Ultra wideband (UWB) through wall radar (TWR) can detect human targets behind strong bodies by using the penetration characteristics of low-frequency electromagnetic waves, and is widely used in urban operations, anti-terrorism conflicts, disaster rescue, criminal investigation, search and rescue and other fields. In these application fields, human motion recognition is one of the most challenging topics in moving target recognition through walls. Because of the influence of the wall, such as attenuation, refraction and multipath effects, it will bring obvious distortion to the echo signal of the ultra wideband through wall radar, resulting in a significant decline in the accuracy of human motion recognition, a significant increase in the calculation cost of available models, and a very challenging system deployment. This forces us to develop algorithms with faster convergence speed and higher reliability, while the accuracy can still be maintained at a higher level under the same background conditions.

We built this site to record and present our work more visually, and to provide peers with links to Google Drive, Baidu Cloud and Github for some of our open source work at the same time. Since some of our core work is a team effort, it is not open-source permitted. Work done by me personally and not involving a conflict of interest will be open to download links. Hope more peers can join the open source cooperation!