A wearable biosensing system with in-sensor adaptive machine learning for hand gesture recognition

Moin, Ali; Zhou, Andy; Rahimi, Abbas; Menon, Alisha; Benatti, Simone; Alexandrov, George; Tamakloe, Senam; Ting, Jonathan; Yamamoto, Natasha A. D.; Khan, Yasser; Burghardt, Fred; Benini, Luca; Arias, Ana Claudia; Rabaey, Jan M.

doi:10.1038/s41928-020-00510-8

Cited by 422 publications

(296 citation statements)

References 47 publications

Supporting

Mentioning

272

Contrasting

Order By: Relevance

“…Table 1 shows the classification accuracy in the noise-free case. A support vector machine (SVM) with linear kernel and cost parameter on pre-processed, flattened features in float-32 precision with dimension 320 (64 channels 5-gram) [ 49 ] as well as an HD classifier with multiplicative mapping [ 22 ] serve as baselines. Both HD classifiers operate at a dimension of .…”

Section: Case Study: Hybrid Near-channel Classification and Data Transmission In Emg-based Gesture Recognitionmentioning

confidence: 99%

“…Both HD classifiers operate at a dimension of . The SVM marginally outperforms the HD classifiers by 0.14% and 2%; however, in contrast to the HD classifiers, the SVM does not support online updates of the model, which is crucial for practical deployment of EMG applications [ 49 ]. The bipolar feature embedding using the CiM instead of the float-based multiplicative mapping in the HD classification yields only a small accuracy degradation (95.99% vs. 94.13%).…”

Section: Case Study: Hybrid Near-channel Classification and Data Transmission In Emg-based Gesture Recognitionmentioning

confidence: 99%

See 1 more Smart Citation

Near-channel classifier: symbiotic communication and classification in high-dimensional space

et al. 2021

Self Cite

View full text Add to dashboard Cite

Brain-inspired high-dimensional (HD) computing represents and manipulates data using very long, random vectors with dimensionality in the thousands. This representation provides great robustness for various classification tasks where classifiers operate at low signal-to-noise ratio (SNR) conditions. Similarly, hyperdimensional modulation (HDM) leverages the robustness of complex-valued HD representations to reliably transmit information over a wireless channel, achieving a similar SNR gain compared to state-of-the-art codes. Here, we first propose methods to improve HDM in two ways: (1) reducing the complexity of encoding and decoding operations by generating, manipulating, and transmitting bipolar or integer vectors instead of complex vectors; (2) increasing the SNR gain by 0.2 dB using a new soft-feedback decoder; it can also increase the additive superposition capacity of HD vectors up to 1.7$$\times$$ × in noise-free cases. Secondly, we propose to combine encoding/decoding aspects of communication with classification into a single framework by relying on multifaceted HD representations. This leads to a near-channel classification (NCC) approach that avoids transformations between different representations and the overhead of multiple layers of encoding/decoding, hence reducing latency and complexity of a wireless smart distributed system while providing robustness against noise and interference from other nodes. We provide a use-case for wearable hand gesture recognition with 5 classes from 64 EMG sensors, where the encoded vectors are transmitted to a remote node for either performing NCC, or reconstruction of the encoded data. In NCC mode, the original classification accuracy of 94% is maintained, even in the channel at SNR of 0 dB, by transmitting 10,000-bit vectors. We remove the redundancy by reducing the vector dimensionality to 2048-bit that still exhibits a graceful degradation: less than 6% accuracy loss is occurred in the channel at − 5 dB, and with the interference from 6 nodes that simultaneously transmit their encoded vectors. In the reconstruction mode, it improves the mean-squared error by up to 20 dB, compared to standard decoding, when transmitting 2048-dimensional vectors.

