Adaptive Neural Network Output-Constraint Control for a Variable-Length Rotary Arm With Input Backlash Nonlinearity

This article studies the problem of deformation reduction and attitude tracking for a rotated and extended flexible crane arm with input backlash-saturation and output asymmetrical constraint. By employing Halmilton's principle, the arm system model is formulated by a set of partial and ordinary differential equations (ODEs). Given the modeling inaccuracy, a radial neural network (RNN) is used to approximate system parameters. To better design the controllers, the backstepping technique is applied to the control design. For input nonlinearities with backlash and saturation, we reversely transform them as an asymmetric saturation constraint via a virtual input. A barrier Lyapunov function (BLF) containing logarithmic terms is constructed to guarantee the asymmetric output constraints and the uniformly ultimate boundedness and stability of the arm system are proved. Finally, to testify the effectiveness of the proposed controllers, numerical simulations are carried out, and responding simulation diagrams are displayed.