Ask any car engineer what's the biggest variable in achieving fuel economy targets, and he'll tell you "the driver." If one human can't understand human driving behavior enough to be certain about an innocuous number like miles per gallon, how is an autonomous car supposed to figure out what hundreds of other drivers are going to do in the course of a day? Ford has enlisted the help of Stanford and the Massachusetts Institute of Technology to find out.
Starting with the automated Fusion Hybrid introduced in December, MIT will be developing algorithms that driverless cars can use to "predict actions of other vehicles and pedestrians" and objects within the three-dimensional map provided by its four LIDAR sensors.
The Stanford team will research how to extend the 'vision' of that LIDAR array beyond obstructions while driving, analogous to the way a driver uses the entire width of a lane to see what's ahead of a larger vehicle in front. Ford says it wants to "provide the vehicle with common sense" as part of its Blueprint for Mobility, preparing for an autonomous world from 2025 and beyond.
There's a press release with more information below.Permalink | Email this | Comments