states.kalman.project.solve_psd
def solve_psd(cov: torch.Tensor, rhs: torch.Tensor, label: str) -> torch.TensorSolve ``cov @ x = rhs`` for a (batch of) PSD covariance(s).
Parameters
| Name | Type | Description |
|---|---|---|
| cov | torch.Tensor | ``(..., D, D)`` positive-semi-definite system matrix. |
| rhs | torch.Tensor | ``(..., D, K)`` right-hand side. |
| label | str | Identifier embedded in the error message on failure. |
Returns
torch.Tensor — ``(..., D, K)`` solution tensor.
Raises
- RuntimeError — If the solve produces non-finite values, indicating a singular covariance or PSD-cone drift. The caller should re-tune process/measurement noise or the initial covariance.
Source: unitrack/states/kalman/project.py:57