Crossref DOIMemberCiteFactoropen accessissn
SJIF Impact Factor 8.659 | ISSN: 2582-3930
IJSREM Logo

International Journal of Scientific Research in Engineering and Management

(Gateway to Global Scholars - An International, Scholarly, Multi-Disciplinary, Open Access Journal)

Peer-Reviewed Journal Designed to Meet UGC's Quality Parameters

HomeCurrent IssuePast IssuesPay OnlineConferencesEditorial BoardAbout UsContact Us
IJSREM

Loading article...

Follow Us

  • Google Scholar
  • ResearchGate
  • Academia.edu
  • Mendeley
  • LinkedIn
  • Facebook
  • Instagram

Journal Policies

  • Peer Review Policy
  • Publication Ethics
  • Copyright Infringement
  • Refund & Cancellation Policy
  • Terms and Conditions
  • Privacy Policy

Journal Identifiers & Licenses

  • ISSN Verified: 2582-3930ISSN Verified: 2582-3930
  • Crossref DOI: 10.55041Crossref DOI: 10.55041
  • SJIF Ranking: 08.659SJIF Ranking: 08.659
  • Licensed under CC BY 4.0Licensed under CC BY 4.0
  • ISO Certification: 9001:2015ISO Certification: 9001:2015
  • Indexed & Cited In: Google ScholarIndexed & Cited In: Google Scholar

Contact Us

International Journal of Scientific Research in Engineering and Management (IJSREM)

#62/1, New No 7, 1st Cross, 2nd Main, Ganganagar, R T Nagar, Bangalore North, Bangalore, Karnataka, India – 560032
[email protected]+91 93911 67991
logos
logos
Working Hours

Mon – Sat: 9:00 AM – 6:00 PM

Sunday: 10:00 AM – 4:00 PM

+91 93911 67991

© 2026 IJSREM – International Journal of Scientific Research in Engineering and Management. This is an open-access journal. Articles are published under the Creative Commons Attribution (CC BY 4.0) License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

DisclaimerPrivacy PolicyTerms and Conditions
ArticlePeer ReviewedOpen AccessPlagiarism Passed
Paper ID:IJSREM68811

TinyML and Edge AI for On-Device Intelligence: A Review of Keyword Spotting and Voice-Controlled Assistive Robots

Stream/Area:

-

Subject:

-

Paper ID:

IJSREM68811

Volume:

10 |

11 |

Nov 2026

Pages:

1-9

Publication Frequency:

Monthly (12 issues per year)

DOI
Digital Object Identifier
10.55041/IJSREM68811

License:International Journal of Scientific Research in Engineering and Management follows the Creative Commons Attribution 4.0 (CC BY 4.0) License, allowing free use and distribution with proper attribution.

Share this article:

Authors

Aishwarya Yuvraj Dhane

Aishwarya Yuvraj Dhane

Department of Electronics & Telecommunication (VLSI & Embedded System)

TSSM's Padmbhushan Vasantdada Patil Institute of Technology, Pune

Download Author Certificate
PDSMK

Prof. Dr. S. M. Kulkarni

Department of Electronics & Telecommunication (VLSI & Embedded System)

TSSM's Padmbhushan Vasantdada Patil Institute of Technology, Pune

Download Co-Author Certificate

Author Services & Downloads

Full ArticlesAcceptance Certificate

Abstract

Voice is a natural control channel for assistive mobile robots. Yet most currently deployed voice interfaces stream audio to cloud-based recognizers, which adds latency, reliance on connectivity, and exposure of privacy. Tiny machine learning (TinyML) offers an alternative in which a quantized neural network performs keyword spotting (KWS) directly on a low-cost microcontroller. This review examines how far recent work from 2024 to 2026 supports the development of a voice-controlled assistive robot based on on-device KWS. We synthesize 40 recent studies covering compact KWS architectures, neural architecture search, quantization and pruning, noise and front-end variability, on-device personalization, speech-driven robotic systems, and safety-aware human-robot interaction. The literature indicates that efficient KWS models now fit within the memory and energy budgets of low-cost microcontrollers, and that personalization and open-set recognition are maturing. Evaluation practice, however, remains uneven: false activations per hour, end-to-end command-to-action latency, and energy per decision are seldom foregrounded, and few systems couple recognition confidence with motion safety logic. Building on these observations, we present a reference architecture for an ESP32-based assistive robot with six functional blocks, a confidence-aware command arbitration layer in which an emergency stop and obstacle checks take precedence over recognized speech, and a six-experiment evaluation protocol. Five distinct research gaps are identified, including speaker-independent validation, the effect of quantization on rare-command recall, and fail-safe behavior under sensor or model faults. The review rests on reported objectives and findings rather than reproduced experiments. It offers a structured basis for designing, benchmarking, and safely deploying offline voice-controlled assistive robots on resource-constrained hardware.

