Fuzzy logic in your ear

Hearing aids process sound, but the output that you want to hear depends on the situation and on your own abilities and preferences. Each situation presents a new set of requirements, and each person’s hearing is different. The whole process is uncertain. Fuzzy logic can help with that.

In traditional logic, values are either true or false. Dr Faiyaz Doctor specialises in fuzzy logic. He gave a talk to the COG-MHEAR teams in which he explained that in real world situations there is often uncertainty, and it is difficult to work with the complexity of this. Fuzzy logic is an established methodology to deal with imprecise and uncertain data. It deals with reasoning by using logic that is approximate rather than fixed and exact (crisp). This enables a system to estimate, rather than having to produce exact responses. Fuzzy systems have been used for decision making and some control applications, for example in the development of automated anaesthesia control to optimise the dosage being delivered to a patient.

Imagine driving a car around a bend in the road. As your eyes perceive the changes in the curvature of the road ahead (inputs) you adjust your steering whilst simultaneously breaking or accelerating (outputs) to maintain a safe speed and distance from the curb. This can be modelled as a fuzzy system in which decision or control responses are gradually adjusted as conditions and circumstances change in an inexact way.

Hearing assistive technology needs to be able to handle uncertainties in human speech, which incorporates sound, vision, movement, and context, as well as rapid changes in environmental conditions such as transient noise and reverberation. Neural networks can be used to analyse this information. These mimic the way in which the human brain works. Hybrid systems can combine neural networks with fuzzy logic, giving a method of working out what is essential in the sensing and processing chain of events that happens within a hearing aid. A key area in which this type of approach can be applied is in the development of context aware and adaptive assistive technologies such as speech enhancement systems for hearing aids.