In this work, an adaptive feedback linearized model predictive control (AFLMPC) scheme is proposed to compensate system uncertainty for a class of nonlinear multi-input multi-output system. Initially, a feedback linearization technique is used to transform the nonlinear dynamics into an exact linear model, thereafter, a model predictive control scheme is designed to obtain the desired tracking performance. A suitable constraint mapping algorithm has been developed to map input constraints to the new virtual input of the proposed control scheme. The proposed control scheme utilizes multiple estimation model and the concept of second-level adaptation technique Pandey et al. (2014) to handle the parametric uncertainty in real-time. Hence, the adaptive term in the control scheme is used to counteract the effect of model uncertainties and parameter adaptation errors. The effectiveness of the proposed AFLMPC control algorithm has been verified successfully in simulation as well as the experimental setup of the TRMS model. The unavailable states of the nonlinear system have been estimated using an extended Kalman filter based state observer. The performance of the proposed control algorithm has been compared with other existing nonlinear control techniques in simulation and experimental validation.

中文翻译：

---

非线性多输入多输出系统的自适应反馈线性化模型预测控制器设计

在这项工作中，提出了一种自适应反馈线性化模型预测控制（AFLMPC）方案来补偿一类非线性多输入多输出系统的系统不确定性。最初，使用反馈线性化技术将非线性动力学转换为精确的线性模型，然后设计模型预测控制方案以获得所需的跟踪性能。已经开发了一种合适的约束映射算法来将输入约束映射到所提出的控制方案的新虚拟输入。所提出的控制方案利用多重估计模型和第二级适应技术 Pandey 等人的概念。(2014) 实时处理参数不确定性。因此，控制方案中的自适应项用于抵消模型不确定性和参数自适应误差的影响。所提出的 AFLMPC 控制算法的有效性已在 TRMS 模型的仿真和实验设置中得到成功验证。非线性系统的不可用状态已使用基于扩展卡尔曼滤波器的状态观察器进行估计。所提出的控制算法的性能已与其他现有的非线性控制技术在仿真和实验验证中进行了比较。