Predefined-time self-structuring neural network H∞ cooperative control for multirobot systems with prescribed performance and input quantization

In this paper, a predefined-time adaptive command filter H_∞ controller with a self-adjusting performance function is proposed for multirobot systems. It guarantees that the tracking error meets the desired performance requirements and solves the vulnerability problem that arises in traditional prescribed performance. First, an asymmetric tan-type barrier Lyapunov function is introduced to establish asymmetric barrier constraints under input saturation and input quantization. Second, a prescribed performance with self-adjustment is introduced in the asymmetric tan-type barrier Lyapunov function, which limits the position error and changes the performance envelope based on its state. Third, a predefined-time adaptive command filter is introduced to address the "complexity explosion" issue and improve the convergence speed. Fourth, a predefined-time self-structuring neural network is introduced to fit the model uncertainty and time-varying disturbances, and a predefined-time H_∞ control strategy is designed to address the strong sudden disturbances. Finally, some simulation examples are provided to test the validity of the above algorithms.