Charge-Balanced Control of Linear Systems: Controllability, Stabilizability, and Periodic Regulation

Abstract

Electrical neuromodulation is a therapeutic approach that regulates neural activity by delivering electrical currents to brain tissue to treat brain disorders, such as epilepsy and Parkinson's disease. A critical safety requirement is charge balance, which ensures zero net injected charge over each stimulation cycle. Despite its clinical importance, the implications of charge-balanced stimulation for system dynamics and control remain poorly understood. This article investigates the controllability, stabilizability, and regulation of discrete-time linear systems under charge-balanced input constraints, motivated by applications in electrical neuromodulation. We introduce novel notions of charge-balanced controllability and stabilizability and establish necessary and sufficient conditions for both properties. We further demonstrate that nonzero steady states cannot be maintained under charge-balanced inputs, motivating the introduction of charge-balanced periodic orbits as alternative control targets. We characterize the conditions under which convergence to such orbits is achievable and develop optimal feedback controllers to stabilize the origin and track optimal charge-balanced periodic orbits. Finally, we validate the proposed control framework using a Hodgkin-Huxley neuron model, demonstrating successful suppression of tonic spiking and highlighting the potential of our approach for controlling nonlinear dynamical systems.