Enabling Vibration-Based Gesture Recognition on Everyday Furniture via Energy-Efficient FPGA Implementation of 1D Convolutional Networks
🤖 AI Summary
Existing vibro-tactile gesture recognition systems rely on high-performance hardware, incur substantial power consumption, and are ill-suited for deployment on resource-constrained edge devices—such as everyday furniture—due to computational and energy limitations.
Method: We propose an end-to-end lightweight solution tailored for low-power FPGAs. It eliminates conventional spectral preprocessing, directly ingesting raw vibration waveforms; employs compact 1D-CNN and 1D-Separable CNN architectures; and integrates hardware-aware neural architecture search, integer quantization, Ping-Pong buffering, and RTL-level automated synthesis.
Contribution/Results: Implemented on an AMD Spartan-7 FPGA, our system achieves 97.0% accuracy with only 1.2 mJ per inference and 6.83 ms latency—over 53× faster than CPU-based acceleration—enabling long-term, real-time multi-user, multi-desk interaction. To the best of our knowledge, this is the first FPGA-native vibro-tactile gesture recognizer that operates without preprocessing while delivering high accuracy and ultra-low power consumption.
Application Category
📝 Abstract
The growing demand for smart home interfaces has increased interest in non-intrusive sensing methods like vibration-based gesture recognition. While prior studies demonstrated feasibility, they often rely on complex preprocessing and large Neural Networks (NNs) requiring costly high-performance hardware, resulting in high energy usage and limited real-world deployability. This study proposes an energy-efficient solution deploying compact NNs on low-power Field-Programmable Gate Arrays (FPGAs) to enable real-time gesture recognition with competitive accuracy. We adopt a series of optimizations: (1) We replace complex spectral preprocessing with raw waveform input, eliminating complex on-board preprocessing while reducing input size by 21x without sacrificing accuracy. (2) We design two lightweight architectures (1D-CNN and 1D-SepCNN) tailored for embedded FPGAs, reducing parameters from 369 million to as few as 216 while maintaining comparable accuracy. (3) With integer-only quantization and automated RTL generation, we achieve seamless FPGA deployment. A ping-pong buffering mechanism in 1D-SepCNN further improves deployability under tight memory constraints. (4) We extend a hardware-aware search framework to support constraint-driven model configuration selection, considering accuracy, deployability, latency, and energy consumption. Evaluated on two swipe-direction datasets with multiple users and ordinary tables, our approach achieves low-latency, energy-efficient inference on the AMD Spartan-7 XC7S25 FPGA. Under the PS data splitting setting, the selected 6-bit 1D-CNN reaches 0.970 average accuracy across users with 9.22 ms latency. The chosen 8-bit 1D-SepCNN further reduces latency to 6.83 ms (over 53x CPU speedup) with slightly lower accuracy (0.949). Both consume under 1.2 mJ per inference, demonstrating suitability for long-term edge operation.
Problem
Research questions and friction points this paper is trying to address.
Enabling vibration-based gesture recognition on everyday furniture using energy-efficient FPGAs
Reducing computational complexity and energy consumption for real-time deployment
Replacing complex preprocessing with raw waveform input to simplify implementation
Innovation
Methods, ideas, or system contributions that make the work stand out.
Uses raw vibration input without complex preprocessing
Deploys lightweight 1D-CNNs on energy-efficient FPGAs
Implements integer quantization and automated RTL generation
🔎 Similar Papers
No similar papers found.
