Team product / Social digital twin

TectoTrack

A living simulation environment for studying how people move, choose, wait, and experience complex buildings.

Role
Computational design & simulation development
Period
2023 — present
Status
Active development
Stack
Unity / C# / Python / BIM
Agent-based modellingRoutingHuman behaviourSpatial analytics
Question

How can a building model include people as active, uncertain participants rather than static occupancy values?

Response

TectoTrack combines a digital building environment with agent-level decisions, routing, queues, services, accessibility rules, and reportable simulation data.

Agent-level behaviour and task logic
BIM → runtime repeatable model pipeline
Live system routing, queues and services
Decision-ready spatial and operational evidence
01

System, not a scene

TectoTrack is not a rendered copy of a building. Its digital environment stays connected to a simulation graph, Function Objects, agent tasks, vertical transportation, service queues, and analytical outputs.

The central design problem is behavioural: an agent must perceive available choices, commit to a destination, move through multiple levels, react to congestion, and still remain explainable when something goes wrong.

02

My contribution

Behaviour architecture

Revising agent decisions, target commitment, retry states, queue lifecycles, and task-to-route translation.

Routing & accessibility

Multi-level paths, stairs, escalators and elevators, including wheelchair-compatible reachability.

BIM-to-simulation

Helping structure the FBX and metadata pipeline so design models enter Unity with consistent coordinates and semantics.

Evidence & documentation

Turning debugging history into architecture decisions, invariants, test gates, and maintainable technical documentation.

03

A behavioural loop

  1. 01

    Read context

    The agent receives a persona, task sequence, location, constraints, and available services.

  2. 02

    Choose & commit

    The decision layer selects a reachable target without confusing intention with intermediate route objects.

  3. 03

    Move & negotiate

    Navigation, queues, level transitions, capacity, and other agents continuously shape the path.

  4. 04

    Record consequences

    Movement, waiting, density, service use, and failure states become analysable evidence.

04

Engineering underneath

Queue as a physical system

Explicit FIFO order, slot ownership, vacancy propagation, overflow policy, and collision-safe movement.

Simulation before spectacle

A persistent simulation shell is separated from visual chunks to control memory without changing agent logic.

Failure is observable

Incomplete routes, stale callbacks, unreachable services, and state conflicts are documented rather than silently hidden.

Scale through contracts

Coordinate, lifecycle, readiness, and ownership contracts keep a large collaborative codebase understandable.

05

What it proves — and what it does not

Current position

The platform demonstrates a substantial, working simulation architecture for complex facilities. Behavioural realism still requires calibration against observed people and project-specific data; a convincing animation alone is not validation.

06

Credits

TectoTrack is a Morphotect team product led by Ali Jabbari. My work sits inside a much larger collaborative Unity, simulation, data, and product effort.

Project animation

Page