ANN-based adaptive motion and posture control of an inverted pendulum with unknown dynamics

In this paper, an artificial neural network (ANN) based control scheme is introduced for the inverted pendulum motion and posture control problem. The adaptive control strategy consists of a Lyapunov stability-based online weights adaptation that provides asymptotic tracking while learning the nonlinear inverted pendulum system's dynamics. Unlike other control strategies, no a priori offline training, weights initialization, or parameters knowledge is required. Experiments for different situations highlight the performance of the proposed controller in compensating for friction nonlinearities, in the form of Coulomb friction. Furthermore, the neural networks inherent parallelism makes them a good candidate for implementation in real-time electromechanical systems.