Adaptive Complexity Model Predictive Control

Abstract--This work introduces a formulation of model predictive control (MPC) which adaptively reasons about the complexity of the model based on the task while maintaining feasibility and stability guarantees. Existing MPC implementations often handle computational complexity by shortening prediction horizons or simplifying models, both of which can result in instability. Inspired by related approaches in behavioral economics, motion planning, and biomechanics, our method solves MPC problems with a simple model for dynamics and constraints over regions of the horizon where such a model is feasible and a complex model where it is not. The approach leverages an interleaving of planning and execution to iteratively identify these regions, which can be safely simplified if they satisfy an exact template/anchor relationship. "[the complex, slow system] is activated when an event is detected that violates the model of the world that [the simple, I. Extending this concept to the field of motion planning yields meta-planning methods which As demand for robotic systems increases in industries change their structure to leverage simple, fast models where like environmental monitoring, industrial inspection, disaster possible and complex, slow ones where the simple model is recovery, and material handling [1-3], so too has the need for inaccurate [5,6]. However, it is not well understood under what motion planning and control algorithms that efficiently handle exact conditions a given dynamical system may leverage a the complexity of their dynamics and constraints.