Motion trajectories contain rich information about human activities. We propose to use a 2D LIDAR to perform multiple people activity recognition simultaneously by classifying their trajectories. We clustered raw LIDAR data and classified the clusters into human and non-human classes in order to recognize humans in a scenario. For the clusters of humans, we implemented the Kalman Filter to track their trajectories which are further segmented and labelled with corresponding activities. We introduced spatial transformation and Gaussian noise for trajectory augmentation in order to overcome the problem of unbalanced classes and boost the performance of human activity recognition (HAR). Finally, we built two neural networks including a long short-term memory (LSTM) network and a temporal convolutional network (TCN) to classify trajectory samples into 15 activity classes collected from a kitchen. The proposed TCN achieved the best result of 99.49% in overall accuracy. In comparison, the TCN is slightly superior to the LSTM network. Both the TCN and the LSTM network outperform hidden Markov Model (HMM), dynamic time warping (DTW), and support vector machine (SVM) with a wide margin. Our approach achieves a higher activity recognition accuracy than the related work.

Human activity detection outdoors is emerging as a very important research field due to its potential application in surveillance, assisted living, search and rescue, and military applications. For such applications it is important to have detailed information about the human target, for example, whether the detected target is a single person or a group of people, what activity a target is performing, and the rough location of the target. In this paper, we propose novel usage of machine learning techniques to perform subject classification, human activity classification, people counting, and coarse localization by classifying micro-Doppler signatures obtained from a low-cost and low-power radar system. Our experiments were performed outdoors. For feature extraction of micro-Doppler signatures, we applied a two-directional two-dimensional principle component analysis (2D2PCA). Our results show that by applying 2D2PCA, the accuracy results of Support Vector Machine (SVM) and knearest neighbors (kNN) classifiers were greatly improved. We also designed and implemented a Convolutional Neural Network (CNN) for the target classifications in terms of type, number, activity and coarse localization. Our CNN model obtained very high classification accuracies (97to the best results obtained by SVM and kNN. Finally, we investigated the effects of the frame length of the sliding window, the angle of the direction of movement, and the number of radars used on the classification performance, providing valuable guidelines for machine learning modeling and experimental setup of micro-Doppler based research and applications.

In this article, a cost-effective ultrawideband (UWB) communication system for gesture recognition in a smart home environment is proposed, which uses gesture trajectories and a deep learning model. Most previous studies of gesture recognition using the UWB technology used electromagnetic signals directly, which may bring problems, such as radar clutter, signal coupling, multipath, fading, and interference. However, instead of using UWB’s high-frequency pulse signals, the proposed method only uses gesture trajectories by data positioning. To this end, first, a data set of four gesture activities was created. Then, this data set was trained using a convolutional neural network (CNN) integrated with a squeeze-and-excitation (SE) block, namely, the SE-Conv1D model. Finally, the system was prototyped to interact with appliances in practical smart homes. The experimental data was used to demonstrate the superiority of the SE-Conv1D model in comparison with four baselines: 1) support vector machines; 2) $K$ -nearest neighbor; 3) random forest; and 4) binarized neural networks. Experimental results show that all collected gesture activities are correctly recognized with an overall accuracy of over 95%, among which the proposed SE-Conv1D model achieves the best accuracy of 99.48%. The proposed system is a complete end-to-end sensing system specifically designed for tracking and recognizing human gestures, which is robust against interference and changes in distance or direction. In addition, the proposed system can tackle the device selection problems for smart homes, which means it is reliable for real-world applications.

Radar-based human activity recognition is attracting a wide range of interest from both industry and academia because of its through-wall ability, privacy-preserving capability, and device-free detection. Currently, most radar-based systems consider signal analysis and feature extraction in the frequency domain or the temporal domain independently without fusing them together. In this paper, in order to model both frequency properties and temporal profiles of human activity, we proposed a spectro-temporal network (STnet) that integrates a temporal convolutional network (TCN) and a convolutional neural network (CNN). It can extract temporal patterns and micro-Doppler features from radar signals for human activity recognition. In the experiments, two radar sensors and one base station were used to build a low-power wireless radar sensor network. Fifteen activities were investigated in a real kitchen scenario by using this radar sensor network. Frequency spectrograms were obtained after signal processing using a short-time Fourier transform (STFT). They were further segmented using a short sliding window (2.5s), which enables a very small latency. The proposed STnet achieved 99.64% overall accuracy in testing, which is superior to the other three networks that we implemented in this work. Our work also can be used as a generic solution to other sensor-based (wearable sensors, WiFi CSI, etc.) activity recognition.

Integrated sensing and communication technologies provide essential sensing capabilities that address pressing challenges in remote health monitoring systems. However, most of today’s systems remain obtrusive, requiring users to wear devices, interfering with people’s daily activities, and often raising privacy concerns. Herein, we present HealthDAR, a low-cost, contactless, and easy-to-deploy health monitoring system. Specifically, HealthDAR encompasses three interventions: 1) symptom early detection (monitoring of vital signs and cough detection); 2) tracking and social distancing; and 3) preventive measures (monitoring of daily activities, such as face-touching and hand-washing). HealthDAR has three key components: 1) a low-cost, low-energy, and compact integrated radar system; 2) a simultaneous signal processing combined deep learning (SSPDL) network for cough detection; and 3) a deep learning method for the classification of daily activities. Through performance tests involving multiple subjects across uncontrolled environments, we demonstrate HealthDAR’s practical utility for health monitoring.