TANGO turns the virtual agents of a GAMA simulation into real, moving robots on a physical model. It is an open, frugal framework for tangible agent-based simulation — and for the two-way dialogue between a researcher's hands and their model.
Tangible agent-based models (ABM) help people feel a simulation in space. Until now, that tangibility stopped at the screen. TANGO is an exploratory step further — bringing the agents themselves off the screen and onto the model.
Most tangible ABMs use projected simulation on a 3D mock-up. The maquette is real; the simulated agents are not. Users see the model react to placed objects, but the agents inside it remain immaterial overlays.
TANGO replaces projected overlays with low-cost, modular robots that embody the agents. GAMA pilots their trajectories live; congestion, coordination, and emergent behaviours can be observed in the room — not on a screen.
A thin, generic pipeline: an ABM in GAMA pushes agent states to a synchronisation layer; the layer translates them into commands for a fleet of robots moving on a physical model — and reads their real positions back into the simulation.
The agent-based simulation defines the agents, their behaviour, and the environment. Existing GAMA models can plug into TANGO without rewriting their logic.
A generic connector translates agent states into motion commands, monitors real-world drift, and corrects positions so digital and physical stay coherent.
Small modular robots, 3D-printed and Wi-Fi-connected, embody the agents on a real model surface — visible, hackable, and (eventually) interactive.
TANGO is not a closed product. It is an opinionated reference implementation that other researchers can fork, extend, or replace piece by piece. Three architectures define the project.
Modular robots designed for FDM 3D printing. Documented shells, off-the-shelf electronics, and a hackable mechanical base so any lab can build, customise, and repair its fleet.
A generic GAMA ↔ robot connector built on standard practices. Plug an existing model in; get a fleet moving without writing robotics code. Source, schemas, and CAD released openly.
The hard problem: keeping the simulated and the real in lockstep. TANGO embeds an error-correction loop so agent positions stay coherent even when the world disagrees with the model.
Tangibility is most useful where space, coordination, and emergence matter — where a 2D screen flattens what the eye should immediately see. Three scenarios anchor the project.
Test crowd routing and signage placement against the real spatial constraints of a neighbourhood model — bottlenecks become visible as bodies, not heatmaps.
Optimise delivery flows and priority rules with a live fleet of robots circulating on a district mock-up. See congestion appear; iterate on coordination algorithms in the room.
Rehearse coordinated interventions of service robots through damaged terrain. Confront theoretical planners with friction the screen never shows.
TANGO is a focused, ten-month exploratory project. The aim is not a finished platform but a credible, documented proof — enough to convince a community that tangible ABMs can be embodied, not just projected.
Hardware purchase, lab setup, GAMA model selection for the demonstrator scenario.
First printable chassis, motor control, on-board comms. One agent, one robot, end-to-end.
Generic GAMA connector, position tracking, drift correction loop validated at scale > 1.
Full mock-up + fleet, running a logistics or evacuation scenario at minimal latency.
Documentation, CAD plans, source code, and a paper submitted to JASSS / AAMAS.
Three artefacts, each meant for a different audience: a working demo for the lab, a reusable framework for researchers, and a paper for the field.
A mock-up, a robot fleet, and a coupled GAMA model running a complete tangible-simulation scenario — fully documented end to end.
CAD plans, firmware, the GAMA connector, and the synchronisation layer — released openly for the community to fork, audit, and extend.
A submission to an international venue at the intersection of agent-based modelling and tangible interfaces — JASSS or AAMAS as primary targets.
TANGO is run by two researchers based at the ACROSS Lab in Hanoi, Vietnam, with prior work on tangible ABMs (Hoan Kiem Air, CityScope Hanoi, etc) and on the GAMA platform itself.
TANGO is a deliberate next step from a lineage of tangible ABM and swarm-robotics research. A selection of the work the project builds on.