COVID-19 Updates

Information is available for the campus community on the Return to Learn website. Please get vaccinated and stay up to date with County and State guidelines as well as CDC recommendations.

Improved maps for self-driving vehicles win Research Expo 2021

May 25, 2021-- From monitoring the structural integrity of airplane wings, to improving lithographic 3D printing processes, to creating better maps for self-driving cars, this year’s Research Expo symposium showcased the depth and breadth of work done by graduate engineering and computer science students at the University of California San Diego Jacobs School of Engineering.

Computer science Ph.D. students David Paz-Ruiz and Hengyuan Zhang took the top prize, the Lea Rudee Outstanding Poster Award, in addition to the Best Poster Award for the Department of Computer Science and Engineering. Their work focuses on improving the maps that self-driving vehicles use to navigate the world. They worked under the direction of Henrik Christensen, director of UC San Diego’s Contextual Robotics Institute and a professor in the computer science department.

David Paz-Ruiz Hengyuan Zhang

Paz-Ruiz and Zhang are tackling a problem that makes it hard for self-driving cars to scale up: most rely on high-definition (HD) maps, with resolution as small as one centimeter, which are manually labeled.

The computer scientists instead created what is known as semantic maps--2D data based on camera images overlapping with point cloud maps. The maps contain information about drivable areas, crosswalks, sidewalks, and lane markings. 

“Unlike HD maps, our maps do not include centimeter level annotations for the trajectories that the vehicle can follow,” Paz-Ruiz said. “Instead, we seek to generate these in real time.” 

“Our semantic maps are generated without human labeling effort and can be continuously updated if the environment was to change in the future,” Zhang added. 

Researchers can build a semantic map of campus in just a few hours. By contrast, an HD map would take weeks. In addition, the researchers’ maps can be used across different robots and different vehicles. 

“At UCSD we are studying how smart vehicles can be used in urban areas such the university campus or a parking lot at the airport to assist people with daily tasks” adds Prof. Christensen 

This year, the 39th annual Research Expo event was held virtually, with students submitting video presentations ahead of time, and answering questions about their work during the live event from judges, alumni and industry representatives. In all, 250 people attended the event. 

In addition to Paz-Ruiz and Zhang, one research project from each of the Jacobs School’s six departments was recognized. They are listed below, with links to their video presentations: 

Bioengineering

Sandhya Sihra

"A Heart-to-Heart Worth Having"

Computer Science and Engineering

David Paz and Hengyuan Zhang

"Towards Autonomous Vehicle Navigation without HD Maps"

Electrical and Computer Engineering

Ross Greer

“Trajectory Prediction in Autonomous Driving with a Lane Heading Auxiliary Loss”

Mechanical and Aerospace Engineering

Jessica Medrado

"Fundamentals of water-evaporation processes for renewable energy application"

NanoEngineering

David Wirth and Justin Hochberg

"A Highly Expandable Foam for Lithographic 3D Printing"

Structural Engineering

Adrielly Hokama Razzini

"Structural Health Monitoring of an Airplane’s Wings Using Gaussian Process Regressors"


 

Media Contacts

Ioana Patringenaru
Jacobs School of Engineering
858-822-0899
ipatrin@ucsd.edu