Radar Data Cube Processing for Human Activity Recognition Using Multisubspace Learning

Erol, Barış; Amin, Moeness G.

doi:10.1109/taes.2019.2910980

Cited by 77 publications

(36 citation statements)

References 26 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Raw radar data contains hierarchical information [13], which can be used to measure range and velocity. Data from the FMCW radar can be mapped to the Range-Time domain with a Fast Fourier Transform (FFT), whereas the Doppler-Time domain or radar spectrogram is generated by performing a Short-Time Fourier Transform (STFT) on the range profiles.…”

Section: A Feature Fusion With Conventional Classifiersmentioning

confidence: 99%

Sequential Human Gait Classification With Distributed Radar Sensor Fusion

Mehul

Kernec

et al. 2021

IEEE Sensors J.

View full text Add to dashboard Cite

Section: A Feature Fusion With Conventional Classifiersmentioning

confidence: 99%

Sequential Human Gait Classification With Distributed Radar Sensor Fusion

Mehul

Kernec

et al. 2021

IEEE Sensors J.

View full text Add to dashboard Cite

“…

σ

is the radar cross‐section,

L_{a}

and

L_{s}

indicate the atmospheric and system loss, respectively. The complex baseband signal is generated by the process of de‐chirping and expressed in terms of the in‐phase and quadrature components as

s false(t false) = I false(t false) + normalj Q false(t false) = A e^{ψ false(t false)}

where

()(, ψ false(t false))

is the signal phase [25, 30, 31]. This signal is used in all follow‐on analysis.…”

Section: Radar System and Data Analysismentioning

confidence: 99%

“…The 2D PCA has shown to be very effective in motion classification. It outperforms hand‐crafted‐based motion classifications and offers competitive results to convolution and deep neural networks [14, 25, 26]. Both the target micro‐Doppler signature, provided by the spectrograms and the target range‐map are input to the 2D PCA.…”

Section: Introductionmentioning

confidence: 99%

Radar classifications of consecutive and contiguous human gross‐motor activities

Amin

Guendel

2020

IET Radar, Sonar & Navigation

Self Cite

View full text Add to dashboard Cite

“…Hence, numerous studies in the literature have investigated the use of radar sensing for human activities classification, personnel recognition, and presence sensing, even in through-the-wall conditions [10]- [16]. The radar information can be represented in a 3D space, containing range (physical distance), time, and velocity (measured through the Doppler effect), sometimes referred to as "radar cube" [17]. Among these different domains of radar information, micro-Doppler is typically used, exploiting the small modulations on the received radar signal caused by "micro-motion" of individual body parts (e.g., limbs, torso, head) [18].…”

Section: Introductionmentioning

confidence: 99%

Bi-LSTM Network for Multimodal Continuous Human Activity Recognition and Fall Detection

et al. 2020

View full text Add to dashboard Cite

This paper presents a framework based on multi-layer bi-LSTM network (bidirectional Long Short-Term Memory) for multimodal sensor fusion to sense and classify daily activities' patterns and high-risk events such as falls. The data collected in this work are continuous activity streams from FMCW radar and three wearable inertial sensors on the wrist, waist, and ankle. Each activity has a variable duration in the data stream so that the transitions between activities can happen at random times within the stream, without resorting to conventional fixed-duration snapshots. The proposed bi-LSTM implements soft feature fusion between wearable sensors and radar data, as well as two robust hard-fusion methods using the confusion matrices of both sensors. A novel hybrid fusion scheme is then proposed to combine soft and hard fusion to push the classification performances to approximately 96% accuracy in identifying continuous activities and fall events. These fusion schemes implemented with the proposed bi-LSTM network are compared with conventional sliding window approach, and all are validated with realistic "leaving one participant out" (L1PO) method (i.e. testing subjects unknown to the classifier). The developed hybrid-fusion approach is capable of stabilizing the classification performance among different participants in terms of reducing accuracy variance of up to 18.1% and increasing minimum, worst-case accuracy up to 16.2%.

show abstract

Radar Data Cube Processing for Human Activity Recognition Using Multisubspace Learning

Cited by 77 publications

References 26 publications

Sequential Human Gait Classification With Distributed Radar Sensor Fusion

Sequential Human Gait Classification With Distributed Radar Sensor Fusion

Radar classifications of consecutive and contiguous human gross‐motor activities

Bi-LSTM Network for Multimodal Continuous Human Activity Recognition and Fall Detection

Contact Info

Product

Resources

About