Applications of virtual reality

Design, train, care, explore, meet

Étienne Peillard
IMT Atlantique

What is virtual reality used for?

Five application families, one question: what does VR add to the activity?

3.1

Design and industry

Which decision can be made before manufacturing?

Can this part be assembled?

Access to the fastener

Can the tool and hand reach the target?

Assembly order

Which part must be positioned first?

Decision: change the design before manufacturing.

Can this part be assembled?

Virtual prototypes allow design questions before a physical prototype exists. The examples include product appearance, seat insertion, manufacturing processes and workstation ergonomics. Fidelity must support the question being asked: correct dimensions and reach may matter more than surface appearance. A model can omit physical constraints unless those constraints are explicitly simulated.

Existing digital models provide the starting point. VR allows inspection at body scale: can a worker reach a fastener, insert a tool and assemble parts in the intended order? The lecture contrasts modifications to a physical assembly mock-up with inspection of a digital version. The key question is which decision becomes possible before manufacturing. Access constraints and collisions must be represented for that inspection to be meaningful.

Architecture and spatial design

Human scale

Experience proportions and circulation

Alternative proposals

Compare design choices

Immersive authoring

Create and modify in space

From CAD model to immersive visit

Stage	Design question
CAD model	Which elements support the task?
Scene preparation	Which details can be simplified?
Headset trial	Is the visit fluid and readable?

Existing 3D content still requires preparation.

Understanding a structure generated in 3D

Spatial question

Where does the flow go? Which layers intersect?

Exploration

Change viewpoint or representation

Visual comparison of simulated phenomena

Understanding also depends on tools and expertise.

Understanding a structure generated in 3D

The examples include flow, temperature and scientific structures. Spatial viewing can help reveal relationships, but a 3D display is not automatically the clearest way to answer every data question. Discuss the role of abstraction, comparison, colour and task-specific interaction. These are historical demonstration clips, not claims about current products.

The lecture emphasises preserving the spatial dimension of data generated in 3D. Screen projections require some mental reconstruction, particularly when layers overlap. Immersion offers another way to inspect these relationships. Comparisons must account for expertise with existing tools: interpreting 2D slices is itself a learned skill. VR should be assessed through understanding achieved on a specific task.

Screen or immersion: the same spatial question

Same task: locate a flow	Screen	Immersion
Access the structure	Slices, rotation, views	Moving viewpoint, spatial exploration
Assess the contribution	Same question and same data	Compare errors, time and effort

Choose a representation for a specific task.

BMW: production planning at scale

Shared factory data

Layouts, equipment and logistics

Before physical changes

Simulate clearances and production flows

Reported scale

Digital twins of over 30 production sites

BMW Virtual Factory simulation of a vehicle moving through a production line — © BMW Group · Virtual Factory, 2025

Separate the industrial simulation from the immersive interface.

BMW: production planning at scale

BMW reported scaling Virtual Factory applications across digital twins of more than 30 production sites in June 2025. The platform combines industrial data and simulation to support planning. This is evidence of an operational digital-twin workflow, not a controlled demonstration of the added benefit of a headset. Immersive review is one possible interface to such a world; automated collision checks and desktop tools do not require VR.

The lecture introduces a digital twin as a virtual representation connected to information about a physical system, its state and activity. A geometric model describes shapes; a twin project adds data for monitoring or simulation. Specify what is updated and how frequently rather than assuming a perfectly exact, instantaneous copy. VR can provide a spatial interface to this information. A factory-planning video alone does not demonstrate complete live synchronisation.

Model, simulation and digital twin

Model

Describe shapes and positions

Simulation

Compute behaviour within a scenario

Digital twin

Connect the model to physical-system data

Which data are updated, from where and how often?

Design: from geometry to a decision

Target decision

Access, assembly, circulation or data interpretation

Useful fidelity

Preserve constraints needed for the decision

Rivian design review Physical and virtual components Modelisar

3.2

Training and skills

What is learned in the simulator?

Training for rare or risky situations

Repeatability

Practise the same situation

Controlled variation

Change task conditions

Reduced exposure

Simulate danger or unavailable equipment

What is learned in the simulator?

