Forum Discussion

Altera_Forum's avatar
Altera_Forum
Icon for Honored Contributor rankHonored Contributor
16 years ago

Efficient Matrix Inversion

Hello,

I need to implement Kalman Filter in one FPGA.

The botleneck for the implementation of the Kalman filter is the Matrix inversion.

Someone can suggest me which algorithm would be better according with the architecture of usual FPGAs.

Thank you.

Diego Botero.

3 Replies

  • Altera_Forum's avatar
    Altera_Forum
    Icon for Honored Contributor rankHonored Contributor
    16 years ago

    it's strongly ligated to your matrix size. For small matrices (2x2 or 3x3), you could use the determinant and the conjugate matrix. This approach uses multipliers and a divider circuit. For bigger matrices it's recomended a QRD-RLS decomposition.

  • Altera_Forum's avatar
    Altera_Forum
    Icon for Honored Contributor rankHonored Contributor
    13 years ago

    why do you need to do it by hardware, I've done it in software in a NIOS2