diff --git a/inverse/__init__.py b/inverse/__init__.py
new file mode 100644
index 0000000..e69de29
diff --git a/inverse/sample_sensors.py b/inverse/sample_sensors.py
new file mode 100644
index 0000000..d588c70
--- /dev/null
+++ b/inverse/sample_sensors.py
@@ -0,0 +1,112 @@
+"""
+Campionamento sensori dalla soluzione FDM.
+
+Legge results/fdm/fdm_solution.csv (o lancia il solver FDM se non esiste),
+chiede le posizioni dei sensori e la frequenza di campionamento, salva
+results/fdm/measurements.csv con colonne: x, t, T.
+"""
+
+import sys
+import os
+
+sys.path.insert(0, os.path.dirname(os.path.dirname(os.path.abspath(__file__))))
+
+import numpy as np
+import pandas as pd
+
+FDM_CSV  = "results/fdm/fdm_solution.csv"
+OUT_CSV  = "results/measurements.csv"
+
+
+def _load_or_solve():
+    if os.path.exists(FDM_CSV):
+        print(f"Carico soluzione FDM da {FDM_CSV} ...")
+        df = pd.read_csv(FDM_CSV)
+        x_vals = np.sort(df["x"].unique())
+        t_vals = np.sort(df["t"].unique())
+        T_matrix = df.pivot(index="x", columns="t", values="T").values
+        return T_matrix, x_vals, t_vals
+    else:
+        print(f"{FDM_CSV} non trovato. Lancio il solver FDM ...")
+        from fdm.solver import solve, save_csv
+        T_matrix, x_vals, t_vals = solve()
+        save_csv(T_matrix, x_vals, t_vals, FDM_CSV)
+        return T_matrix, x_vals, t_vals
+
+
+def _ask_sensors(L: float) -> list[float]:
+    print(f"\nDominio spaziale: [0, {L}] m")
+    print("Inserisci le posizioni dei sensori separate da virgola (es: 0.2, 0.5, 0.8):")
+    while True:
+        raw = input("  x sensori [m]: ").strip()
+        try:
+            positions = [float(v.strip()) for v in raw.split(",") if v.strip()]
+            if not positions:
+                raise ValueError
+            out_of_range = [p for p in positions if p < 0 or p > L]
+            if out_of_range:
+                print(f"  Valori fuori range [0, {L}]: {out_of_range}. Riprova.")
+                continue
+            return positions
+        except ValueError:
+            print("  Input non valido. Usa numeri separati da virgola.")
+
+
+def _ask_freq(dt_fdm: float) -> float:
+    max_freq = 1.0 / dt_fdm
+    print(f"\nFrequenza massima (passo FDM): {max_freq:.2f} Hz")
+    print("Inserisci la frequenza di campionamento [Hz]:")
+    while True:
+        raw = input("  f [Hz]: ").strip()
+        try:
+            f = float(raw)
+            if f <= 0:
+                raise ValueError
+            if f > max_freq:
+                print(f"  Frequenza troppo alta (max {max_freq:.2f} Hz). Viene usata la frequenza massima.")
+                f = max_freq
+            return f
+        except ValueError:
+            print("  Input non valido. Inserisci un numero positivo.")
+
+
+def main():
+    print("=" * 40)
+    print("  Campionamento sensori FDM")
+    print("=" * 40)
+
+    import config
+    T_matrix, x_vals, t_vals = _load_or_solve()
+
+    sensor_positions = _ask_sensors(config.L)
+    freq = _ask_freq(t_vals[1] - t_vals[0])
+
+    # Indici temporali al passo di campionamento
+    dt_sample = 1.0 / freq
+    t_sampled = np.arange(t_vals[0], t_vals[-1] + dt_sample * 0.5, dt_sample)
+    t_indices = [int(np.argmin(np.abs(t_vals - t))) for t in t_sampled]
+    t_indices = sorted(set(t_indices))
+
+    # Campionamento
+    rows = []
+    for x_s in sensor_positions:
+        i_x = int(np.argmin(np.abs(x_vals - x_s)))
+        x_actual = x_vals[i_x]
+        for i_t in t_indices:
+            rows.append({"x": x_actual, "t": t_vals[i_t], "T": T_matrix[i_x, i_t]})
+
+    df_out = pd.DataFrame(rows).sort_values(["x", "t"]).reset_index(drop=True)
+
+    os.makedirs(os.path.dirname(os.path.abspath(OUT_CSV)), exist_ok=True)
+    df_out.to_csv(OUT_CSV, index=False)
+
+    n_sensors = len(sensor_positions)
+    n_times   = len(t_indices)
+    print(f"\nSalvato: {OUT_CSV}")
+    print(f"  Sensori: {n_sensors}  |  Istanti: {n_times}  |  Righe totali: {len(df_out)}")
+    print(f"  Posizioni effettive: {sorted(df_out['x'].unique().tolist())}")
+    print(f"  Intervallo T: [{df_out['T'].min():.2f}, {df_out['T'].max():.2f}] °C")
+
+
+if __name__ == "__main__":
+    main()