Particle identification (PID) efficiency plays a pivotal role in hadron spectroscopy study at hadron colliders, as demonstrated by the substantial increase in efficiency of $\Lambda_b\to J/\psi pK$ in LHCb RUN II compared to RUN I. By introducing PID variables, the signal yield has increased tenfold; five times due to luminosity and an additional two times due to enhanced PID efficiency. In the upcoming LHCb upgrade, the higher luminosity presents challenges in maintaining or improving the performance of the RICH detectors. To tackle these challenges, a novel technique incorporating an advanced detector description and pattern recognition using artificial neural networks has been developed. In this talk, we will explore the importance of PID efficiency in physics analysis and delve into the innovative solutions employed to optimize the performance of the RICH detectors under increased luminosity conditions.