# MDS-MAP localization simulation

This project simulates localization in an ad-hoc/sensor network using MDS-MAP
and two anchor-only baselines: circular lateration (nonlinear least squares)
and square lateration (min-max bounds).

## Install and run

Use Python 3.10 or newer:

```powershell
python -m pip install -r requirements.txt
python scripts/run_single.py
python scripts/run_remote_carlo.py
pytest
```

The default experiment places 4 anchors and 16 unknown nodes in a 100 m
square, uses a 55 m radio range, Gaussian ranging noise, and 30 repetitions.
Outputs are written to `results/figures` and `results/tables`. Edit
`config/default.yaml` to change the experiment. A connected graph is required
for shortest-path completion; increase `radio_range` if a random run is
disconnected.

The position figure compares actual unknown positions with all estimates and
anchors. The error-distribution figure is a histogram over all unknown nodes
and repetitions. The metrics CSV contains RMSE, MAE, mean, standard deviation,
median, maximum, and 95th percentile errors for every repetition and method.