The UC San Diego Center for Wireless Communications (CWC) Research Review was all about connection: Connecting people, connecting industry and academia, connecting technology across various sectors, and ultimately, connecting the world.

The thoughtful and provocative daylong discussion featured talks by CWC researchers about new applications and user experiences made possible by the exciting new developments in connected technology, including circuits for 5G communications, mmWave, networking, wireless augmented and virtual reality, connected health, connected and autonomous Vehicles, the path to 6G networks and more.

CWC Director Sujit Dey, delivered opening remarks describing a broad sense of enthusiasm among researchers, industry executives and civic leaders for the hyper-connected world promised by the Internet of Things (IoT). But for IoT to become a reality, these stakeholders must work to develop innovative business models that will allow for a seamless transition to connected cities, connected transportation, connected health and beyond.

One example of the role that the CWC is playing in this realm is its work to develop the sensor fusion, artificial intelligence (AI), computing and communications innovations necessary for connected vehicles. In collaboration with the Smart Transportation Innovation Program (STIP), a university-industry collaboration launched at UC San Diego last year, the researchers of the CWC are developing a collaborative Data Sharing and Predictive Intelligence Platform for applications enabling smart-safe-sustainable transportation and mobility. This platform—comprised of multi-modal data fusion, enhanced perception, intent analysis and predictive intelligence—also prioritizes security, privacy, trust, ethics, community-centered design and planning policies all based on an innovative computing and communications infrastructure.

“Twenty years back, computing and computing innovations were all about new kinds of computing devices, and new kinds of computing architectures” said Dey. “But especially in the last 10 years, computing has led to innovations everywhere."

“We are enabling live-streaming of data from various kinds of sensors: Involved vehicle sensors, street IoTs, pedestrians and EV charging stations and grids,” he continued. “Imagine there is live streaming of all this data going on, live data fusion, data intelligence and predictive intelligence to enable a lot of different applications, in the mobility space and beyond.”

This wide array of topics discussed hints at the difficulties inherent to developing autonomous vehicles, from the computer perception challenges created by rainy, foggy, smoky or dusty driving conditions (cameras and LiDAR sensors, for example, only work in good weather) to the dangers of inaccurate maps, poor GPS access in urban canyons and problematic merging conditions.

Yet for every one of these challenges, CWC researchers and their industry partners are investigating potential solutions, such as creating novel signal processing applications and replacing cameras or LiDAR sensors with multiple radars combined with learning algorithms to create a log-likelihood point cloud. This provides for much more accurate and higher coverage perception from within the vehicle, allowing an autonomous vehicle to operate in all weather.

The CWC-STIP Concept Vehicle is equipped with various cameras, edge computing nodes, and different kinds of radios and radars.

“We are not only creating concept vehicles that deploy various cameras, but also edge computing nodes, different kinds of radios and radars,” added Dey. “We are developing a massively distributed, hierarchical cloud computing infrastructure and also very novel communications techniques called hybrid V2x—bringing together Sub 6 and millimeter wave communications to enable a new kind of V2x to support this massive data sharing,” explained Dey. “All this live data sharing and live data fusion needs support from very agile computing and communications. In about nine months of efforts, we have acquired two vehicles, and now these concept vehicles are open to our member companies to play with to collect a lot of real data.”

Dey said that the hybrid computing paradigm being developed by the CWC can be successful for various other applications “community-wide and city-wide, including in the medical space, smart cities and so on.” Noting that the CWC is partnering with AT&T and a new UC San Diego Department of Urban Studies and Urban Planning, Dey added that “we are not just doing these simulations in the labs—we want to actually deploy this in the cities, go to the communities and really make this work.

“Our plan is to build a community of developers who will work on this platform and create applications we cannot imagine today.”

Some of those applications were described in a session at the Research Review on “Applications of 5G and Beyond: Use of Wireless Devices, AI, and Edge Computing for Connected Health and Wireless VR.”

Connected Health has been a focus of the CWC for a number of years, supported by a number of grants from the National Science Foundation and the National Institutes for Health and collaborations with Scripps Research, Teradata, FitBit, Kaiser Permanente and Samsung Digital Health.

The CWC’s focus on IoMTs (Internet of Medical Things), for example, enables personalized and predictive chronic care for patients by connecting vital data, behavioral data and contextual data in real time, all the time.

“By using that data,” explains Dey, “we are essentially trying to do two things. One is that we want to predict target variables, such as blood pressure, and we also want to figure out what might be the important features that are contributing to your blood pressure… If I can give you that input, maybe you can change (your lifestyle), and your blood pressure might be under control. Our early trial results look promising”

To study this particular example, CWC researchers equipped volunteers with FitBits to create a predictive model for blood pressure vis-a-vis various lifestyle factors such as heart rate, walking and running distance, walking and running speed, sleep duration etc. The next step was to launch a clinical study with the help of UC San Diego Health, the UC San Diego Altman Clinical and Translational Research Institute and the Samsung Galaxy watch, which is equipped with a PPG sensor—a trial that is still ongoing.

They then coded the model created with FitBit to real patient data, which, according to Dey, proved to be challenging given the fluctuating quality of patient data. Nonetheless, the resulting error rating of the resulting predictive model was comparable to that of existing medical devices, and made it possible to provide the patients with the top factors impacting their blood pressure.

“We are encouraged, and we are hoping that in the next three months we will learn much more,” said Dey. “We are trying to affect the lifestyle of these patients, and keep them off medication. That’s the whole purpose, and this is one example of where sensing, computation and communication can come in and really change lifestyles.”

MEDIA CONTACT

Tiffany Fox
tmfox@ucsd.edu