The LISA DFACS: Model Predictive Control design for the test mass release phase

Abstract

This paper presents a Model Predictive Control (MPC) design for the test mass release phase of the LISA space mission. LISA is a gravitational wave observatory consisting of a triangular constellation of three spacecraft. The gravitational waves are detected by measuring the relative distance between free falling test masses by means of a laser interferometer. Each test mass is a cubic body located inside an electrostatic suspension that is initially locked by a clamp mechanism. Once the plungers are retracted, the test masses are released with high initial offsets and velocities. To detect the gravitational waves, each test mass must be accurately positioned at the cage centre and its attitude must be aligned with the local cage frame. However, the low actuation authority of the electrostatic suspension along with the critical initial conditions, make the attitude and translation control a difficult task. MPC is a suitable technique for this application because it can systematically account for command saturations, state constraints and can provide optimal (or sub-optimal) control inputs by solving an optimization problem online. In this paper, an MPC controller is designed and validated by means of Monte Carlo simulations, achieving satisfactory results.