Archives
International Journal of Zoology and Applied Biosciences Research Article
Accelerometer-assisted human gesture detection using arm architecture
Thangasubha T, Nirmala B, Lavanya R, Swetha M and Geetha C
Year : 2025 | Pages: 547-550
Received on: 27/09/2025
Revised on: 25/10/2025
Accepted on: 26/11/2025
Published on: 01/12/2025
-
Thangasubha T, Nirmala B, Lavanya R, Swetha M and Geetha C( 2025).
Accelerometer-assisted human gesture detection using arm architecture
. International Journal of Zoology and Applied Biosciences, 10( 6), 547-550.
-
click to view the cite format
Abstract
This study presents an accelerometer-assisted human gesture detection system designed using ARM microcontroller architecture for real-time embedded applications. The proposed system leverages a tri-axial accelerometer (ADXL335/ADXL345) integrated with an ARM Cortex-M based controller to capture hand motion, extract dynamic features, and classify gesture patterns. A threshold-based pattern recognition algorithm is implemented to identify commonly used gestures such as tilt, shake, directional movements, and predefined control actions. The system demonstrates high responsiveness, low power consumption, and improved accuracy through digital filtering and adaptive calibration. Experimental evaluation indicates a gesture recognition accuracy of 94.2%, outperforming traditional 8-bit microcontroller-based systems by 19% in terms of processing speed and 23% in noise immunity. The findings emphasize the potential of ARM-based embedded gesture interfaces in human–machine interaction, IoT, assistive technologies, and smart automation.
Keywords
Gesture Recognition, Accelerometer Sensor, ARM Microcontroller, Embedded System.
-
Full Article PDF (
11)
- View HTML Article
Copy Rights
© The Author(s) 2025. This article is published by International Journal of Zoology and Applied Biosciences under the terms of the Creative Commons Attribution 4.0 International License (creativecommons.org), which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
