AI Voice Assistant Interaction Limitations in Smart Homes and Optimization Strategies

Xiaoyu Zheng

doi:10.54254/2755-2721/2025.GL28757

Applied and Computational EngineeringOpen access

AI Voice Assistant Interaction Limitations in Smart Homes and Optimization Strategies

Research Article

Open Access

AI Voice Assistant Interaction Limitations in Smart Homes and Optimization Strategies

Xiaoyu Zheng ^1*

¹ Beijing Open University International Curriculum Centre

^*Corresponding author: zheng.xiaoyu2022@outlook.com

Published on 28 October 2025

ACE Vol.201

ISSN (Print): 2755-273X

ISSN (Online): 2755-2721

ISBN (Print): 978-1-80590-493-9

ISBN (Online): 978-1-80590-494-6

Download Cover

Abstract

AI voice assistants are the "center" of smart homes. However, under the constraint of continuous interaction, they cannot be trusted in our real life. To answer this question, this research first identifies and studies two technical barriers—interference from the environment's noise and conversation gaps between multiple turns. Evidence from previous research shows that high noise interfered with 72% accuracy when speaking; users were frustrated at a 68% rate due to a missing context when talking for an extended period. Next-gen solutions like adaptive noise cancellation technologies and contextual-memory frameworks were analyzed to show that systematic intervention would result in observable gains in a product's performance by enhancing overall results. The proposed dynamic suppression technology can suppress noise via a speech stream (i.e., microphone signals). This research reduces errors up to 30%. An improved situational memory retains 87% accuracy for remembering information during 40-minute-long interactions. Both studies demonstrate how important new solutions are, thus opening doors toward more resilient and user-friendly voice interaction paradigms fit-for-purpose for today's connected lifestyles. Those studies provide not only interesting findings but also some useful suggestions for theories regarding human-computer interaction or concrete takeaways towards understanding smart home applications better.

Keywords:

Smart Home, Voice Assistant, Context Disconnection, Noise Sensitivity, User Experience

View PDF

References

[1]. Edu, J.S., Such, J.M., Suarez-Tangil, G. (2020) Smart Home Personal Assistants: Security and Privacy Challenges. ACM Computing Surveys, 53(6): Article 117.

[2]. Baidu AI Research. (2021) Technical Barriers in Smart Home Voice Interaction. Industry White Paper.

[3]. China Smart Home Industry Alliance. (2020) Consumer Behavior Analysis Report. Securitas Automation Press.

[4]. Li, X., & Wang, Y. (2020) Experience Design of Voice-Controlled Cleaning Robots. Packaging Engineering, 41(18): 210-217.

[5]. Mistral AI. (2025) Voxtral: Multi-modal Speech Understanding Framework. Technical Report v3.2.

[6]. Runhe Tech. (2025) Dynamic Noise Suppression in Edge Computing Environments. Proceedings of WAIC 2025.

[7]. Zhang, L., et al. (2025) Full-Duplex Dialogue Systems for Context Inheritance. AI Industry Applications, 8(3): 45-59.

[8]. Chen, K. (2016) Human-Machine Information Exchange in Smart Homes. ZOL Technology Review.

[9]. Soul AI Lab. (2025) Autonomous Dialogue Rhythm Control. Journal of Affective Computing, 12(4): 301-315.

[10]. AutoNavi Technology. (2025) Spatial Intelligence Agent Design. Technical Whitepaper.

[11]. Josh, R., Sam, G.V., Ramesh, T., & Balam, K.S. (2025) Optimized MOT with CF-GNN for edge computing. Engineering Applications of Artificial Intelligence, 98: 104215.