Model Predictive Control of Active Brownian Particles

Control of active matter systems offers significant potential for various applications, such as directed self-assembly and microactuation. In this study, we propose a feedback control framework for controlling noninteracting Active Brownian Particles (ABPs) at the population level. To achieve control, we adopt Model Predictive Control (MPC), an optimization-based approach that tunes model inputs to achieve desired objectives. The model governing our MPC scheme is based on the Smoluchowski equation, with additional terms accounting for self-propulsion and an actuated external field that influences particle orientations. Leveraging this setup, we apply our feedback control framework to a Brownian dynamics simulation of ideal ABPs. The outcomes of our control simulations demonstrate the versatility and effectiveness of the proposed framework. Notably, our controller successfully achieves multiple objectives, including particle tracking of a specified target, splitting of the particle population into distinct groups, and control over the average particle velocity. These results highlight the potential of MPC in steering active matter systems and pave the way for exciting new opportunities in the field.