Efficient Dual-Type Reference Governor for Model Predictive Control With Vehicle System Applications
提出一种双类型参考调节器，解决模型预测控制中预测时域内参考不可达问题，通过初始和终端两种调节器设计确保跟踪性能与可行性，并在自适应巡航和换道控制仿真中验证了其低计算负担优势。
In this work, a novel dual-type reference governor (DRG) is proposed to handle the unreachable reference within the prediction horizon for the model predictive control (MPC). The proposed method, featuring two types of reference governor (RG) designs, ensures efficient reference tracking and MPC feasibility throughout the entire process. The first type, the initial RG, is specifically designed to adjust references when the controller receives a new target reference. This addresses a gap in existing research on command/feasibility governors (FGs), which overlooks this scenario. The second type, named terminal RG, ensures the convergence of auxiliary references to the current target reference. By introducing control variables, it expands the feasible set of auxiliary references compared to existing methods, thereby further accelerating the convergence speed. The recursive feasibility, finite-time convergence, and asymptotic stability of the combined DRG+MPC closed-loop system with the proposed algorithm are demonstrated in the article. To validate the effectiveness of the proposed algorithm, numerical simulations were conducted in two vehicle application scenarios: adaptive cruise control (ACC) and lane changing control (LCC), using their respective vehicle models. The results indicate that, compared to existing methods such as classical MPC, tracking MPC (TMPC), and FG+MPC, the proposed approach guarantees both reference tracking performance and low computational burden.