Sensing

metasurface_py.sensing

Sensing models: radar, localization, and ISAC objectives.

JointCommsSensingObjective dataclass

Weighted combination of comms and sensing objectives.

objective = alpha * comms_obj + (1-alpha) * sensing_obj

Parameters:

Name	Type	Description	Default
comms_objective	Any	Communication objective callable.	required
sensing_objective	Any	Sensing objective callable.	required
alpha	float	Weight for comms (0 to 1). Default 0.5.	0.5

__call__(state, surface, freq, **kwargs)

Evaluate weighted joint objective.

MaxDetectionSNRObjective dataclass

Maximize radar detection SNR for a point target.

Returns negative SNR (dB) for minimization.

Parameters:

Name	Type	Description	Default
target_pos	Position3D	Target position.	required
tx_power	float	Transmit power [W].	1.0
noise_power	float	Noise power [W].	1e-12
target_rcs	float	Target RCS [m^2].	1.0

__call__(state, surface, freq, **kwargs)

Evaluate: returns negative detection SNR in dB.

bistatic_rcs(surface, state, freq, theta_inc, phi_inc, angles_obs)

Compute bistatic RCS pattern.

Parameters:

Name	Type	Description	Default
surface	Metasurface	Metasurface object.	required
state	SurfaceState	Surface configuration.	required
freq	float	Frequency [Hz].	required
theta_inc	float	Incident polar angle [rad].	required
phi_inc	float	Incident azimuthal angle [rad].	required
angles_obs	AngleGrid	Observation angles.	required

Returns:

Type	Description
DataArray	Bistatic RCS as xr.DataArray [m^2].

crlb_position(surface, state, target_pos, freq, snr=20.0)

Compute Cramer-Rao Lower Bound on position estimation.

CRLB = diag(J^{-1}) gives minimum variance for each position coordinate.

Parameters:

Name	Type	Description	Default
surface	Metasurface	Metasurface object.	required
state	SurfaceState	Surface configuration.	required
target_pos	Position3D	Target position.	required
freq	float	Frequency [Hz].	required
snr	float	Signal-to-noise ratio (linear).	20.0

Returns:

Type	Description
NDArray[floating[Any]]	(3,) array of CRLB values [m^2] for (x, y, z).

detection_snr(surface, state, target_pos, freq, tx_power=1.0, noise_power=1e-12, target_rcs=1.0)

Compute radar detection SNR for a point target.

Uses simplified radar equation: SNR = Pt * G^2 * lambda^2 * sigma / ((4*pi)^3 * R^4 * Pn)

Parameters:

Name	Type	Description	Default
surface	Metasurface	Metasurface acting as radar aperture.	required
state	SurfaceState	Surface configuration.	required
target_pos	Position3D	Target position.	required
freq	float	Frequency [Hz].	required
tx_power	float	Transmit power [W].	1.0
noise_power	float	Noise power [W].	1e-12
target_rcs	float	Target radar cross-section [m^2].	1.0

Returns:

Type	Description
float	Detection SNR (linear).

fisher_information_matrix(surface, state, target_pos, freq, snr=20.0)

Compute Fisher Information Matrix for position estimation.

Uses a simplified far-field signal model where the FIM depends on the array geometry and element responses.

Parameters:

Name	Type	Description	Default
surface	Metasurface	Metasurface object.	required
state	SurfaceState	Surface configuration.	required
target_pos	Position3D	Target position to localize.	required
freq	float	Frequency [Hz].	required
snr	float	Signal-to-noise ratio (linear).	20.0

Returns:

Type	Description
NDArray[floating[Any]]	3x3 Fisher Information Matrix.

monostatic_rcs(surface, state, freq, angles)

Compute monostatic RCS pattern.

For each observation direction, the RCS is proportional to |AF(theta, phi)|^2 where the illumination and observation directions are the same.

Parameters:

Name	Type	Description	Default
surface	Metasurface	Metasurface object.	required
state	SurfaceState	Surface configuration.	required
freq	float	Frequency [Hz].	required
angles	AngleGrid	Observation angles.	required

Returns:

Type	Description
DataArray	RCS pattern as xr.DataArray [m^2], dims (theta, phi).