Low Cost Virtual Reality Platform Done by: Peter Fang, Kevin Feng & Karen Wai Supervised by: Prof. Edwin Blake & Dave Maclay.

Low Cost Virtual Reality Low Cost Virtual Reality Platform Platform

Done by: Peter Fang, Kevin Feng & Done by: Peter Fang, Kevin Feng & Karen WaiKaren Wai

Supervised by: Prof. Edwin Blake & Supervised by: Prof. Edwin Blake & Dave MaclayDave Maclay

IntroductionIntroduction

• VR is expensive

• Benefit of cheap VR:– Education– Museum– Medical– etc

High end VRHigh end VR

• Requires:– CAVE-like projection system – High speed graphics workstation– Stereo shutter glasses– Motion tracking device– Very expensive!

Aim of this projectAim of this project

• This project:– Cheap VR– PCs– Use high-end PC graphics accelerators– Distributed network– Explore the use of PS2 in VR

Work allocationWork allocation

• 3 parts in this project:– PC VR subsystem (Kevin)– PS2 VR subsystem (Peter)– Distributed network subsystem (Karen)

PC-Driven VR subsystemPC-Driven VR subsystem

• 4 projectors

• Passive stereo

• Corner projection

• 3 PCs connected by distributed network system

• 2 Dual output display card

• Each display PC control 2 projectors to produce stereo effect

Control 5.1 Surround Sound System

L2

Display 1Hub

L1

Left

10/100 Mbps Ethernet

R2

Display 2

R1

Right

Diagrammatic representation of the PC-Driven Diagrammatic representation of the PC-Driven SubsystemSubsystem

Advantage & DisadvantageAdvantage & Disadvantage

• Advantage– Easy to implement

• Disadvantage– Higher cost

DevelopmentDevelopment

• Torque engine used to construct VE

• MS VC++

• Windows XP

Possible ExperimentsPossible Experiments

• Compare PS2 and PC presence

• Compare PS2 and PC performance

• Different hardware configuration

RiskRisk

• Delivery of hardware may delay

• Our strategy:– Schedule the implementation so only 2

projectors are required until the final stage

• Alternate the use of projectors for stereo and corner projection

PlayStation2-Driven VR subsystemPlayStation2-Driven VR subsystem

• 3 PS2 connected by distributed network• No dual output on PS2• Can only produce stereo or corner projection

with 3 PS2• 2 weeks of investigation to decide which is more

appropriate and feasible• Goal: demonstrate feasibility of VR using PS2• Outcome: one VR demo using PS2

DevelopmentDevelopment

• How to program a PS2:– Linux Kit– Consist of HD,

Network adaptor, mouse, keyboard, Linux, cable

– Available from www.linuxplay.com

L1

Display 1

Left

Hub

R1

Display 2

Right

Control

10/100 Mbps Ethernet

Diagrammatic representation of the PS2-Diagrammatic representation of the PS2-Driven SubsystemDriven Subsystem

Advantages & DisadvantagesAdvantages & Disadvantages

• Advantage of PS2– Low cost– Equal performance of hardware

• Disadvantage– No audio in the proposed VR system– High software implementation complexity

Possible ExperimentsPossible Experiments

• Comparison with PC in hardware performance

• Comparison of user presence level

Risk 1Risk 1

• Lack of expertise in the department -> less support

• More difficult to program than PC because of different architecture

• Skills of PC programming cannot be easily transferred to PS2 programming

• Expect minimum system can be built• Analysis of the system will constitute the

major part of the project

Risks 2Risks 2

– No game engine -> no easy programming– OpenGL slow on PS2 due to different

architecture– Expect slow performance if no optimization in

PS2 assembly– However, we try to demonstrate feasibility,

therefore not a priority in this project– Confident that algorithm optimization will give

us interactive frame rate with PS2

Distributed network subsystem Distributed network subsystem

• PC driven VR system– 2 displaying computers & 1 control

• PS 2 driven – 3 PS 2– 1 for control, others for display

• Links by 10/100 Mbps Fast Ethernet

ProblemProblem

• Synchronisation– Imbalance CPU & Graphics card– Equal performance PCs but with different

complexities for displaying scenes

• Need– Consistent frame rate – Synchronized display of graphics

GoalGoal

• Synchronization of the different projector displays with mean out of sync of less than 0.1 second

Possible ExperimentsPossible Experiments

• Performance analysis of the system in relation to # of computers

• Response time of the display PCs and synchronization overheads

CostCost

Total estimated cost

≈ R151218 Detail in proposal

Hardware

Software

Others

92%

6.5% 1.5%

ResourcesResources

• Hardware– 3 PCs with dual output graphics card

(R31000)– 3 PS2s with Linux kits (R15000)– Projectors (R80000)– Monitors, Network cards, Polarising filters etc.

ResourcesResources

• Software– Torque Engine (US $ 200 ≈ R1652)– Windows XP Professional, Visual C++.Net

• Others– IP addresses etc.