For monostatic radar/unidirectional antenna pattern/scattering reflection, we need to have different pattern for TX & RX?

[mlouielu]

Consider the following monostatic radar/unidirectional antenna pattern/vertical polarization/scattering reflection case:

If we ignore the offset (y=0.05) between TX/RX, then the incident and scatter vector will be ki=[1, 0, 0], ko=[-1, 0, 0], as shown in the figure.

In field_calculator.py, we calculate the receive antenna pattern as the following:
https://github.com/NVlabs/sionna-rt/blob/f3a944cbb7b4945975fad21353413b3b7c8a477e/src/sionna/rt/path_solvers/field_calculator.py#L326-L330

        # Receive antenna pattern
        tgt_patterns = [antenna_pattern_to_world_implicit(tgt_antenna_pattern,
                                                          tgt_to_world,
                                                          -ki_world,
                                                          direction="in")
                        for tgt_antenna_pattern in tgt_antenna_patterns]

Consider the following unidirectional pattern:

def v_unidirectional_pattern_v2(theta: mi.Float, phi: mi.Float) -> mi.Complex2f:
    """Simplified version 2 - more combined."""
    h_width_rad = dr.deg2rad(20.0)
    e_width_rad = dr.deg2rad(10.0)

    pattern = (dr.sin(theta) * dr.exp(-2.77 * ((theta - dr.pi / 2.0) / e_width_rad) ** 2) * # E-plane
               dr.exp(-2.77 * (phi / h_width_rad) ** 2))                             # H-plane

    return mi.Complex2f(dr.sqrt(10.0) * pattern, 0.0)

Then, we calculate the antenna_pattern_to_world_implicit(), which will then yield 0 for target pattern:

to_world = np.eye(3)
k_world = -ki   # ki = [1, 0, 0], -ki = [-1, 0, 0]

to_local = to_world.T
k_local = mi.Vector3f(to_local @ k_world)

# theta_local = 90deg, phi_local =180deg
theta_local, phi_local = theta_phi_from_unit_vec(k_local)

# c_theta =  c_phi = 0
c_theta, c_phi = pattern(theta_local, phi_local)

So, even if I put RX in front of the TX (rx=[0.8, 0, 0]), and only enable LOS, which still will yield ki=[1,0,0], ko=[-1,0,0], and give us c_theta = c_phi = 0.

---

Are my observations correct? Are there any assumption I made it wrong? And is it safe to just mirror the TX antenna pattern and assign to RX antenna? Thanks!

[faycalaa]

Hi @mlouielu ,

If you place the RX in front of the TX along the X-axis, and assuming that both radio devices have an orientation of $$(0,0,0)$$, then you should get zero gain for the LoS path with the antenna pattern you show.