In case you haven't heard, self-driving (or 'autonomous', for a very specific value of the word) cars are the Next Big Thing. Every car-maker is working on one, and if they're not, they're looking at companies like Uber and Mobileye and Bosch who are. Ford has its own autonomy plans well underway, and the latest version of its self-driving Fusion Hybrid packs in a bunch more high-tech sensors to understand the world around it in real time.
The latest iteration of the autonomous test platform built onto a Fusion Hybrid chassis re-jigs the car's existing LIDAR distance-sensing setup, and uses two sleeker sensors with a tighter field of vision than the previous four -- while still drawing in an equal volume of data.
Three years in the making since the last significant update to the road-going prototype, the new Fusion Hybrid has much more processing power, allowing it to make decisions faster and with more granularity per second -- closer to the reaction times and flexibility of a human driver. Ford says there are five elements that make its new self-driving car SAE Level 4-capable:
- Sensors — LiDAR, cameras and radar
- Algorithms for localization and path planning
- Computer vision and machine learning
- Highly detailed 3D maps
- Computational and electronics horsepower to make it all work
All that technology is interpreted by what Ford calls a "virtual driver" -- the computer-driven processor inside the car that draws on all those data points -- and that makes decisions approaching the quality of what a human driver could or, in some instances, exceeding it. The computer, hidden in the trunk, requires more power than a regular car battery and alternator could produce, so relies on the Fusion Hybrid's larger high-voltage pack and a standalone inverter.
The entire background of Ford's autonomy project is detailed in a blog post on Medium by chief program engineer Chris Brewer -- it's a great read. The company plans to build a fully autonomous vehicle for sale by 2021. [Medium]