Haibo Zhao

Haibo Zhao

Auto-generated profile, based on publicly available information

Haibo Zhao

School of Microelectronics, the State Key Laboratory of Advanced Materials for Intelligent Sensing the Key Laboratory of Imaging and Sensing Microelectronic Technology, and Tianjin Digital Information Technology Research Center, Tianjin University, Tianjin, China

Research Focus & Metrics

Academic Career Length

4Years (2023-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
30 June 2023 in IEEE Internet of Things Journal
Verified
Journal Article
Closed access
HZHaibo Zhao
Haibo Zhao
YMYongtao Ma
Yongtao Ma
YHYuxiang Han
Yuxiang Han
CTChenglong Tian
Chenglong Tian
XHXinyue Huang
Xinyue Huang
The four stages (first heart sound (S1), systole, second heart sound (S2), and diastole) of heartbeat sounds recorded by contact seismocardiogram (SCG) reflect the health of the heart, but these stages are challenging to measure by noncontact millimeter wave radar. If the sampling rate of millimeter wave radar is increased, this will increase the amount of data storage needed for the long-term monitoring of human vital signs. This article presents an algorithm for reconstructing the envelope of high-frequency heart sound signals using low-frequency millimeter wave radar signals, as well as a heart sound envelope segmentation algorithm based on peak points. Its design principle is a combination of signal processing and a transformer network, which is called T-HSER. This technique maps the low-frequency radar signal into a high-frequency heart sound envelope signal through the transformer network and determines the four different stages of the heart sound using appropriate thresholds. Based on the training of more than 30000 heartbeats of 25 healthy subjects and the prediction evaluation of six subjects, the T-HSER algorithm is shown to reconstruct the high-frequency heart sound envelope signal with high correlation. Moreover, the mean correlation can reach 0.85 on one minute of data, which is higher than that of the bidirectional long short-term memory algorithm, and can effectively distinguish the four stages of the heart sound so that the mean absolute error (MAE) between the predicted value and the ground truth of S1 and S2 is within a tolerable range (70 ms). At the same time, the algorithm is suitable for low sampling rate radar, which greatly reduces the amount of data storage required.

2 October 2023 in IEEE Sensors Journal
Verified
Journal Article
Closed access
HZHaibo Zhao
Haibo Zhao
YMYongtao Ma
Yongtao Ma
YLYiwei Lu
Yiwei Lu
KLKaihua Liu
Kaihua Liu
In the field of human-computer interaction, millimeter-wave radar has attracted considerable attention as a contactless and private approach to hand gesture recognition. However, single-target gesture recognition scenes are generally simplistic and not representative of real human-computer interactions. Therefore, this research examines the feasibility of using a single millimeter wave radar to recognize hand gestures in double-target scenes by combining radar theory with deep learning. First, a dynamic range angle image (DRAI) of the double-target gesture is composed using the weights and DRAIs of two single-target gestures. Thus, a gesture separation network is studied to separate double-target gestures into two single-target gestures. Then, a convolutional neural network and long short-term memory (CNN+LSTM) model is applied to classification recognition. Finally, experiments are conducted to show that the proposed double-target gesture separation and classification recognition (DGSCR) system has a high recognition accuracy for gestures in new environments and positions. The viability of this method is verified using a public dataset. The CNN+LSTM model is validated using test set, which shows that the maximum accuracy across different positions is 99%. Moreover, the average accuracy after separation across the different environments for two targets in staggered arrangements is 93%. Furthermore, when facing unknown gestures in all samples with double targets, the gesture separation network also has good adaptability, with an average accuracy of 81.6%.

30 April 2024 in IEEE Internet of Things Journal
Verified
Journal Article
Closed access
HZHaibo Zhao
Haibo Zhao
YMYongtao Ma
Yongtao Ma
XWXiaofeng Wang
Xiaofeng Wang
Basic life activities of the human body can be classified as either static or dynamic. Millimeter wave radars can monitor physiological signals in the static state and posture in the dynamic state. These biometrics can be used for identity recognition to address matching vital sign monitoring with the user identity information. However, existing studies are based on single-modal information for identity recognition using either static or dynamic information. This study proposes a framework called MN-UIV for enabling user identify verification using static and dynamic information via millimeter wave radar. The design method entails a combination of signal preprocessing and multimodal neural networks. First, the algorithm uses two radars: Radar 1 obtains physiological signals (in the time and frequency domains of breathing and heartbeat) in the static state and Radar 2 obtains the dynamic range angle image (DRAI) of the posture in the dynamic state. Multimodal information is later formed to extract features from multiple aspects. Here, the user identity information is verified using two designed multimodal neural networks (CNN+LSTM_Encoder (CLE), ResNet_Encoder (RNE)). Finally, based on the predictive evaluation of five users, the average recognition rates of the two multimodal neural networks were 96% and 96.4%, respectively. Compared to the average recognition accuracy of single-modal neural networks, it has a maximum improvement of 45.2% and 31%, respectively. It can continuously monitor the signs and posture of the body, thereby reducing system idle time and improving time utilization.

