BBefore America had seatbelts or the Interstate Highway System, before Elon Musk was even a wink in Silicon Valley, the dream of fully autonomous vehicles revved up its engines. General Motors unveiled its Futurama exhibit at the 1939 World’s Fair, conveying visions of a utopian future. Tens of thousands of eager visitors queued up to drive around a model of the city of the future that envisioned a world of 14-lane freeways, flying vehicles and, you guessed it, self-driving cars.
We’ve come a long way over the past eight decades when it comes to autonomous vehicles. Tesla Autopilot, Ford BlueCruise and GM Super Cruise are commercially available cars with some autonomous functions such as steering, braking and acceleration under limited conditions. Waymo, which grew out of Google’s self-driving car project, currently has self-driving taxis that can drive autonomously under certain circumstances. and Just do operates a fleet of self-propelled semitrailer trucks along supply routes in some regions.
But these caveats—under restricted conditions, in certain circumstances, in some regions– are signs that fully autonomous vehicles are not yet a reality. There’s still a long way to go before consumers can nap or watch a blockbuster on their evening commute. Despite decades of advancement, autonomous vehicles are still not safe enough for the general public. Federal regulations designed specifically for autonomous passenger vehicles were only introduced this year. Waymo, Teslaand TruSimple Vehicles have all been involved in accidents, worrying consumers. While elements of automation promise to reduce human error leading to injury or death on the road, the technology isn’t quite there yet.
The companies like that Spark.ai, Quantum Computing, Inc.and Ephel come to bridge the gap between the dream of autonomous vehicles and reality.
“The world is complex and constantly changing,” he says Michael KohenFounder and CEO of Spark.ai, based in New York, which emerged from Zoox, the autonomous vehicle company acquired by Amazon. “It’s really difficult to build an AI system that works 100 percent well. You wouldn’t get in a self-driving car that’s 95 percent accurate. That last five percent is critical for most autonomous applications.”
Because AI lacks cognition—the human ability to use common sense, intuition, prior experience, and learning to use split-second decisions—self-driving cars continue to encounter situations they don’t understand. This confusion can lead to deadlocks and delays at best and failures or malfunctions at worst. Spark.ai solves this problem by making real-time human cognition available to AI.
Suppose an autonomous vehicle encounters an unfamiliar stretch of road. It recognizes orange flags and cones everywhere, people waving their arms in the middle of the street, and unfamiliar signs. The car is confused and does not know how to behave in this situation. With Spark.ai’s technology, the vehicle can call a human to assess the situation. This mission specialist can identify the car in a construction site in a matter of seconds and determine the safest way forward.
Not only self-driving cars can benefit from occasionally putting people in the driver’s seat again. Kohen believes that empowering artificial intelligence with human cognition can be a game changer for any industry trying to overcome plateaus in their AI.
One of the areas where autonomous vehicles outperform human drivers is in their superior senses. Today, autonomous vehicles are equipped with an array of cameras, radar, and LiDAR (Light Detection and Ranging) sensors that enable them to navigate in conditions that would be difficult for humans, such as driving through a fog bank or navigating pitch black darkness highway. However, the number of sensors required for an autonomous vehicle to fully cover road conditions and have full situational awareness has been inefficient and costly.