Newbie to the embedded systems here ;) We have a Cortex M4 with DSP.
We've got a linear system of equations to solve. Say, we have 8 variables to solve, so vectorize it into AX=B.
So it's required that A is (n,8) sized, X=(8,1), and B=(n,1)(with n>=8).
If A is a square matrix with a rank of 8, then I could've easily gotten X by inv(A).B, where arm_mat_inverse_f32.c and arm_mat_mult_f32.c could be directly invoked.
But, if we got an A matrix with n>8 and ranked 8, how to solve the X? Is there a way to automatically do something like elementary row transformations?
Appreciate everything. Please bear with my noobness ;)