21 July 2023 in IEEE Sensors Journal
Verified
Journal Article
Closed access
YLYiwei Lu
Yiwei Lu
YMYongtao Ma
Yongtao Ma
WNWanru Ning
Wanru Ning
HZHaibo Zhao
Haibo Zhao
Radio Tomographic Imaging (RTI) is a passive localization method that has important applications in Device-Free Localization (DFL). RTI estimates the location of a target by using the impact of the target entering a Region of Interest (ROI) on the radio link. However, research on RTI in the field of passive UHF RFID is still insufficient, and there are many issues that need to be addressed. In the RTI method, targets typically appear at the pixel with the highest value in the reconstructed image, displaying the location of the target intuitively by the variation in Received Signal Strength (RSS) as the input. However, multipath effects, noise interference, and the presence of unknown targets in the ROI can lead to the appearance of multiple local maxima in the reconstructed image, making it very difficult to accurately identify the number and location of the targets. We proposed a method to establish feature regions in image reconstruction and introduced the method of link-tracking which use more robust known link geometric features than RSS measurement values to match the false targets with the real targets one by one. The algorithm traverses all the feature region combinations in a specific order and outputs the localization result, distinguishing targets from false targets and obtaining more accurate results. Experimental results verified the convergence and advantages of the proposed method.

12 September 2024 in IEEE Internet of Things Journal
Verified
Journal Article
Closed access
ZWZemin Wang
Zemin Wang
YMYongtao Ma
Yongtao Ma
XLXiuyan Liang
Xiuyan Liang
YCYicheng Chu
Yicheng Chu
HZHaibo Zhao
Haibo Zhao
KLKaihua Liu
Kaihua Liu
With the rapid development of the Internet of Things (IoT), an increasing number of industrial demands have become urgent. Radio frequency identification (RFID) system plays a crucial role in addressing these challenges. It has now become a significant direction for the development of industrial automatic identification and data collection technology. However, the existing three-dimensional (3D) high-precision absolute localization methods have issues that need to be resolved. For example, the antenna sampling position needs to be used as known prior information, and the timestamp information of the samples is not fully utilized. Additionally, relying on multiple antennas or reference tags as auxiliary measures reduces the system flexibility. To overcome these challenges, we propose a single-antenna, multi-target, and 3D localization method based on an attention mechanism and neural network regression model inspired by the synthetic aperture (SA) method. This method suggests using a single antenna moving uniformly on a slider for continuous motion and sampling. By utilizing sample phase and time-domain information, it achieves 3D localization of cargo boxes in intelligent warehousing environments. In comparison with the current state-of-the-art localization solutions, our method doesn’t give the use of reference tags and utilizes a reduced number of antennas. The experimental results prove the feasibility of using only mobile antennas without reference tags to perform 3D multi-target positioning tasks. The combined dimensional average positioning error of 3D indoor positioning achieves a high-precision positioning, reaching 2.04cm and exhibits good robustness.

9 January 2025 in IEEE Transactions on Instrumentation and Measurement
Verified
Journal Article
Closed access
YLYiwei Lu
Yiwei Lu
YMYongtao Ma
Yongtao Ma
YHYuxiang Han
Yuxiang Han
LZLianfu Zhao
Lianfu Zhao
HZHaibo Zhao
Haibo Zhao
In the realm of smart homes, the emphasis lies in fostering seamless interaction between user-device and device-device within their environment, bringing convenience to both work and life. However, users may not always carry signal transceiving devices to establish connections with the environment. In such cases, device-free localization (DFL) plays a pivotal role. DFL involves identifying targets, precisely locating them, and providing feedback to other devices, thereby filling the information gap for targets without carried interactive devices. Most traditional passive UHF RFID localization methods can only achieve single-target localization at a time, while many application scenarios require multi-target localization. To address this challenge, this paper proposes a device-free multi-target Link Fingerprint (DMLF) localization method. DMLF analyzes the impact of targets on the UHF RFID links, extracts target-specific features from the observation data. A novel approach to link fingerprint localization is introduced, involving the compression and reconstruction of feature information into link fingerprints. Furthermore, considering the characteristics of link fingerprints, virtual reference fingerprints and dynamic fingerprint clustering are introduced to enhance localization accuracy and reduce computational overhead. There is no need for personnel to collect fingerprint data at sampling points during the offline phase, compared to received signal strength indicator (RSSI) fingerprints, link fingerprints provide a more accurate reflection of the target’s position and exhibit stronger robustness. This method reduces the storage requirements for the fingerprint database and computational complexity. Experimental results indicate that, in comparison to traditional localization methods, the DMLF algorithm achieves higher accuracy. The average localization accuracies for scenarios with 1 to 5 targets are 0.2679m, 0.3962m, 0.4937m, 0.7637m, and 1.2226m, respectively, with link fingerprints contributing to the enhanced precision of localization.