ACED Fab: 240-GHz Energy-Efficient CMOS MIMO Radar

University Of California-Berkeley, Berkeley CA

Investigators

Abstract

Radar technology is essential in modern automotive applications, enabling automatic cruise control and semi-autonomous driving. Using advanced multiple antenna array of radars has the promise to offer even more functionality and applications, including camera-less detection of living people/animals and objects, three dimensional localization, and gesture recognition to enhance user interfaces. Despite these many important applications, these radar systems still require considerable research to improve their performance, and to reduce the size and energy footprint. While automotive applications can tolerate the high-power consumption, other applications, such as sensors in buildings and homes and mobile gesture recognition require the radar module to be smaller and to consume less power. To address these many challenges, our proposed research will explore low-cost integrated circuit technology for next generation radars operating at 240 GHz. Successful realization of a low power CMOS 240 GHz radar would enable high resolution radar and imaging. Next generation communication or “6G” will likely benefit from the availability of spectrum above 100 GHz, and thus our research would impact broadband connectivity, thus broadening wireless access, which is a key national priority. Operation over 100 GHz also has potential new applications in biomedical imaging, cancer detection, spectroscopy, and security. Our proposed research will explore low-cost CMOS technology for next generation MIMO radars operating at 240 GHz. Operating at 240 GHz has many benefits, 3-4X smaller wavelength than 60 GHz and E-band (today's radars) allows miniaturization and higher angular resolution, and large MIMO arrays for full imaging. At the same time, operating at such high frequencies comes with many challenges due to the lower power gain and activity limits of CMOS technology. Our team’s focus is on key building blocks, in particular the 240 GHz receiver, the transmitter, and the frequency generation and distribution. Additionally, we will address system and antenna/packaging challenges, since power, performance and cost are highly sensitive to packaging. The integrated circuit needs an efficient means to couple power to the antenna array, otherwise the performance benefits will be lost in transition losses. By tackling the entire system, from baseband to antenna, and by working closely, we plan to pave the path for next generation sub-Thz radars. Our proposed solution must also be designed in the context of next generation applications. Emerging applications for mm-wave radar, including gesture recognition and automotive arrays, require sensitivity for nearby objects, and undesired large objects near the target, such as tables and walls, produce large reflections that can easily swamp out the desired signal. This means that next generation radar systems need to have even higher performance in terms of linearity and phase noise. This is also true for next generation autonomous L4/L5 driving systems, which require an order-of-magnitude higher resolution, sub 1° to detect pedestrians from hundreds of meters away. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.