Learning user volume control preferences in hearing aids

Abstract

The advancements in the learning of user control preferences to minimize the amount of adjustments being made to the hearing aids over time are discussed. The concept of self-learning hearing aids address the issues of fine tuning and optimal adjustment by learning the settings a user prefers in environments that are presented on a daily basis. The learning algorithm learn different user behaviors and behaviors were simulated to supplement any real data. The performance error measures are the average over all phases in a user profile of the difference between the learned value for a phase. The parameters for the three adaptive algorithms are tuned to the best values for the profile and compared to the optimum fixed time constant for the profile. The optimum parameters of hearing aids perform as good as the best fixed time constant on a profile-by-profile basis of a user control.