Hybrid state machine model for fast model predictive control: application to path tracking

摘要: Cyber-Physical Systems (CPS) are composed of computing devices interacting with physical systems. Model-based design is a powerful methodology in CPS the implementation control For instance, Model Predictive Control (MPC) typically implemented applications, e.g., path tracking autonomous vehicles. MPC deploys model to estimate behavior system at future time instants for specific horizon. Ordinary Differential Equations (ODE) most commonly used models emulate continuous-time (non-)linear dynamical A complex may comprise thousands ODEs which pose scalability, performance and power consumption challenges. One approach address these complexity challenges frameworks that automate development model-to-model transformation. In this paper, we introduce generation framework transform ODE Hybrid Harmonic Equivalent State (HES) Machine equivalents. Moreover, tuning parameters introduced reconfigure adjust its accuracy from coarse-grained critical situations fine-grained scenarios safety paramount. learning techniques applied adopt run-time applications. We conduct experiments on closed-loop using vehicle dynamics model. analyze when applying our HES The proposed compared state-of-the-art ODE-based models, terms execution accuracy. Our experimental results show 32% reduction return 0.8% loss