show abstract

Section: Case Study: Hybrid Near-channel Classification and Data Transmission In Emg-based Gesture Recognitionmentioning

confidence: 99%

Section: Case Study: Hybrid Near-channel Classification and Data Transmission In Emg-based Gesture Recognitionmentioning

confidence: 99%

Near-channel classifier: symbiotic communication and classification in high-dimensional space

et al. 2021

Self Cite

View full text Add to dashboard Cite

show abstract

“…With the rapid development of wearable electronics for the applications of motion monitoring, health monitoring, internet of things, virtual reality and sensory enhancement, the need for flexible and wearable energy supply system arouses considerable interest in recent years. [1][2][3][4] The human body, as a constant heat source, can potentially act as an in situ energy base if the waste body heat can be conveniently collected and efficiently converted. 5 As a green and sustainable energy technique, thermoelectric generators (TEGs) can direct convert waste heat into electricity without the involvement of moving parts or the emission of waste gas/liquid.…”

Section: Introductionmentioning

confidence: 99%

A Wavy-Structured Highly Stretchable Thermoelectric Generator with Stable Energy Output and Self-Rescuing Capability

et al. 2021

View full text Add to dashboard Cite

Thermoelectric generators (TEGs) demonstrate great potential for flexible and wearable electronics due to the direct electrical energy harvesting from waste heat. Good wearability requires high mechanical flexibility and preferable stretchability, while current TEGs are primarily developed with rigid or non-stretchable components, which cannot well conform to human skin or accommodate human motions, thus hindering further applications. Herein, a wavy architecture is proposed for the fabrication of stretchable TEGs, where a stretchable and self-healable hydrogel is employed as device substrate, and intrinsically flexible highperformance TE films are attached to form wavy morphologies. The wavy-structured TEG can be readily stretched to 300%, and in the meantime can sustain a stable energy output with more than 90% TE performance retained, which outperforms most of the reported state-of-the-art TE materials and TEGs. It also demonstrates a desired device-level self-rescuing capability originated from the effective self-healing of the hydrogel and the wavy structure of TE legs, thereby providing persistent energy supply upon injuries or damages. This work offers a promising approach for the design of next-generation stretchable and wearable TEGs.

show abstract

“…Hand gesture recognition has been an active field of research due to the quest to provide a better, more efficient, and intuitive mechanism for human-computer interaction (HCI). Many different sensor modalities have been used to address this challenge, such as radar [1][2][3], cameras [4,5], dynamic vision sensors (DVS) [6][7][8][9], or electromyography (EMG) systems [10,11]. Moreover, several different machine learning techniques have been proposed to address HGR, such as a convolutional neural network (CNN) [12], long short-term memory (LSTM) [12,13], and spiking neural networks [8,11].…”

Section: Introductionmentioning

confidence: 99%

Radar-Based Hand Gesture Recognition Using Spiking Neural Networks

et al. 2021

View full text Add to dashboard Cite

We propose a spiking neural network (SNN) approach for radar-based hand gesture recognition (HGR), using frequency modulated continuous wave (FMCW) millimeter-wave radar. After pre-processing the range-Doppler or micro-Doppler radar signal, we use a signal-to-spike conversion scheme that encodes radar Doppler maps into spike trains. The spike trains are fed into a spiking recurrent neural network, a liquid state machine (LSM). The readout spike signal from the SNN is then used as input for different classifiers for comparison, including logistic regression, random forest, and support vector machine (SVM). Using liquid state machines of less than 1000 neurons, we achieve better than state-of-the-art results on two publicly available reference datasets, reaching over 98% accuracy on 10-fold cross-validation for both data sets.

show abstract

A wearable biosensing system with in-sensor adaptive machine learning for hand gesture recognition

Cited by 422 publications

References 47 publications

Near-channel classifier: symbiotic communication and classification in high-dimensional space

Near-channel classifier: symbiotic communication and classification in high-dimensional space

A Wavy-Structured Highly Stretchable Thermoelectric Generator with Stable Energy Output and Self-Rescuing Capability

Radar-Based Hand Gesture Recognition Using Spiking Neural Networks

Contact Info

Product

Resources

About