Work

<Internship> Concordia NGCI - CITYplayer ARTM

Digital Twin
Gamification
Sandbox Game

This research project was developed with the understanding that achieving the efficient decarbonization of transportation within a community needs solutions rooted in bottom-up feedback .

A Photo of Montreal downtown

Introduction

1. High-Fidelity Digital Twin Construction

Overview: Developed a realistic digital twin of Montreal’s downtown core to serve as a simulation platform for urban decarbonization. The project quality garnered significant attention from the Montreal Transit Authority (ARTM).

Lighting & Atmosphere (Unity 6 Features): Leveraged the latest Unity 6 features to achieve photorealism without compromising performance.

  • Adaptive Probe Volumes (APV): Implemented APV to handle dynamic global illumination for the dense city geometry, replacing traditional lightmaps to allow for faster iteration.
  • Environmental Simulation: Programmed a custom Day-Night cycle blending system and weather simulation, ensuring the lighting and skybox reacted dynamically to simulation parameters.
Montreal Downtown Digital Twin Day Montreal Downtown Digital Twin Night
Unity 6 Rendering: Dynamic Day/Night Cycle with APV

2. Shader Optimization: Fake Interiors

The Problem: The legacy project relied on procedural geometry to generate building windows and interiors. This approach caused an explosion in GameObject count and Draw Calls, creating a severe CPU bottleneck for the city-scale simulation.

The Solution (Interior Mapping): I customized an existing shader to implement Interior Mapping tailored to our building facades.

  • Technique: Utilized Parallax Occlusion Mapping (POM) logic driven by a depth map (height map). This calculates UV offsets based on the view angle to simulate perspective-correct 3D rooms on a completely flat 2D quad, creating a convincing illusion of volume.
  • Impact: Drastically reduced the polygon count and Draw Calls, shifting the load from CPU to GPU, which stabilized the frame rate for complex urban scenes.
Legacy: Physical Geometry Windows (High Cost) New: Shader-based Interior Mapping (Low Cost)
Optimization: Replaced physical meshes with Interior Mapping Shader

3. Automated Texture Pipeline & Team Training

Pipeline Automation: To texture over 60+ unique buildings in the downtown area efficiently, manual UV mapping and painting were not feasible. I designed a semi-automated pipeline:

  1. Capture: On-site 360° photography of building facades.
  2. Processing: Automated scripts for Photoshop/Lightroom de-distortion and batch cropping.
  3. Generation: Batch export through ShaderMap to generate PBR maps (Normal, Specular, AO).

Result: This pipeline accelerated asset production speed by roughly 300%, allowing a small team to populate the entire downtown core in weeks rather than months.

Texture Pipeline Workflow Diagram Texture Pipeline Showcase
Automated PBR Production Pipeline

Collaboration & Leadership: Standardized the modeling workflow and conducted team training sessions on Unity ProBuilder. This empowered level designers to perform modular building adjustments and texture replacements independently, reducing dependency on the core tech team.