If you live in certain neighborhoods in San Jose, California, you can now get some of your groceries delivered by a self-driving car. Bright green cars from the self-driving startup AutoX–with a backup driver inside–will pull up with your order in the trunk. In the backseat, the window rolls down to offer a shelf of extra food to buy.
“Last mile delivery for grocery and food will be faster and cheaper enabled by self-driving cars,” says Jianxiong Xiao, CEO of AutoX. The vehicle is equipped with high-resolution cameras to spot pedestrians or pets in the road–a system that Xiao, who trained at MIT’s Computer Science and Artificial Intelligence Lab, says works better than Lidar laser arrays and other sensors used on other self-driving cars, but is cheaper to build. Fully outfitted, the car costs $80,000 and can operate for several years; if regulations and technology get to the point that the backup driver isn’t necessary, that will be cheaper than a regular car with a human driver.
While other autonomous delivery companies like Nuro, with a cute toaster-shaped van, or Marble, with a sidewalk delivery robot, are designing vehicles customized for delivery, AutoX is adding its technology to existing cars. “We are focusing on providing a better self-driving car service rather than building a car,” says Xiao. “What we are building is an AI driver, not a car.” That helps keep cost down, he says, and a regular vehicle can carry multiple orders and has enough room for equipment to control the temperature of the food.
A handful of states, including California, Arizona, Texas, and Nevada, have set up or are in the process of setting up legislation to support tests and deployment of fully unmanned vehicles. Xiao believes that it will only take one or two years to commercialize fully unmanned delivery in some locations.