NTSB: Uber’s AV in fatal crash ‘had software issues’

The US National Transportation Safety Board (NTSB) has found that an Uber autonomous vehicle which killed Elaine Herzberg last year had software flaws. NTSB released a report which says the Volvo XC60’s autonomous system software classified the pedestrian as an unknown object and determined that an emergency braking manoeuvre was needed to mitigate the collision. Uber confirmed that emergency braking manoeuvres must be carried out manually and the system is not designed to alert the driver. Data
November 6, 2019 Read time: 2 mins

The US 5628 National Transportation Safety Board (NTSB) has found that an 8336 Uber autonomous vehicle which %$Linker: 2 External <?xml version="1.0" encoding="utf-16"?><dictionary /> 0 0 0 link-external killed false https://www.itsinternational.com/categories/utc/news/uber-ends-self-driving-programme-in-arizona/ false false%> Elaine Herzberg last year had software flaws.

NTSB released a %$Linker: 2 External <?xml version="1.0" encoding="utf-16"?><dictionary /> 0 0 0 link-external report false https://dms.ntsb.gov/pubdms/search/hitlist.cfm?docketID=62978&amp;CurrentPage=2&amp;EndRow=30&amp;StartRow=16&amp;order=1&amp;sort=0&amp;TXTSEARCHT= false false%> which says the 609 Volvo XC60’s autonomous system software classified the pedestrian as an unknown object and determined that an emergency braking manoeuvre was needed to mitigate the collision.

Uber confirmed that emergency braking manoeuvres must be carried out manually and the system is not designed to alert the driver.

Data retrieved from the autonomous system revealed that the vehicle operator engaged the steering wheel less than a second before impact while the vehicle was travelling at 39mph.

The NTSB reviewed Uber’s autonomous system cameras which revealed that Herzberg did not look in the direction of the vehicle just before impact.

An inward-facing video showed the vehicle operator glancing down toward the centre of the vehicle several times before the crash. A post-crash interview with NTSB investigators revealed that she was monitoring the autonomous system’s interface.

The NTSB is working with Uber, Volvo Cars and the 6576 Arizona Department of Transportation to compile a complete account of the crash.

Uber is not the only company whose AV trials have led to a fatality. Last year, a Tesla vehicle operating in %$Linker: 2 External <?xml version="1.0" encoding="utf-16"?><dictionary /> 0 0 0 link-external autopilot false https://www.itsinternational.com/categories/utc/news/tesla-car-crash-in-california-kills-driver-while-running-on-autopilot/ false false%> in California caught fire after crashing into a roadside barrier that killed the driver. 

For more information on companies in this article

Related Content

  • ID TECH releases new MagStripe and smart card mobile reader
    November 20, 2013
    ID TECH’s new addition to its audio jack family of secure mobile card readers, UniPay, combines contact chip card reading and magnetic stripe reading in a single device. EMV Level 1 approved with AES and TDES encryption algorithms, and DUKPT key management, it ensures that consumers’ financial information is securely transmitted to the host device. Able to connect to various mobile devices such as Apple and Android platforms with an SDK, UniPay transforms any mobile device with an audio jack into a POS sy
    Read More
  • iMobility Challenge event
    November 12, 2014
    The next iMobility Challenge event, "Intelligent Mobility, Smart Cities" takes place in Brussels on 18 November 2014. Experts from innovation companies and researchers will share the latest ideas and developments in the field of intelligent mobility, sharing insights into how we will move forward in cities tomorrow, and what future vehicles will look like. Confirmed keynote speakers include John Baekelsmans, CTO at Cisco; other key speakers will be Rupert Fausset, Principal Sustainability Advisor at F
    Read More
  • Sensefields’ wireless sensors simplify sensing
    March 24, 2014
    Sensefields’ traffic monitoring system uses easily installed wireless sensors to determine vehicle speed and, in urban situations, also for categorisation. Information from the sensor is sent in real time to the data processing station to determine the capacity (vehicles per hour) in each lane, average speed, speed distribution, average vehicle length, length distribution, density, average headway between vehicles and occupancy (%).
    Read More
  • Meet RideScout
    September 10, 2014
    A new mobile app called RideScout provides Transit as a Service (TaaS) by consolidating multiple public, private and social transit options in a single platform.
    Read More

Highlighted Content