Asking the fleet what it is doing…
monad-knowledge Wi-Fi sensing lab · FIIT STU
SIM507 active 3 / 12 weeks

Simulation Toolkits for Sensing & Crowd Research

A twelve-week, tool-agnostic course on how to operate a family of research simulators and the reproducible workflow that turns a simulator run into defensible scientific evidence. The course teaches the class of each tool — an agent-based pedestrian engine, a trajectory-measurement library, a ray-traced channel simulator, and the notebook-as-simulator pattern — rather than any single product, then shows how a knowledge platform (campaigns, sessions, runs, vault notes) makes a sprawl of runs reproducible and citable.

The throughline: a simulation is a scientific instrument. An instrument you cannot reproduce, version, and validate is not evidence — it is an opinion with a plot attached.

The 12-week arc

3 of 12 lectures published — each a self-contained read with a slide deck and runnable workbook. Start at the first published week and work forward.

Wk Theme & focus Read
01 What is a research simulator? start here The four simulator classes; determinism, seeds, the run / artefact / manifest contract read → slides · workbook
02 Pedestrian dynamics on the platform Operating an agent-based engine + a measurement library (JuPedSim → PedPy) end-to-end read → slides · workbook
03 Ray-traced channel simulation Scenes, emitters, scatterers; synthetic CSI/BLE; occlusion and hardware impairments. Must reflect the IP-119 reparameterisation: occupancy diffuse power is channel-relative Rician K (K_dB(N) = k0 − slope·N, per-link diffuse = \ read → slides · workbook
04 The notebook as a simulator Parameterised, papermill-executed notebooks; interactive replays as artefacts planned
05 Coupling simulators Crowd → channel co-simulation; data contracts between pipeline stages planned
06 Reproducibility & provenance Pinned digests, git SHA, seeds, deterministic artefacts, manifest schema checks planned
07 Parameter sweeps & experiment design Grids, factors, seeds, confounds; what a sweep can and cannot tell you planned
08 Supervised research campaigns Brief → sessions → runs → synthesis; success criteria; negative results planned
09 Artefact stores & cloud runs Object-store layout, signed URLs, fetching and inspecting run artefacts planned
10 Validation & the sim-to-real gap Comparing synthetic to measured; metrics; where fidelity breaks planned
11 Scaling & orchestration CI dispatch, container isolation, cost and budget control planned
12 Capstone Design, run, and synthesise a small reproducible campaign end-to-end planned

Capstone

Week 12: pose a research question answerable by simulation, write a one-page brief with explicit success criteria, run a small sweep or campaign on the platform, synthesise the result (including any negative result), and hand in a package a classmate can re-run from the manifest alone.