July 17, 2002
JACOBS SCHOOL SHOWCASES MULTI-USE SENSOR NETWORKS ON CORONADO BRIDGE FOR HOMELAND SECURITY, STRUCTURAL AND SEISMIC MONITORING
San Diego, July 17, 2002—A multidisciplinary group of researchers from UCSD recently deployed an ad hoc network of cameras and other sensors on Coronado Bridge. They were linked via high-speed wireless to a makeshift control center in the Computer Vision and Robotics Research laboratory in the SERF building on the UCSD campus. The goal: demonstrate a single monitoring system that could be used simultaneously by academic researchers and agencies as diverse as Caltrans, SPAWAR and the U.S. Coast Guard.
The temporary installation in May was part of an ongoing effort at UCSD to focus sensor network research to respond to the heightened interest in technology for emergency preparedness, including homeland security. "It's really the multi-use of these sensors and the synchronized data transmission," said Frieder Seible, interim dean of the Jacobs School of Engineering and principal investigator on the project. "We can look at everything as a whole in a holistic approach--what's going on on the bridge, and really using it not just for dumb engineering data, but also for security monitoring of important structures."
The monitoring equipment was installed at Pier 14, on the bridge's understructure, roughly 200 feet above sea level. "The bridge is entirely wirelessly connected from Coronado Bridge to Mount Soledad and from there to UCSD," said Hans-Werner Braun, a senior researcher at the San Diego Supercomputer Center (SDSC) and director of the High Performance Wireless Research and Education Network (HPWREN). "On the bridge itself the wireless hub is surrounded by Ethernet connected devices."
Researchers from UCSD's Structural Engineering Department and the Scripps Institution of Oceanography deployed two seismic EpiSensors and two accelerometers to gauge the extent of movement on the bridge in the face of earthquakes--or even heavier-than-usual traffic. They also installed two digital high-resolution (3 megapixel) still cameras that are also viewable over the Internet. Researchers from the Jacobs School's Electrical and Computer Engineering department put up an omni-directional 360º camera and a standard video camera and web server. Meanwhile, SDSC researchers used off-the-shelf 802.11b high-speed wireless technology to get the seismic and visual data captured by the sensors to the HPWREN which then transmitted the data through its 45 megabit per second backbone to its facility on Mount Soledad 12 miles away. The data was then sent 3 miles to the supercomputer center, and from there to the control center inside the computer vision lab on campus.
The result: multiple streams of data from half a dozen sensors on the bridge were transmitted simultaneously and wirelessly at very high speeds. Researchers and invited guests were instantly able to view and aggregate different data sets over the Internet in real time. "We have at a remote site, multiple cameras, multiple sensors and structural monitoring sensors, all of them utilizing a high-bandwidth wireless backbone," said Mohan Trivedi, professor of electrical and computer engineering and the Cal-(IT)² layer leader in charge of "intelligent transportation" research at UCSD. "Meanwhile, at a completely different location, they all converge and we can have a control room type of experience."
Combined with televiewing—viewing from a remote location--Trivedi's omni-directional camera allows someone in the command center to zero in on a small craft a mile away, while someone else is using the same 360-degree video feed to monitor the bridge's own bearings. Computer vision and pattern recognition technology also allow a computer to automatically process what is visible and "watch" for possible danger signs:
Said Jacobs School dean Seible: "We can monitor and see formations and compare one image to the next image—this is pattern recognition—and when we see a change we can send an alarm."
"What Frieder Seible wants for the structural part are the same types of signals I want for seismic monitoring," noted Frank Vernon, a senior researcher at the Scripps Institution of Oceanography’s Institute of Geophysics and Planetary Physics. The data from seismic sensors on the bridge can also be integrated with the output of cameras, resulting according to Vernon, in "one data stream. Also integrated into the same data stream is photo imagery, taken every 15 seconds. This has never been done before. This is the first time we've ever tried these two types of imagery together. So, one, it gives us a view of how the pier is moving, and the other, what the bearing looked like."
The sensors and wireless technology worked flawlessly, effectively demonstrating the feasibility of multi-use, multimedia networked monitoring systems. Concluded HPWREN’s Braun: "If people from Caltrans or SPAWAR or other institutions determine there is real value, then all of us need to sit down and design a real system. At that time, wireless is probably just one piece of it."
The same cameras that maintain a security perimeter around Coronado Bridge and nearby Navy assets, could also be used to monitor the main shipping channel for the Coast Guard. Researchers also see extending the network with environmental sensors to monitor air and water quality for pollution as well as biological or chemical weapons. And the same type of cameras to monitor ship and structural activity could allow computers to "watch" and analyze traffic patterns on the bridge—and automatically alert a human operator in the event of a problem.
Now, investigators say, the next stage is to discuss the added value
of such systems, and if possible, produce an end-product that serves the
specific needs of various agencies while also fitting into UCSD’s own