New technology from Stanford University may make driverless cars even more advanced in their maneuvering skills. While self-driving cars have been in the industry for a while now, recent developments continue to make such automobiles even more intelligent than they already are. Stanford’s new imaging technology improves on the concept of self-driving, making them all the more competent in their capabilities.
New Tech Lets Driverless Cars See Objects Hidden From View
Essentially, Stanford’s tech allows self-driving cars to see objects that may not be in their peripheral. With this, such cars may have improved reception of hazards or other obstacles on the road.
For the tech to work, a laser is placed next to a highly sensitive photon detector, which will then shoot pulses of laser light at a wall. These pulses will bounce off objects around corners. After such, the pulses will bounce back to the wall and to the photon detector. Upon completion of such scans, the tech’s algorithm will straighten out paths of the captured photons to later create sharper and more distinguishable images.
Stanford explains that such a scan can span two minutes to an hour. The amount of time is dependent on external conditions, such as lighting and a hidden object’s reflectivity. The algorithm, on the other hand, is much quicker, and is efficient enough to run on a regular laptop. Researchers are currently looking into improving the speed of the algorithm, going so far as to aiming for a more instantaneous result when producing images.
Researchers at Stanford claim that such technology may be capable of performing adequately if placed on a car at this point in time. It could “easily detect things like road signs, safety vests or road markers.” However, the team also adds that the system may “struggle with a person wearing non-reflective clothing.” Stanford’s team is still improving on the system and adding tweaks so as to make their tech even more competent, especially when it comes to detecting other variables in the real world.
Other Driverless Cars
Most recently, this year’s Mobile World Congress unveiled the capabilities of the Huawei Mate 10 Pro’s artificial intelligence (AI). The Android phone was able to drive a Porsche Panamera all on its own thanks to its object recognition AI.
To make this work, an external camera was installed on the roof of the car so as to provide a means of surveying the road ahead. The camera was then connected via HDMI to the Mate 10 Pro, acting like a dash cam that receives video capture in real time. Upon transmission of footage, the phone’s AI with object recognition is utilized, followed by self-driving commands being conveyed to the vehicle.
It’s worth noting that despite such technology, Huawei insisted that it was not going to delve in the market of driverless cars. The project was simply meant to show off the Mate 10 Pro’s capabilities and the advancements of AI in this day and age.