X

Uber Resumes Self-Driving Tests In Pittsburgh After Fatal Crash

Uber has announced that it will restart the testing of its self-driving vehicle fleet, with the new public road trials set to take place in Pittsburgh and with the vehicles operating in manual mode. The ride-sharing company halted all of its autonomous driving tests across the country after a self-driving Uber vehicle hit and fatally injured a woman in Tempe, Arizona, last March. According to a preliminary investigation conducted by the National Transportation Safety Board (NTSB), the self-driving Volvo XC90 from Uber’s autonomous fleet saw a jaywalking woman six seconds before hitting her, raising concerns about the configuration Uber used for testing its self-driving vehicles in the state. It was also found that Uber disabled the SUV’s emergency brake so as to minimize the potential for erratic behavior of the vehicle.

While Uber decided to permanently shut down its self-driving operations in Arizona following the fatal incident, terminating the jobs of some 200 employees, the company reiterated that it won’t be exiting the self-driving segment following the incident. It was reported in May that the company’s autonomous driving operations would be restarted in Pittsburgh in the future as soon as the federal probe into the Tempe incident is concluded. Eric Meyhofer, Head of Uber Advanced Technologies Group, said in a recent Medium post that the ride-sharing company plans to redeploy self-driving vehicles in Pittsburgh with a backup driver sitting behind the wheel and controlling the vehicle hands-on while a passenger in the backseat records all notable circumstances.

Uber is also taking into consideration the things it learned from the recent fatal crash in order to maintain vehicle safety while testing its self-driving cars in Pittsburgh. Specifically, the startup will install a real-time driver-monitoring system in all of its autonomous vehicles to alert backup drivers and help them maintain focus. Uber’s self-driving fleet will also feature a native collision avoidance system that will stay on during the entire course of every autonomous trip. The system is designed to automatically brake under specific circumstances and warn the backup driver about an imminent collision. The goal of the testing is to create virtual simulation scenarios in real time to improve system performance and develop accurate HD maps meant to help the company’s vehicles navigate any environment.