Major automakers are competing to come up with a technology which allows vehicles to drive itself. Some automakers are using sensors and some are using LiDAR technology for the self driving vehicles to detect the environment surround them and make decision based on that. An accurate virtual map also takes an important role in order to let the autonomous system know which way that it needs to take in order to bring its passengers to their destination.
A team at the University of Cambridge has invented a new system named SegNet which is designed to analyse the road and its various features such as street signs, road markers, people and even the sky. Based on the official release, the system takes an image of the street in RGB mode and classifies it accordingly in real time. There are a total of 12 different categories, such as roads, street signs, pedestrians, buildings and cyclists. On top of that, it also can differentiate between light, shadow and night time environment.
SegNet is also able to allow the vehicle to detect an object and assess its location and orientation within a few metres and a few degrees. On top of that it does not require any wireless connection to analyse and report a position.
Professor Roberto Cipolla, the research leader said that in the short term, they were more likely to see this sort of system on a domestic robot such as a robotic vacuum cleaner, for instance. He also added that it would take time before drivers could fully trust an autonomous car, but the more effective and accurate they could make these technologies, the closer they were to the widespread adoption of driverless cars and other types of autonomous robotics.