Has Anki Solved The Autonomous Car Problem?
Opening the 2013 Worldwide Developer Conference (WWDC) for Apple developers was a keynote speech that featured a demo from Anki. Anki Drive is an iOS controlled racing game where autonomous cars drive around a track racing. These vehicles are approximately the size of Hot Wheels or Matchbox cars, and feature very little intelligence on them. They have sensors and low-powered Bluetooth sensors to locate each other, and then the iOS app performs all of the processing.
Despite a small hiccup during the demo, the vehicles took to the track. One of the presenters then introduced a rogue element to the pattern, and the other vehicles immediately become aware of its existence and then avoided it. Then the engineers ordered the other vehicles to attempt to block the faster vehicle from passing. Immediately, they used the location data being processed by the iOS application to begin blocking the car from passing. Finally, in typical Spy Hunter fashion, the rogue vehicle was provided with “machine guns” and shot the other cars off the road.
Google has been operating a fleet of autonomous vehicles for years, and they contain a sophisticated processing system to avoid collisions. However, Google has not yet developed an effective way for vehicles to communicate with each other, which is really the big hurdle in preventing accidents. The Anki vehicles, though on a much smaller scale, are aware of each others’ location. They then communicate that to the app. The app then coordinates all of the vehicles; mimicking the role of an air traffic controller.
It is conceivable that a cloud-based solution could be developed from this. New cars sold would contain the sensor equipment (which is minimal in this scenario), and would maintain a data connection to the cloud where the information would be processed. Many people probably looked at that demo and asked themselves, “Why is this company on stage demoing a child’s toy?” It is clear though Anki wants to pursue more advanced robotics control, and if executed properly, could have some serious influence over the future of autonomous vehicles.