Goal	Practice	Assessment
Understand the procedure	Order the actions	Omissions in the real situation
Adjust the movement	Connect action and sensory feedback	Accuracy and task errors
Adapt to a case	Vary events	Performance on a new variant

Procedures and team coordination

Individual skill

Tools and action sequences

Shared activity

Roles, timing and communication

Debriefing

Review decisions and outcomes

From surgical observations to a collaborative training scenario

Trial records inform debriefing

GoalChoose a skill to practise
Trial and recordsIdentify an omitted step or late handover
DebriefingExplain the decision and its effect
Next trialAdapt the scenario to the skill needing practice

A needle-insertion simulator

Required fidelity

Tool path and tissue resistance

Useful modalities

Vision and precise force feedback

Communication

A trainer may guide the learner

A needle-insertion simulator

A needle-insertion simulator illustrates task-dependent fidelity. The force felt during contact or penetration may carry information essential to the skill, while surrounding visual detail may contribute less. Training, remote instruction and actual teleoperation also impose different requirements for accuracy, delay and feedback. The appropriate interface follows from which part of the procedure the person needs to learn or control.

This example distinguishes understanding a task from finely adjusting a movement. A headset alone does not provide the resistance and contact required for every manual skill. The lecture discusses haptic interfaces and specialised simulators that align controls and feedback with a target task. Correspondence must be described for each device. A simulation demonstration alone does not establish skill transfer.

Lufthansa: VR in recurrent crew training

An existing workflow

Part of annual cabin-crew training

A repeated task

Search a virtual aircraft cabin

An operational constraint

Access without occupying a real aircraft

Routine use answers “is it used?”; transfer testing answers “does it work?”

How is transfer assessed?

Test outside the simulator

Does the procedure or movement succeed on the target task?

Relevant comparison

Against which training and for which learners?

The intended outcome determines the measurement.

Example: movement and force feedback Lufthansa demonstration

How is transfer assessed?

A simulator score is a proximal measure. Assess whether learning persists and transfers to the intended activity, preferably against an appropriate training alternative. Simulator fidelity should follow the skill: visual detail, forces and team communication have different importance for different tasks.

The oral discussion distinguishes understanding how to perform a task, learning a movement and adapting to a particular case. These goals need different assessments. Virtual preparation can precede real practice, but its contribution must be measured in that practice. The oral claim of faster subsequent real learning needs a specific study before being presented as a general result. For manual skills, evaluation should also examine errors and correspondence of sensory feedback.

3.3

Healthcare and clinical uses

What is controlled and how is the person supported?

Which situation parameters can be controlled?

Scenario parameters

Crowd density, proximity, space size

Observations

Emotion and behaviour in the scene

A historical exposure-therapy research example

The scenario belongs to a defined intervention and assessment.

Which situation parameters can be controlled?

A simulated situation can evoke emotion even when the participant knows that it is artificial. Research on phobias and post-traumatic stress uses this relationship within specified protocols. Interpretation depends on the population, exposure and outcome measure. An emotional reaction in a demonstration illustrates responsiveness to the scenario; it does not by itself establish a clinical benefit.

In the lecture's teaching example, the virtual environment offers adjustable parameters such as crowd density, space size or stimulus proximity. The scenario illustrates control and reproducibility of presented conditions. It provides neither a care protocol nor evidence for a particular indication. Use the example to identify controlled variables rather than prescribe a universal therapeutic progression.

What feedback accompanies movement?

Action

Perform a movement during the exercise

Visible feedback

Compare performed and intended movement

Measurement

Relate records to a functional objective

Movement tracking and care follow-up answer different questions.

What feedback accompanies movement?

Research on body perception includes altered body-size perception and virtual representations of missing limbs. These examples examine how sensory feedback can affect the representation of one’s body. A change in a perceptual measure is distinct from a broader therapeutic outcome. The interpretation depends on which experience was induced, what was measured and the conditions of the study.

The lecture develops distinct questions: supporting movement exercises, connecting action with visual body feedback and studying body perception. These should not be treated as a single clinical application. Phantom-limb and mirror examples recall the perception–action relationship here. Mechanisms and therapeutic outcomes require specific references. Projects concerning body perception remain research questions without generalisation to a diagnosis or population.

How can remote follow-up be organised?

Home exercisePerform the planned activity
Movement recordsMake the session available for review
Professional follow-upObserve live or review the trial
Feedback to the personAdapt instructions for the next session

Remote group rehabilitation: exercise with other patients in a shared virtual space.

RelieVRx: a defined clinical indication

Who?

Adults with chronic lower back pain

What?

Prescription, home-based adjunctive treatment

How?

Immersive behavioural skills and exercises

A clinical claim belongs to an intervention, population and outcome.

Healthcare: from mechanism to clinical use

Mechanism

Attention, graded exposure or feedback

Evaluation

Defined population, comparison and outcome

Care pathway

Indication, supervision and follow-up

Healthcare: from mechanism to clinical use

These are complementary steps, not a guarantee of progression. Laboratory changes in perception or emotion do not automatically establish clinical benefit. Discuss where evidence is needed for the particular rehabilitation, exposure or pain intervention under consideration.

