This paper deals with the problem of detecting and classifying a radar signal (target signal or jamming) in the presence of Gaussian disturbance. The jamming signal is modelled as belonging to a cone whose centre vector is the true target steering vector. Two different detection/ classification approaches are analysed (the well-known ACE and GLRT approaches), yielding both to a two block device. The performance of the two decision algorithms are evaluated analytically, where possible, and by Monte Carlo simulation, otherwise. Some numerical results are reported along with some concluding remarks. (5 pages)