Waymo is one of the companies at the forefront of autonomous car technology and isn’t afraid to brag about it, as evidenced by this video.
In the Alphabet portfolio of companies, which also owns Google, it doesn’t manufacture self-driving vehicles. Waymo creates the artificial intelligence that guides the machine, converting such cars as a Chrysler Pacifica Hybrid and a Toyota Prius for its fleet.
There are so many different scenarios an autonomous car encounters and has to safely navigate, it’s pretty mind boggling. And that’s just dealing with other vehicles on the road. What about pedestrians and cyclists, two separate entities that can be equally, if not more so, be unpredictable in their actions?
On its blog, Waymo released a video showing one of its self-driving cars recognizing a cyclist and anticipating the safe action based on the circumstance. In this case, the cyclist had to shift out of the bike lane into traffic to get around a parked trailer. The Waymo car predicted the move, slowed to give the cyclist the right-of-way to finish the maneuver and then resumed to the flow of traffic.
In order for autonomous cars to make safe decisions, they need to identify which type of road objects they’re encountering, whether pothole, pedestrian or cyclist. Each has its own expected ways of “behaving”.
While a pothole doesn’t move, cyclists and pedestrians do. But, a pedestrian presents more of a moving target than a cyclist because she is prone to suddenly changing directions. In a previous demonstration, Waymo showed its vehicle correctly identifying children in a crosswalk and taking the appropriate measures.
In order for the public to have faith in self-driving cars, they need to trust a machine’s ability to make decisions based on identifying sometimes very subtle distinctions, like between a cyclist and a pedestrian. Waymo has shown its systems are up to this task.