A5021893881- Indoor and Outdoor Localization Technologies
- Microwave Imaging and Scattering Analysis
- Target Tracking and Data Fusion in Sensor Networks
- Advanced Optical Sensing Technologies
- Advanced SAR Imaging Techniques
- Geophysical Methods and Applications
- Robotics and Sensor-Based Localization
- Sparse and Compressive Sensing Techniques
- Non-Invasive Vital Sign Monitoring
- Radar Systems and Signal Processing
- Gait Recognition and Analysis
- Spectroscopy and Laser Applications
- Image and Object Detection Techniques
- Hand Gesture Recognition Systems
- Ultra-Wideband Communications Technology
University of Electronic Science and Technology of China
2021-2025
Non-line-of-sight (NLOS) detection is an enduring topic as it provides a powerful tool to monitor visually blocked areas. Currently, the NLOS requires precise prior knowledge of building layout, which limits its further applications in practice. In this paper, we consider problem joint estimation layout and target location scenario by exploiting multipath returns. Specifically, first, simplified into combined linear equations with unknown parameters. way, establish parametrized propagation...
Localization of non-line-of-sight (NLOS) targets in the complex urban environment have attracted significant attention recent years. However, requirement for precise prior information about is idealistic. It challenging to know environmental blind area vision advance practical applications. This paper proposes a joint estimation algorithm building layout and target position L-shaped scene without any information. Specifically, round-trip multipath propagation model first developed cases...
Among various types of sensors, through-the-wall imaging radar (TWIR) is widely used in concealed targets detection and urban environment perception. Especially, multi-view TWIR has been received more attention recent years for its ability to provide accurate position estimation reduce the blind areas. However, when exploiting detect complicated indoor with an unavailable global positioning system, relative radars hard be obtained acquired non-uniform images cannot aligned automatically. To...
Non-line-of-sight (NLOS) radar has the feature of detecting targets which are in blind area, while existing time-of-arrival (ToA)-based methods limit its further development multi-target environments. To deal with it, we investigate problem multiple localization behind L-shaped corner for multiple-input-multiple-output (MIMO) millimeter wave (MMW) radar. Specifically, multipath signal model that accounts range, Doppler, direction departure (DoD) and arrival (DoA) is established first. After...
Through-wall human motion recognition is suffered from the problems of too few samples and large model parameters. In this letter, we propose a multi-domain fusion through-the-wall radar (TWR) based on lightweight network transfer learning. Specifically, in order to make full use target information, multiple parallel feature pyramid (FPN) first proposed extract detailed information time–frequency map range profile. After that, MobileNetV3 learning proposed. The pre-trained public ImageNet...
Detecting nonline-of-sight (NLOS) targets is a crucial concern in both military and civilian applications. Currently, NLOS detection techniques generally focus on the range relationship of multipath, which limits its further applications multitarget environments. To deal with it, this article raises novel target localization method inspired by distribution characteristics multipath different domains. Specifically, first, echo model for L-shaped scenario via ultrawideband (UWB)...
This letter studies the problem of ultrawideband (UWB) tomographic imaging for unknown building layouts where multipath-rich condition is considered. Specifically, first, multiple propagation paths UWB signal are analyzed, and a delay estimation algorithm proposed to estimate direct path (DP) from multipath signal. Then, projection model established by mapping relationship between DP relative permittivity layout. Besides, modified total variation method developed reconstruct layout, which...
This paper proposes a method for locating non-line-of-sight (NLOS) target in long L-shaped scenario with ultra-wideband (UWB) radar. Specifically, refined multipath signal model adapted to the under consideration is established. Next, we propose two-stage localization method, which first exploits diffraction path initially determine area where might exist. Then an algorithm based on grid matching executed acquire precise localization. Finally, numerical simulation proves feasibility of this method.
We proposed a robust real-time human activity recognition method based on attention-augmented Gated Recurrent Unit (GRU) using radar range profile, namely Attention-Augmented Sequential Classification (AASC). use attention mechanism to capture the temporal relationships inherent in profile signatures. Therefore, our model can learn long-term correlation of without increasing depth or width recurrent neural network. The weights are adaptively generated features extracted by GRU Finally,...