A student describes a project supporting rehabilitation after breast cancer, with movement tracking and information sent to clinicians. This classroom account illustrates a possible workflow and is not a clinical evaluation. The lecture also discusses home exercises, live or later supervision and remote group practice. These examples raise questions about follow-up: who observes, when, and what feedback reaches the patient? The social dimension is a design direction discussed in class whose effects require evaluation.

3.4

Culture and entertainment

What does the experience offer its audience?

Games and heritage: two reasons for immersion

Games

Act with the body; account for space and session length

Heritage

Explore a lost, remote or inaccessible place

Virtual environments for public experiences

The headset choice depends on the intended experience.

Games and heritage: two reasons for immersion

Public applications range from entertainment to education and heritage interpretation. The learning or cultural purpose should guide interaction and content. Access to a reconstructed location does not remove uncertainty about the reconstruction. Distinguish documented evidence, interpretation and creative invention when presenting historical worlds.

The lecture asks how VR fits gaming habits: available space, effort, session length and interaction with people nearby. Scepticism about universal headset gaming is a position discussed in class rather than an established market forecast. Heritage offers a different argument: accessing and spatially exploring an absent, lost or inaccessible place. Distinguish a VR visit from tablet-based augmentation through the actual operation of the device.

Versailles: explore a vanished place

A shared visit

Free-roaming, multi-user VR

A reconstructed setting

Three lost spaces from the 1682 gardens

A public run

25 March 2025 – 4 January 2026

Reconstructed Royal Menagerie in the Versailles Lost Gardens experience — Royal Menagerie · © GEDEON / Small Creative / VIVE Arts / Versailles

Versailles: Lost Gardens — the VR experience

Culture: design the visitor journey

Content

What is explored or understood?

Participation

What can a visitor actually do?

Operation

Guidance, accessibility and visitor flow

Narrative and sensory perspectives Art and performance Public installations

3.5

Social virtual reality

Which activities do people share in VR?

Meet

Converse in a shared virtual place

Create and participate

Play, learn or join a performance

Represent oneself

Act through an avatar and its gestures

Second Life: a historical social-world example

The shared activity gives social VR its purpose.

Which activities do people share in VR?

Social virtual worlds organise interaction between people through spatial content and representations of participants. The metaverse idea, associated with Snow Crash, adds broader expectations about persistence, identity and connections between services. VR, social worlds and these narratives overlap without being equivalent. Headset use, interoperability, content governance and persistent identity are separate properties of a platform.

The discussion distinguishes promises of general adoption from practices in particular communities. The lecturer describes an investigation with an anthropologist involving VRChat users. In the interviews discussed, virtual activities extend artistic and social activities carried out elsewhere. This account invites examination of continuity between practices rather than an assumption of social isolation. Profiles and usage totals mentioned orally cannot characterise all users; the study reference and methods remain to be documented.

How can other people be made present?

Voice and gesture

Identify who is speaking and acting

Space and objects

Understand what is shared

Avatar

Express identity and intention

Place rules

Organise access and coexistence

VRChat: continuity of social practices

Investigation discussed in class · Privateer, Peillard and Loetscher, 2025

Activities

Extend an artistic or social practice

Places and identities

Choose a setting and representation

Research question

How do these practices connect in everyday life?

Examples reported in class; distinguish the inquiry’s scope from all users’ practices.

Conference contribution

Social VR: which activity, with whom?

Situation

Who meets, and for which activity?

Contribution to examine

What do shared space and avatars enable?

Social-world archives Shared learning Telepresence research

3.6

Choosing and evaluating a use

Which activity justifies choosing VR?

Compare evidence, not impressions

Family	Object of comparison	Possible criteria
Industry	Design decision	Errors and effort
Training	Transfer to the real task	Omissions and accuracy
Healthcare	Defined intervention and population	Target outcome and follow-up
Culture	Visitor journey	Experience and understanding
Social VR	Activity and participants	Coordination and shared experience

A building model: why choose VR?

Discussion case: you have a 3D model of a building.

Activity

Which task and which people?

Expected contribution

What does immersive exploration enable?

Limit

Which information or constraint is missing?

Comparison

How would you assess it against another solution?

From applications to system design

Start with the activity

Users, setting and intended outcome

Specify the interaction

Cues, actions and shared state

Select the evidence

Baseline, measurement and deployment context

Further examples and research Return to the five families

3.7

Further examples and research

Additional demonstrations and open questions

Resources for exploring each family

Additional demonstrations

Consult an example for the question being studied

Research

Observe, coordinate and explore new interfaces

Design and industry Training and skills Healthcare and clinical uses Culture and entertainment Social virtual reality Research questions Family map

Industrial design demonstration

What is physically touched? What is simulated?

