots will offer news, weather,

recipes, [and] will clean homes" *o Robots maintain connection to a servero As much processing of robot functions

as possible is done server-sideo Results in less expensive robots

Background

* http://robotgossip.blogspot.com/2006/06/korea-picks-ubiquitous-robot-companion.html

Page 5: Software Testing for Intelligent Robots

•The testing process is based on ISO 9126 and ISO 14598-5.

•9126 specifies attributes to strive for.o The researchers took the 30 most

important specifications from this standard (trimmed down from about 80).

•14598-5 specifies how to test these attributes.

Applying ISO Standards

Page 6: Software Testing for Intelligent Robots

ISO/IEC 9126 consists of six standards:

•Functionality: how capable is the software of carrying out its specified requirements?o Submetrics: Suitability, Accuracy,

Interoperability, Security, Functional Compliance

ISO/IEC 9126: Quality Attributes

Page 7: Software Testing for Intelligent Robots

•Reliability: how well does the software maintain its specified performance, when used under normal conditions?o Submetrics: Maturity, Fault Tolerance,

Recoverability, Reliability Compliance

•Usability: how easy is it for users to understand and use the software?o Submetrics: Understandability,

Learnability, Operatability, Attractiveness, Usability Compliance

ISO/IEC 9126: Quality Attributes

Page 8: Software Testing for Intelligent Robots

•Efficiency: how well does the software manage performance of the robot?o Submetrics: Time Behavior, Resource

Behavior, Efficiency Compliance

•Maintainability: how much effort is required to change the code over time?o Submetrics: Analyzability, Changability,

Stability, Testability, Compliance

ISO/IEC 9126: Quality Attributes

Page 9: Software Testing for Intelligent Robots

•Portability: how ready is the software to be transferred to a different environment?o Submetrics: Adaptability, Co-Existance,

Replacibility, Compliance

ISO/IEC 9126: Quality Attributes

Page 10: Software Testing for Intelligent Robots

1.Analysis of requirements: what are specific features that customers want?

2.Evaluation specification: for each requirement in step #1, how will we measure it?

ISO 14598-5: Testing the Attributes

Analysis of evaluation

requirementsEvaluation

specification

Design and definition of

evaluation plan

Execution of evaluation plan

Evaluation report

Page 11: Software Testing for Intelligent Robots

3. Design/definition of evaluation plan: create a specific, repeatable testing plan.

4. Execution of plan.5. Evaluation report.

ISO 14598-5: Testing the Attributes

Analysis of evaluation

requirementsEvaluation

specification

Design and definition of

evaluation plan

Execution of evaluation plan

Evaluation report

Page 12: Software Testing for Intelligent Robots

Robot Software Components

Others...

Visual recognition of user

faces/gestures

Audible recognition of user

voice

Human_Robot Interactions (HRI)

Smart Action Technology

Environmental mapping

Physical object manipulation

Page 13: Software Testing for Intelligent Robots

•In developing test cases, the authors specify that each case should:o reduce the number of additional cases

needed by more than one. Unclear about how exactly to do this.

o detect as many errors as possible (even if finding them is not the specific goal of the test case). Some people might find this approach too broad.

Test Case Scenarios

Page 14: Software Testing for Intelligent Robots

1.Identify the requirements in the specification. [ISO14598-5, step #1]a. Examples: The robot shall...

i. ...map its surroundings using a supersonic sensor.

ii. ...recognize the faces of humans it has already been programmed to know at least 95% of the time.

iii. ...distinguish between the voices of humans it has already been programmed to know.

Creating Test Cases

Page 15: Software Testing for Intelligent Robots

2. Create a test scenario targeting at least one requirement (more if possible). [ISO14598-5, step #2]

3. Continue this until all requirements have been covered.

4. Create a specific, repeatable test plan for the scenario. [ISO14598-5, step #3]

Creating Test Cases

Page 16: Software Testing for Intelligent Robots

•The paper focuses on presenting an approach to URC Robot Software Components testing, by adapting both ISO 9126 and 14598.

•Beginning of an initiative to establish a standard in robot software testing.

•Good demonstration of applying abstract standards to a practical project.

Conclusion


Recommended