As the world has seen with the recent Tesla Smart Summon foibles and other serious incidents involving autonomous vehicles, the technology behind driverless cars still needs a lot of improvement before it is put to widespread use. Now, researchers from MIT have developed technology that allows autonomous vehicles to effectively see around corners, Gizmodo reports.
The value in being able to see what can’t be seen is huge in just about any application. However, it’s especially important in the fledgling autonomous vehicle industry. Another team of MIT researchers recently developed a sort of X-ray device to see what human eyes can’t.
That device, however, needs a special camera, lasers, and other instruments. These features would drive up the price of an already expensive autonomous vehicle. So, MIT engineers went back to the drawing board and came up with something more cost-effective.
The new technology piggybacks on a previous system called ShadowCam. Rather than using X-ray technology or lasers, MIT’s new system uses video cameras. These focus on the ground in front of the car, specifically on the area where two roadways meet.
The cameras clock a series of video frames that are intelligently processed while the vehicle is in motion. This allows the system to detect shadows, heralding an approaching vehicle in real-time. Furthermore, the computer can also analyze the changes in lighting and predict the position and speed of the approaching object, whether it be another vehicle or a pedestrian popping out from behind a parked car.
The system is quick enough to then slow down or completely stop the car, hopefully avoiding a collision. Furthermore, the camera system proved faster than LIDAR-based systems by half a second. The MIT team plans to unveil its new system at the upcoming International Conference on Intelligent Robots and Systems in Macau, China.
While the new MIT system is ready for the conference floor, it might be a while before it hits vehicles and roadways. The team has only tested it in parking garages with the vehicle’s lights turned on to recreate nighttime driving conditions. The system will need a lot more tweaking before it can handle the complex lighting conditions of the outdoors.
Not Just for Cars
MIT’s new technology isn’t just for cars. The team also sees applications for the system in the medical field. In hospitals of the future, delivery robots may zip around chaotic hallways and patients may get from place to place with autonomous wheelchairs. The system could allow these robo-helpers to safely navigate hospital corridors. As such, the team also tested its system with autonomous wheelchairs that navigated through office hallways.
“For applications where robots are moving around environments with other moving objects or people, our method can give the robot an early warning that somebody is coming around the corner, so the vehicle can slow down, adapt its path, and prepare in advance to avoid a collision,” study co-author Daniela Rus said in a press release. “The big dream is to provide ‘X-ray vision’ of sorts to vehicles moving fast on the streets.”