Virtual sensor fusion for high precision control

High performance control requires high loop gain and large control bandwidth. However, the spurious resonances at the higher frequencies can limit the performance of such type of systems. This drawback can be overcome by using sensor fusion technique. In sensor fusion, two or more sensors are combined in synergy such that good performance is achieved at lower frequencies while ensuring robustness of the system at higher frequencies. This paper presents a new technique, termed as virtual sensor fusion, in which only one of the sensors is physically installed on the system while the other sensor is simulated virtually. The virtual sensor is selected based on desired high frequency response. The effectiveness of the proposed technique is demonstrated numerically for a case of active seismic isolation. A robustness analysis of virtual sensor fusion is also carried out in order to study its stability in the presence of spurious resonances. Finally, the technique is experimentally verified on active isolation of pendulum system from ground motion. The results obtained demonstrate good isolation performance at lower frequencies and robustness to plant uncertainties (spurious resonances) at higher frequencies. This technique can be effectively used for high precision control of sensitive instruments. (C) 2020 Elsevier Ltd. All rights reserved.

Automated summary

This paper introduces a new virtual sensor fusion technique to improve high performance control and robustness of systems by physically installing one sensor and simulating another sensor virtually. The technique demonstrates good results in active seismic isolation experiments, showing good isolation performance at lower frequencies and robustness at higher frequencies.