A camera that can see around corners and reveal hidden objects could be put to real-world use in as little as four years, according to the Morgridge Institute for Research. The technology was first developed at the Massachusetts Institute of Technology and demonstrated back in 2012. Apparently, it caught the interest of the U.S. Defense Department Advanced Research Projects Agency (DARPA), which just provided a $4.4 million grant to fast-track research and development of the camera.
Andreas Velten pioneered the technology at MIT, and now leads a team of researchers at Morgridge and the University of Wisconsin-Madison, where both institutions are working together to optimize it. The DARPA grant will allow them to explore the limitations of the technology and hopefully develop fully-functional hardware that can be used in the field. UW-Madison is one of eight universities to receive 2016 DAPRA grants to investigate different forms and applications of non-line-of-sight imaging.
The camera works by shining a rapid pulse of laser light into a room. When the laser hits a wall or ceiling, the light scatters, bouncing off of other surfaces and objects. Many photons eventually make it back to the camera, arriving at different times depending on how far they’ve traveled. The camera can then use that information to digitally reconstruct an image of the room, including any people that may be inside of it.
Recording the photons that bounce back requires highly specialized equipment, so don’t expect this technology to come to a cameraphone near you. The sensor must be extremely sensitive, especially if the camera is to be able to see around multiple walls, a goal the team is currently working toward.
“The more times you can bounce this light within a scene, the more possible data you can collect,” Velten said. “Since the first light is the strongest, and each proceeding bounce gets weaker and weaker, the sensor has to be sensitive enough to capture even a few photons of light.”
The DARPA grant will fund four years of research, with the first two dedicated to investigating the full potential of the technology. The second two years will be spent developing the hardware, making it viable for production and implementation.
Non-line-of-sight imaging has numerous applications, from search and rescue, to exploring shipwrecks, and potentially even mapping moon caves. It won’t, however, have any effect on standard photography. “We are interested in capturing exactly what a traditional camera doesn’t capture,” Velten said.
San Francisco startup Dimensional advertised for people to walk robot dogs through city streets. The New York Times reported that the job could include carrying a roughly 33-pound robot back to the office if it overheated. A lift home is an awkward requirement for a labor-saving machine.
Keeping the hardware running is only one challenge. Anthropic's robotics tests found that language models mostly failed when controlling joints directly, but did better when directing trained controllers. Giving an AI agent a body doesn't teach it to handle ordinary household objects.
Soon, a mattress protector may be more than a bed cover. Australian medtech company Sleeptite and researchers at RMIT University have received a US patent for FlexiSense, a smart bedding technology that tracks body position and posture to help caregivers identify when older residents may be at risk of falling.
The technology is being developed for Sleeptite’s in-bed monitoring system, REMi, and was recently on display at The MedTech Conference in Boston. A peer-reviewed study in Flexible and Printed Electronics says the sensors can also be made at scale without sacrificing performance or durability.
AI companies spend enormous amounts of time convincing us their technology is safe and secure. But what happens when those same companies become targets of increasingly sophisticated AI-powered attacks? Anthropic, the company behind Claude, is now facing that question. According to a Washington Post report, cybersecurity startup Cathedral discovered vulnerabilities that let its researchers access sensitive information in Anthropic's internal systems.
The researchers used AI agents to identify potential entry points in less than 12 hours. Interestingly, they didn't need to crack Claude itself. Instead, they exploited weaknesses in third-party business applications that Anthropic uses for everyday operations. Anthropic's security testing program disclosed the findings, which allows outside researchers to identify and report vulnerabilities.