Research Article
Bilateral Vibrotactile Feedback for Accurate Lateralization in Hearing Instrument Body Area Networks
@INPROCEEDINGS{10.4108/icst.bodynets.2011.246910, author={Bernd Tessendorf and Peter Derleth and Manuela Feilner and Tobias Gr\aa{}mer and Daniel Roggen and Michael Spuhler and Thomas Stiefmeier and Gerhard Tr\o{}ster}, title={Bilateral Vibrotactile Feedback for Accurate Lateralization in Hearing Instrument Body Area Networks}, proceedings={6th International ICST Conference on Body Area Networks}, publisher={ICST}, proceedings_a={BODYNETS}, year={2012}, month={6}, keywords={hearing instrument vibrotactile feedback user study}, doi={10.4108/icst.bodynets.2011.246910} }- Bernd Tessendorf
Peter Derleth
Manuela Feilner
Tobias Grämer
Daniel Roggen
Michael Spuhler
Thomas Stiefmeier
Gerhard Tröster
Year: 2012
Bilateral Vibrotactile Feedback for Accurate Lateralization in Hearing Instrument Body Area Networks
BODYNETS
ICST
DOI: 10.4108/icst.bodynets.2011.246910
Abstract
Hearing Instruments (HIs) have emerged as true body area networks (HI-BANs). They connect wirelessly to accessories such as remote controls and Bluetooth devices. Multimodal sensor data from a HI-BAN is a way to adapt the HI behavior to the user's current hearing situation. As a potential future HI-BAN component we investigate bilateral vibrotactile feedback to support localization of sound sources. As a foundation we investigate different feedback and vibration patterns. We implemented two approaches for encoding lateral target angles: Continuous Guidance Feedback (CGF) and 6 variants with evolving complexity of Quantized Absolute Heading (QAH). In a user study with 16 normal hearing participants we evaluate lateralization error and response time. For QAH results show a trade off between the minimal quantization error due to the encoding and the number of user errors due to misinterpretation of presented patterns. Results also show a trade off between response time and minimum lateralization error. Choosing the most suitable bilateral vibrotactile encoding schemes is application-specific: For QAH a minimal average lateralization error of 27 degree was achieved with eight 45 degree-segments. A minimal average user response time of 900ms was achieved with four 45 degree-segments. CGF guides the user within a given tolerance margin to the target at the cost of higher response time. For complex encoding schemes the performance is person-specific.
