New input-output matrices based frequency interval balanced truncation

The existing frequency-limited interval-based model reduction techniques can generate unstable models for a stable continuous-time linear system. Further, some current methods resolve the instability problem, yielding a reduced model with a significant approximation error. The suggested approach is based on novel limited-interval Gramians, ensuring stability and producing a low approximation error. The simulated numerical examples illustrate the method's efficacy using singular value plots and infinity norm error.