Minor fix in CartesianToElevationBearingRangeRate inverse_function.

[gawebb-dstl]

An extra test with a platform to test the fix

[oharrald-Dstl]

Is this model invertible? Inverting range rate back to state space leads to incorrect velocities in some instances. Would it be better to make an inheritance of 'semi-invertible' measurement models, whereby we can extract a subset of state-space elements?