AccessiRoom

Inspiration

Indoor accessibility is deeply personal: a room that works well for one person may create serious mobility barriers for another. However, evaluating a space often requires specialist knowledge, manual measurements, and imagining how furniture changes might affect movement.

We built AccessiRoom to make that process more approachable. Our goal was to help a person, caregiver, family member, or occupational therapist scan a room, identify potential barriers, and explore improvements using the mobility needs of the actual Room Occupant.

What it does

AccessiRoom is a native iPad app that turns a LiDAR room scan into an interactive accessibility assessment.

The user first creates and confirms a Mobility Profile containing the Room Occupant’s essential needs and preferences, such as minimum passage width and required turning space. They can then scan a room with Apple RoomPlan or use a bundled demo room.

After reviewing the scan, the user confirms important details such as entrances, movable objects, required destinations, approach areas, and turning zones. AccessiRoom then:

• Identifies mobility barriers and limiting clearances
• Evaluates routes from entrances to required destinations
• Produces an explainable Arrangement Status and Layout Score
• Distinguishes clear results from measurements that need verification
• Displays findings on a top-down Accessibility Map
• Lets the Operator move, rotate, or propose removing furniture
• Automatically reassesses every proposed change
• Compares the observed and proposed arrangements side by side

AccessiRoom does not claim to certify a room as accessible. It provides transparent, person-specific evidence to support better decisions.

How we built it

We built AccessiRoom as a native SwiftUI application for LiDAR-equipped iPads.

Apple RoomPlan captures the room’s architectural features, detected objects, dimensions, and placements. ARKit and RealityKit support spatial capture and 3D visualization, while a custom deterministic geometry engine converts the scan into a floor-plane model.

The assessment engine uses grid-based pathfinding and spatial collision checks to evaluate:

• Door and passage clearances
• Continuous routes through the room
• Approach and turning zones
• Furniture-to-furniture conflicts
• Furniture crossing architectural boundaries

A conservative measurement tolerance accounts for uncertainty in captured dimensions. Borderline results are reported as “Needs Verification” instead of being forced into a pass or failure.

Challenges we ran into

One of our biggest challenges was translating RoomPlan’s 3D output into reliable 2D geometry for routing, collision detection, and visualization. Detected objects can have imperfect dimensions, inconsistent categories, and complex coordinate transformations.

Accomplishments that we're proud of

The prototype supports live LiDAR room capture, on-device storage, configurable Mobility Profiles, deterministic route analysis, interactive Accessibility Maps, editable furniture arrangements, automatic reassessment, and side-by-side comparisons in a nice looking application.

What we learned

From a technical perspective, we gained experience with RoomPlan, ARKit, RealityKit, SwiftUI, coordinate transformations, computational geometry, pathfinding, and persistent state.

What's next for AccessiRoom

A next step for AccessiRoom is to provide more detailed information on object heights, interaction types, and better fidelity shapes of room objects in order to provide stronger scores representative of occupant profiles.

Built With

• arkit
• realitykit
• roomplan
• swiftui