A binaural single object sensor as a mobility aid for the blind

This thesis describes a feasibility study of a binaural single object sensor as a mobility aid for the blind. Firstly, arguments are advanced for a mobility aid presenting the spatial information in a manner similar to the natural echo­ location phenomenon that is experienced by many blind people. It is further argued that the aid should preferably detect only the nearest object in its illuminating field. A pulsed single object sensor (SOS) is then designed with rectified echo pulses as the auditory stimulus. Using the sing-around principle as the mechanism for switching the transmitting pulses, the distance to an object is coded as inversely proportional to the rate of repetition of the audio pulses. Auditory lateralization with an experimental model of the SOS shows that interaural amplitude difference is the most suitable lateralization mechanism. The presentation of dichotic sounds such that they provide the "out-of-the-head" sensation is also studied. This involves theoretical and experimental examination of the role of pinna in auditory localization. It was found that the pinna is the most important factor in externalizing a sound image. An analogue cross correlator with spatial resolution in the order of 3-4mm was constructed to measure the pinna transfer function. It was found that, while the pinna produces 3 groups of distinct delayed replicas of the original sound, the variation of the amplitude of the replicas with direction was too complex to be realized in a portable aid. This, together with a comparison of locomotion control under headphones and free-field listening condition, tends to justify the use of lateralization as the mechanism to determine the direction to an object in the SOS. Laboratory studies of spatial perception with the aid were carried out, with particular reference to the perception of change in the distance to an object, the use of loudness as an extra distance cue, the object recognition capability, and the control of locomotion in various locomotive tasks. Locomotion performances were also compared with Kay's Binaural Sensory Aid and unaided performances. The problems involved in evaluating a sensory aid are discussed, and an analysis of mobility activities is provided. The possible usefulness of the aid in these activities is indicated