Robust Real-Time Embedded EMG Recognition Framework Using Temporal Convolutional Networks on a Multicore IoT Processor

• Published in: IEEE transactions on biomedical circuits and systems Vol. 14; no. 2; pp. 244 - 256
• Main Authors: Zanghieri, Marcello, Benatti, Simone, Burrello, Alessio, Kartsch, Victor, Conti, Francesco, Benini, Luca
• Format: Journal Article
• Language: English
• Published: United States IEEE 01.04.2020; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Accuracy; Adult; Classification; Control systems design; Convolution; Convolutional neural networks; Datasets; Electromyography; Electromyography - instrumentation; Embedded systems; Energy efficiency; Equipment Design; Gesture recognition; Gestures; Hand - physiology; Human motion; Humans; Internet of Things - instrumentation; Male; Man-Machine Systems; Microprocessors; Motion perception; multi-layer neural networks; neural network hardware; Neural Networks, Computer; Performance degradation; Prosthetics; Real-time systems; Recognition; Reliability; Robustness; Signal processing; Signal Processing, Computer-Assisted - instrumentation; Support vector machines; temporal convolutional networks; Topology; Wearable technology
• ISSN: 1932-4545; 1940-9990; 1940-9990
• DOI: 10.1109/TBCAS.2019.2959160

Hand movement classification via surface electromyographic (sEMG) signal is a well-established approach for advanced Human-Computer Interaction. However, sEMG movement recognition has to deal with the long-term reliability of sEMG-based control, limited by the variability affecting the sEMG signal. Embedded solutions are affected by a recognition accuracy drop over time that makes them unsuitable for reliable gesture controller design. In this paper, we present a complete wearable-class embedded system for robust sEMG-based gesture recognition, based on Temporal Convolutional Networks (TCNs). Firstly, we developed a novel TCN topology (TEMPONet), and we tested our solution on a benchmark dataset (Ninapro), achieving 49.6% average accuracy, 7.8%, better than current State-Of-the-Art (SoA). Moreover, we designed an energy-efficient embedded platform based on GAP8, a novel 8-core IoT processor. Using our embedded platform, we collected a second 20-sessions dataset to validate the system on a setup which is representative of the final deployment. We obtain 93.7% average accuracy with the TCN, comparable with a SoA SVM approach (91.1%). Finally, we profiled the performance of the network implemented on GAP8 by using an 8-bit quantization strategy to fit the memory constraint of the processor. We reach a <inline-formula><tex-math notation="LaTeX">\mathbf {4\times }</tex-math></inline-formula> lower memory footprint (460 kB) with a performance degradation of only 3% accuracy. We detailed the execution on the GAP8 platform, showing that the quantized network executes a single classification in 12.84 ms with a power envelope of 0.9 mJ, making it suitable for a long-lifetime wearable deployment.