A physical support can complement a virtual model.

Interior and HMI simulation: Modelisar

Industrial design: complementary views

Which dimensions must be faithful for the decision?

Spatial authoring interfaces

Microsoft Maquette

Prototype a space and its interfaces

Gravity Sketch

Create and share a 3D model

Collaborative exploration of scientific data

Scientific data in a spatial setting

Interactive inspection and thermal data

What information does spatial viewing add?

Rivian: review a vehicle inside the headset

A life-size model

Review proportions, surfaces and the interior

A shared decision

Bring design and engineering into the same scene

An iterative workflow

Change the digital model before physical mock-ups

BMW Virtual Factory: the industrial workflow

A shared surgical simulation

Task

Coordinate a procedure

Interfaces

Vision, communication and contact

Evaluation

Accuracy, cooperation and comfort

Training contexts

What is practised, and how would learning be assessed?

Lufthansa: extended reality crew training

How can body perception be studied?

Manipulation

Change a virtual-body property

Observation

Measure a defined perceptual response

A virtual limb controlled through physiological input

Narrative and sensory perspective

Notes on Blindness

A world organised through sound

Carne y Arena

An embodied spatial encounter

Traveling While Black

A situated narrative experience

Art, performance and virtual identity

Spectre: an immersive performance setting

Gender Swap and mediated perspective

Representation

A different visible body

Correspondence

Movement and viewpoint relation

Interpretation

Experience and its limits

Mediated experience and identity

Carne y Arena

Spatial installation: trailer

Traveling While Black

360° documentary: excerpt

How does each work place the participant within its subject?

Exhibitions and public space

Care / Cure

Architecture and an exhibition context

Débord de Loire

An event in a shared public setting

Immersive design

Content, physical access and mediation

Public installations and their setting

History and heritage settings

The physical venue remains part of the experience

Perceptual stages and projected environments

A staged environment and its physical construction

Projection and spatial illusion

Designing an immersive installation

Display geometry

Coverage, seams and viewpoint

Audience

Individual or collective experience

Physical setting

Movement and access

Projected content in a collective installation

Marketing and product experience

Showrooms

Explore configurations and scale

Demonstrations

Experience a proposed service

Assessment

Separate persuasion from task performance

Social virtual places: a historical comparison

Shared learning and web environments

What can participants share besides conversation?

Three research lenses

Observe people

Control worlds and measure responses

Coordinate people

Study shared space and remote action

Reconfigure interaction

Explore tangible and multisensory interfaces

VR as an experimental setting

Control

Manipulate the environment

Observation

Record actions and responses

Validity

Check the relation to the target situation

Controlled situations in virtual environments

Shared touch: studying social perception

Manipulation

Feel touch while observing a virtual agent

Study

Within-participant experiment · n = 32

Question

How does shared sensation affect social response?

Figure 1 from the shared body sensations study, showing an observer receiving haptic stimulation corresponding to a virtual agent — Tao, Egelman & Bailenson · ISMAR 2024 · Figure 1

Research lens: coordinate shared action

Representation

How are people and actions made visible?

Shared space

Which viewpoints and objects are common?

Coordination

How do people divide and combine control?

Collaborative virtual environments

Shared state

Participants affect a common world

Communication

Voice, gesture and spatial cues

Coordination

Awareness of others and their actions

Telepresence

Remote people

Represented in a shared environment

Spatial communication

Pose, gaze and gestures

A shared task

An activity beyond conversation

Telepresence archive

Sharing a projection room

Viewpoint

Several independently moving observers

Display

Separate view channels where supported

Coordination

Physical and virtual workspace

Multiple-viewer projection

Co-habitation in a CAVE

Co-located object manipulation

Physical support

A shared tangible relation

Virtual action

A common object or activity

Coordination

Who controls each part?

Co-located manipulation

Physical co-presence and virtual activity

Physical space

Participants occupy a common room

Virtual context

The displayed activities may differ

Interaction

Bodies can become part of another person’s constraints

Shared-space interaction research

Remote collaboration and group telepresence

Shared state

Objects, tools and participant actions

Representation

Simplified avatars or captured people

Group activity

Coordination across sites

Fleury et al. (VRST 2012)

Beck et al. (IEEE VR 2013)

Research lens: reconfigure the interface

Physical objects

Arrange and manipulate information

The body

Move around a shared workspace

Computation

Make relationships visible and executable

Dynamicland: shared physical computation

Physical objects

Manipulable parts of an activity

Projection

Digital information in the workspace

Collaboration

People work around a shared setting

Dynamicland archive

Computing with objects and the body

Arrange

Objects become part of a computation

Move and compare

The body participates in the activity