Millimeter-wave (mmWave) imaging is increasingly being adopted across the supply chain and security industries to see through occlusions and detect hidden objects. Yet, most industrial mmWave imaging relies on a static, bulky, and expensive infrastructure to achieve precisely measured positions required for synthetic aperture imaging. In contrast, a handheld mmWave scanner can provide a compact, mobile and cost-effective alternative for high-resolution mmWave imaging, but it suffers from severe phase errors and image distortions due to motion inaccuracy.

mmWave radar based sensors are increasingly being deployed for robotics and automobile applications to sense the environment. Unfortunately, specular reflections of smooth and planar surfaces makes mmWave sensing unreliable at detecting objects and vehicles in the real world. In fact, due to the small wavelength of mmWave radars, most surfaces are smooth reflectors for the radar signals. This paper addresses this fundamental limitation of detecting specular reflectors by capturing the energy reflected away from the radar.

A general consumer around the world gets general awareness about the causes of mechanical, electrical or digital problems in their house (or neighborhood) and how to fix them through common life experiences or learning from their peers. However, despite similar level of integration into everyday life, the understanding of wireless technologies lags significantly. This bottleneck leads to consumer dissatisfaction and reduced profits for cellular providers.

A general consumer around the world gets general awareness about the causes of mechanical, electrical or digital problems in their house (or neighborhood) and how to fix them through common life experiences or learning from their peers. However, despite similar level of integration into everyday life, the understanding of wireless technologies lags significantly. This bottleneck leads to consumer dissatisfaction and reduced profits for cellular providers.

The fundamental bottleneck in adapting data-driven wireless solutions to real world is the lack of tools for augmenting good quality data which is environment-specific. Indeed, much of the data required for popular data-driven wireless communication and sensing systems requires domain expertise beyond the reach of an average consumer. This demo presents a radical new vision for generating synthetic data in consumer-specific environments by leveraging the power of modern compute.