Electrotactile feedback is a promising modality for wearable haptics because it can deliver

tactile cues through thin, low-power interfaces. However, its wider adoption is limited

by a persistent quality of sensation gap: evoked percepts are often described as artificial,

uncomfortable, or inconsistent across users and contexts. This thesis investigates how selected

perceptual qualities of non-invasive fingertip electrotactile feedback can be improved

through biologically informed stimulation design, with particular emphasis on increasing

pressure-like sensations while reducing irritation within practical wearable limits.

To address this challenge, two bio-inspired encoding frameworks are proposed. The

first is a model-driven framework that uses TouchSim-derived afferent activity to map

spatial and temporal characteristics of natural touch onto electrode configuration and

stimulation profiles. The second is an evidence-driven framework based on literatureinformed

burst stimulation, designed to improve neural recruitment through structured

temporal patterning. A coupled computational modelling pipeline was also developed,

combining finite-element modelling of fingertip volume conduction with neural response

modelling. This was used to constrain the practical design space, support the electrode

configuration, and identify an effective operating region for parameters such as pulse width

before human testing.

The resulting stimulation patterns were evaluated in subjective experiments with 22

participants under three progressively richer perceptual contexts: tactile-only, tactile with

visual cues, and tactile with visual cues plus active movement. Sensation quality was

assessed primarily by the type of sensation reported, especially the balance between pressure

and irritation, with naturalness treated as a secondary perceptual indicator. Mixedeffects

modelling showed that stimulation pattern significantly affected both pressure

(χ² = 47.87, p < 6.9 × 10⁻⁶) and irritation (χ² = 186.12, p < 2.2 × 10⁻¹⁶) in pooled

analyses, while naturalness showed a weaker overall effect (χ² = 24.98, p = 0.023). Pressure

probability increased across scenarios from 0.43 in the tactile-only condition to 0.57

with visual cues and 0.71 with visual cues plus movement, whereas irritation remained

comparatively stable (0.45, 0.52, and 0.48, respectively).

Across analyses, biologically inspired waveform design produced a more favourable

pressure–irritation trade-off than less structured alternatives. Co-modulated bio-inspired

patterns emerged as the most robust overall, with the low-frequency co-modulated variant

providing the strongest balance between comfort and pressure saliency across scenarios.

Overall, this thesis shows that selected aspects of electrotactile sensation quality can be

improved through biologically grounded waveform design supported by computational

modelling and evaluation under progressively richer perceptual contexts.