Lagrange programming neural network for radar waveform design

In this chapter, Lagrange programming neural network (LPNN) that is a constrained optimization solver is exploited to design radar probing waveforms under constraints due to unit-modulus, spectral requirement and/or ambiguity function (AF). In the LPNN, there are two types of neurons to compute the optimum solution, namely, variable and Lagrangian neurons, which are responsible for finding a minimum point of the cost function as well as the solution at an equilibrium point and leading the dynamic trajectory into the feasible region. The local stability conditions of the dynamic neuron model are also analyzed. Simulation results show that the LPNN-based approach is a competitive alternative for waveform design.