Section 01
Problem Statement
A site twin needs a model + telemetry + scenario engine. Without a drone and without desktop BIM viewers, the team must turn the phone into the capture, visualisation, and simulation device: take a set of photos walking around a model / building / room to reconstruct a point cloud or mesh (photogrammetry may run on a laptop, but the phone app ingests and displays the mesh); all field interactions — progress toggle, scenario selection, AR overlay — happen on the phone. A twin that only works on desktop cannot be judged as a twin; the phone app is the interface.
Section 02
Solution Requirements
Phone capture photogrammetry
Walk around the site model / a real structure recording a circular video or 25+ photos; GPS + gravity sensor provide camera orientation via SensorManager / ARCore. Photos may be stitched with Meshroom/COLMAP on any computer, but the phone viewer loads the resulting .obj/.pts/.glb. ARCore depth is an accepted alternative.
IFC-lite ingestion on the phone
Use ifcopenshell (or a small server) to translate IFC to simple JSON meshes delivered to the Flutter/Kotlin 3D viewer (three.js in a WebView, or Filament/FBX); the model rotates on the phone screen.
Live progress tagging via camera & QR
Scan a printed QR on a pillar; the scan opens the twin and the user toggles "poured / not poured / late"; status propagates over WebSocket / Firebase in real time and the twin color-codes the corresponding column.
ARCore overlay: "what should be here right now?"
Using ARCore AugmentedImage / Geospatial API, project the BIM slab outline onto the live camera view with a red/green fit check; judges can see both the real wall and the overlay simultaneously.
Scenario "what-if" engine
Pick one of three scenarios (48h rain delay / steel 3wk late / crane moved); a CP-SAT solver (or server) recomputes the completion date; Gantt bars animate and the predicted date delta is shown in large text on the twin screen.
Phone-native weather-risk telemetry
Combine battery temperature, barometric pressure, GPS altitude, and Fused Location speed with an OpenWeatherMap lookup into a "weather risk index" shown live in a corner widget — no external sensor used.
Multi-user sync
Two team phones open the twin; a status change on one appears on the other within ≤ 3s over Firebase / socket.io.
Section 03
Expected Solution / Outcomes
- The 3D mesh rotates in the app, loaded from a real photogrammetric result.
- Tapping a real QR on a pillar marks a column red; a partner's phone turns the same column red over the web.
- A phone pointed at a real wall shows a BIM rebar overlay via ARCore, wall visible behind it.
- Tapping the "rain" card changes the project date live on the phone screen.
- Phone altitude + pressure feed an offline "rain risk" widget.
Section 04
Technology & Hardware Constraints
Data: IFC test files, mesh .obj files, ARCore tracking images / spatial anchors.
Section 05
Scoring
Requirements Completeness
50
Each requirement (R-id) is worth 3–12 marks. A hard-coded or mocked claim earns 0 — only working code counts.
Presentation & Demo
50
3-slide pitch + video ≤ 7 min + live demo. Judges must understand the problem in ≤ 90 seconds.
Real-world Implementation & Feasibility
50
Would a PM, engineer, or worker install this with only a phone on Day 1 of the next project? Must work offline and cheaply. Field-trial videos prove real-environment execution (20+ pts of this score is the field video).
Innovation & Disruption
50
A novel fusion of phone sensors (e.g. IMU + camera + barometer = scaffold-instability detector) is rewarded over simple API-calling. Re-skinning an existing SaaS scores low.
Section 06
Bonus (+10)
"Time-lapse twin": a 10-second phone video per day of the same real site corner; a TFLite image-similarity pixel-diff per day auto-generates a Gantt deviation line per façade element.
Section 07
Deliverables
| Item | Format |
|---|---|
| Video Demo | ≤ 7 min: walk-around capture → 3D viewer → AR overlay → scenario. Mesh compute may run on a computer; the transfer to the phone viewer must be shown. |
| Field Trial Videos | (a) Real walk-around of a parking lot / site perimeter with the phone. (b) ARCore demo on a real wall/door. |
| Demo Link | Android APK download link (installed on the reviewer's own phone) or a mobile-optimised web app URL. |
| GitHub | ifc_importer/, mesh_viewer/, arcore_app/, scenario_solver/, README explaining the PC-side mesh conversion (COLMAP / Open3D) feeding the phone viewer. |
| Pitch Deck | 3-slide PDF (Problem → Solution → Field-Proof & Impact). |
| Technical Brief | 2-page PDF: architecture diagram, phone-sensor map, tech stack, limitations, PRIVACY.md. |
Section 08
Field-Trial Requirements
- Real walk-around capture of an outdoor site/perimeter.
- ARCore overlay demoed on a real wall or door.
Section 09
Privacy Requirements
Any human face captured in a field video must be consented or blurred. A PRIVACY.md file is required in every repository.