In May this year, a robotics expert from Stanford University in Tesla’s home state of California warned that Autopilot should not be used in the vicinity of cyclists, warning that it would put lives at risk. road.cc/content/news/2…
Meanwhile in the UK, a Tesla driver has just killed a very experienced 80 year old cyclist. Unclear whether it was on autopilot at the time.
This matters, because according to a UCL study commissioned by the government, there is next-to-ZERO research into interaction of AVs with pedestrians & cyclists.
Here's UCL's research report for the government (pdf): gov.uk/government/upl…
Here's the key diagram. Out of 50,000 research papers, just 20 deal with 'interaction with other road users'. 
Here's the UCL team's warning to the Government of what needs to be done:
If we dig into the report, we find that out of 50,000 papers there are just 16 papers looking at "autonomous vehicles interaction with cyclists". Sixteen.
Pedestrians do better: there are 18 research papers looking at the interaction of autonomous vehicles with then. Out of 50,000 research papers.
So has the government taken notice of this warning that there is no research into interactions with pedestrians and cyclists in cities?
Not. A. Chance. Phillip Hammond has airily announced that autonomous vehicles will be on UK roads by 2021. Full stop.
So AVs will be coming to a city centre near you shortly. With almost no research into how they react in a crowded high street with lots of pedestrians and cyclists.
They'll be in high streets because most car journeys are between 3 and 5 miles.
So to sum up, an avalanche of money is pushing a new technology onto our streets with almost no research into interactions with vulnerable road users.
So an even better reason to build protected cycle tracks now and use modal filtering to eliminate through traffic in residential areas. /ends.
unroll
