• Fiat Chrysler Automobiles
  • Google
  • Fuel Efficiency and Emerging Technologies
  • Transportation Efficiencies
  • Autonomous Vehicles

Google and Fiat Chrysler Team Up to Build New Fleet of Self-Driving Minivans

Sam Abuelsamid
May 03, 2016

Electric Vehicle 2

For the first time since Google began work on developing autonomous vehicle technology 7 years ago, the company now has an official relationship with an existing automaker. The technology giant is teaming up with Fiat Chrysler Automobiles (FCA) to build an expanded test fleet to accumulate more real-world miles. Engineering teams from FCA and Google will be colocated at an undisclosed facility somewhere in southeast Michigan to develop and build vehicles based on the new Pacifica Hybrid minivan.

Until now, Google has largely worked independently on its self-driving car program, purchasing Toyota Priuses and Lexus RX450s and installing the sensors and computing hardware necessary to have the vehicles drive themselves. More recently, Google contracted with Michigan-based Roush Engineering to build dozens of dedicated self-driving pod vehicles, but these were strictly low-speed electric machines limited to a maximum speed of 25 mph.

Approaching the Automakers

Several automakers have acknowledged off the record that they had been approached by Google over the past several years, but the business conditions set by Google were unacceptable. Essentially, Google wanted a company to build cars and turn them over for installation of a black box control system. Since it’s generally acknowledged now that automakers will be liable for the reliability and performance of autonomous vehicles, no company was willing to cede that much control to Google.

The hiring of former Ford and Hyundai executive John Krafcik as CEO of the Google Self-Driving Cars division last fall likely led to a change in attitude in Mountain View about how to collaborate with incumbent automakers. There had been speculation in late 2015 that Google would announce a partnership with Ford as early as the 2016 CES in Las Vegas last January, but the show came and went without an announcement.

For more than a year, FCA CEO Sergio Marchionne has been looking for a partner to merge with. Marchionne made an especially hard push for a merger with General Motors, but was repeatedly turned away. More recently, Marchionne has publicly stated that FCA would make an excellent partner to manufacture vehicles for Apple should the electronics company decide to get into the automotive business. Given FCA’s limited resources relative to larger rivals in Detroit, Europe, and Asia, a partnership with Google is likely the company’s best course of action right now.

FCA Developments

Navigant Research’s Leaderboard Report: Autonomous Vehicle OEMs from last year ranked FCA 14th among 18 OEMs evaluated for their work on autonomous vehicles. FCA has never publicly discussed or demonstrated an autonomous vehicle program, although it has been surprisingly aggressive in deploying advanced driver assistance systems to its model lineup in the past 3 years. Back in the mid-1990s, Chrysler also developed robotic driving systems that could be used to control vehicles running on a particularly harsh durability test track at its Chelsea, Michigan proving ground. Human drivers could only withstand short periods of driving on the course because of the pounding and it was hoped that an autonomous system could be used to conduct accelerated durability tests. The system was not sufficiently reliable at that time and was eventually abandoned.

The new Chrysler Pacifica Hybrid was unveiled in January 2016 at the North American International Auto Show in Detroit and features FCA’s first in-house developed hybrid drivetrain. The plug-in hybrid features an 18 kWh lithium ion battery pack manufactured by LG Chem in Holland, Michigan and is capable of an approximately 30-mile all-electric driving range.

FCA and Google have not said when the new autonomous minivans will be ready for testing, but the 100 vehicle fleet will enable the two companies to significantly expand their collection of real-world data needed to make autonomous systems more robust.