i think this is a duplicate : https://math.stackexchange.com/a/32602/427318 seems to give an answer. to explain, if r is perpendicular to b, then it's fairly easy to find the inverse, but any component parallel to b will just disappear due to the bilinearity of the cross product, so there may be an arbitrary nonzero b component (no "a" as it must be perpendicular to both r and b, or 0)
splitting up r into the two components- one parallel to b, and one perpendicular to both a, and b, e.g of the form r= t(bxa)+s(b) and then using the vector triple product identity
$\mathbf{a}\times (\mathbf{b}\times \mathbf{c}) = (\mathbf{a}\cdot\mathbf{c})\mathbf{b} -(\mathbf{a}\cdot\mathbf{b}) \mathbf{c}$ + anticommutativity of the cross product to find t should give the answer relatively simply, as most of the terms in the triple product will be zero by orthogonality, thus no need for coefficients
$x$). I'm surprised someone bothered to edit this question without fixing that. – David K Aug 17 '20 at 14:42