Optimisation of a digital autodetector for linear and logarithmic radar video

Procedures are described which lead to empirical optimisation of a digital radar autodetector by adjustment of quantisation and integration parameters. Receiver nonlinearity is dealt with exactly by appropriate transformation of quantisation thresholds. The statistical modifications of quantised data introduced by a digital leaky integrator are treated by a convolution process. Performance is calculated for logarithmic and linear receivers with Swerling 0 and Swerling 2 target models. It is found that judicious adjustment of quantisation thresholds has a marked effect on integration efficiency, particularly in the logarithmic case.