More sensor input, especially at different frequencies which are going to have different characteristics and so different capabilities and limitations regarding

More sensor input, especially at different frequencies which are going to have different characteristics and so different capabilities and limitations regarding what they detect, is certainly going to provide more data. Since LIDAR is on the order of $1000 per vehicle now, it shouldn't really be a cost concern, but the engineering to incorporate and properly merge that additional sensor input is not insignificant. Tesla has certainly been aggressive in its "FSD" market push, and the quantity of accidents absolutely should be a concern. If the NHTSA determines that the system as-is is sufficient (that is, at least better than the average), then they should say so - but the reverse holds true too.