Novel neural network adaptive controller structure for a resonant circuit

This paper addresses the problem of tuning, adaptively, a parallel RLC resonant circuit and introduces a new neural network-based scheme under which the circuit is adaptively tuned. The proposed neural network controller is, in a sense, trainingless as no pre-training phase is required. Furthermore, it has the ability to discover himself, adaptively, the suitable control law needed to maintain the circuit well-tuned at all time.