KEYWORDS

TinyML edge AI keyword spotting voice-controlled robot assistive robotics model quantization ESP32 on-device inference

Cite This Article

Aishwarya Yuvraj Dhane, Prof. Dr. S. M. Kulkarni, "TinyML and Edge AI for On-Device Intelligence: A Review of Keyword Spotting and Voice-Controlled Assistive Robots," International Journal of Scientific Research in Engineering and Management, vol. 10, no. 11, pp. 1-9, Nov 2026. https://doi.org/10.55041/IJSREM68811

Full Article/Conclusion

This review examined 40 studies from 2024 to 2026 to assess how well current TinyML research supports a voice-controlled assistive robot that recognizes commands entirely on a microcontroller. Compact and quantized KWS models, architecture search, open-set rejection, and on-device personalization are advancing quickly, and benchmarking practice is improving. The integration of recognition with obstacle sensing, an emergency stop, and confidence-aware arbitration remains thinly documented. We therefore proposed a six-block reference architecture, an arbitration sequence in which safety conditions precede speech, and a six-experiment protocol that reports false activations, latency, memory, and energy next to accuracy. Five research gaps were identified. Because the reviewed studies were assessed from their reported objectives and findings rather than reproduced, the conclusions should be read as a map of the field. Experimental validation of the proposed architecture on a physical prototype is the natural next step.

References

1. Cruces, A.; Jerez, A.; Bandera, J.P.; Bandera, A. Socially Assistive Robots in Smart Environments to Attend Elderly People—A Survey. Applied Sciences 2024, 14, 5287. https://doi.org/10.3390/app14125287. 2. Garrosa, M.; Ceccarelli, M.; Russo, M.; Yang, B. Design of a Mobile Assisting Robot for Blind and Elderly People. Applied Sciences 2025, 15, 10474. https://doi.org/10.3390/app151910474. 3. Hussein, H.H.; Karan, O.; Kurnaz, S. Enhancing Driving Control via Speech Recognition Utilizing Influential Parameters in Deep Learning Techniques. Electronics 2025, 14, 496. https://doi.org/10.3390/electronics14030496. 4. Padmanabha, A.; et al. VoicePilot: Harnessing LLMs as Speech Interfaces for Physically Assistive Robots. arXiv 2024, arXiv:2404.04066. https://doi.org/10.1145/3654777.3676401. 5. Davila, A.; Colan, J.; Hasegawa, Y. Voice Control Interface for Surgical Robot Assistants. arXiv 2024, arXiv:2409.10225. https://doi.org/10.1109/MHS63891.2024.10856297. 6. Di Leo, S.; De Cicco, L.; Mascolo, S. Real-Time Speech-to-Text on Edge: A Prototype System for Ultra-Low Latency Communication with AI-Powered NLP. Information 2025, 16, 685. https://doi.org/10.3390/info16080685. 7. Torkamani, M.J.; Zarin, I. Adaptive Edge-Cloud Inference for Speech-to-Action Systems Using ASR and Large Language Models. arXiv 2025, arXiv:2512.12769. 8. Huang, P.; Ullah, I.; Wei, X.; Ahanger, T.A.; Hassan, N.; Shah, Z.H. Towards Energy-Efficient and Low-Latency Voice-Controlled Smart Homes: A Proposal for Offline Speech Recognition and IoT Integration. arXiv 2025, arXiv:2506.07494. 9. Wang, X.; et al. Empowering Edge Intelligence: A Comprehensive Survey on On-Device AI Models. arXiv 2025, arXiv:2503.06027. https://doi.org/10.1145/3724420. 10. Somvanshi, S.; et al. From Tiny Machine Learning to Tiny Deep Learning: A Survey. arXiv 2025, arXiv:2506.18927. https://doi.org/10.1145/3776588.

Editorial & Review Transparency

Publication Date

11/6/2026

Peer Review Type

Double-Blind Peer Review

Review Rounds: 2

Documents

Review Report (Confidential)
Copyright Form (Office)

Article Metrics & Impact

Article Views

0

Total views since publication

PDF Downloads

7

Total downloads since publication

Geographic Reach

5

Indexing, Archiving & Preservation

Google Scholar
Academia
ResearchGate
Crossref
SSRN
Semantic Scholar
Thomson Reuters
WorldCat
Mendeley
OpenAlex
EBSCO
Zenodo

Digital Preservation: This article is permanently archived on Zenodo and complies with all data preservation standards across major indexing platforms.

Ethics, Compliance & Transparency

Plagiarism Screening

This manuscript has been checked for plagiarism using TurnitinPlagiarismChecker software.

Conflict of Interest

The authors declare no conflict of interest related to this research work.

Funding Information

No funding information available for this research.

Similar & Recommended Articles

Similar Articles by Topic

Related research in the same subject area

Loading similar articles...