In contrast to design tools for graphics and audio generation from text prompts, haptic design tools lag behind due to challenges in constructing large-scale, high-quality datasets including vibrations and text descriptions. To address this gap, we propose ChatHAP, a conversational haptic system for designing vibrations. ChatHAP integrates various haptic design approaches using a large language model, including generating vibrations using signal parameters, navigating through libraries, and modifying existing vibrations. To further improve vibration navigation, we present an algorithm that adaptively learns user preferences for vibration features. A user study with novices (n=20) demonstrated that ChatHAP can serve as a practical design tool, and the proposed algorithm significantly reduced task completion time (38%), prompt quantity (25%), and verbosity (36%). The study found ChatHAP easy-to-use and identified requirements for chat-based haptic design as well as features for further improvement. Finally, we present key findings with ChatHAP and discuss implications for future work.
https://dl.acm.org/doi/10.1145/3706598.3713441
The ACM CHI Conference on Human Factors in Computing Systems (https://chi2025.acm.org/)