Riders help develop autonomous cars

One of the biggest problems with the development of autonomous cars is the detection of small and vulnerable motorcyclists in traffic.

One of the biggest problems with the development of autonomous cars is the detection of small and vulnerable motorcyclists in traffic.

So the San Francisco chapter of the Iron Order Motorcycle Club recently volunteered the services of seven members to help develop Aurora Driver technology.

They spent a day riding around with cars driven in manual mode by testers from Californian autonomous vehicle company Aurora.

Some of the unique situations tested were detecting riders who were lane splitting, which is only legal in California in the USA, and stopping in an offset position, rather than right in front of a car.

The bikes included four Harley-Davidsons, a KTM, an Indian and a Yamaha cruiser.

One of the biggest problems with the development of autonomous cars is the detection of small and vulnerable motorcyclists in traffic.
Volunteers (Images: Aurora)

Autonomous cars

Detecting motorcycles is a bit of a headache for autonomous cars development.

 

In one incident in San Francisco last year a lane filtering rider was hit by a Chevrolet Bolt electric vehicle being driven in autonomous mode.

And police had the hide to blame the rider!

The Australian Motorcycle Council and other motorcycle representative groups around the world have called on authorities to slow down the testing and introduction of autonomous vehicles.

A motorcycle industry group in the USA called Give a Shift went so far as to say that “the single biggest threat to motorcycling overall (particularly in urban and higher density environments) will be the incompatibility between autonomous vehicles and existing motorcycles”.

However, motorcycle, car and tech companies such as Bosch are continuing to work together to develop systems that better identify and communicate with each other.

So while we still have grave concerns, it appears technology is starting to find solutions that just might make us safer.

In fact, BMW Motorrad spokesman Karl Viktor Schaller declares they will make riding safer because autonomous vehicles (AVs) will be virtually crashproof.

Tesla Autopolit

Electric car company Tesla has launched its Version 9 software update to Autopilot 2.0+ hardware that has a more advanced “neural net” to detect smaller and faster-moving objects around the vehicle.

That is supposed to include lane-filtering motorcycles.

However, Tesla Model 3 owner Scott Kubo shot this video to show the difficulties it has detecting motorcycles.

There are several examples of lane-splitting riders in LA travelling at much higher speeds than the 30km/h lane filtering maximum in Australia.

The system struggles at times to detect them both day and night.

Tesla detects lane filtering riders
Tesla detects lane filtering rider in video, but not on the sidescreen graphic.

In some cases it mistakes a bike for a car and in others the bike is in the next lane and passing right through cars!

Scott says drivers also use their ears to detect motorcycles and suggests an audio sensor to help the camera and radar sensors.

However, the bikes — including a couple of loud Harleys — are only audible for the last couple of seconds as they are coming from behind.

With the crackdown on exhaust noise and the coming wave of quiet electric motorcycles, an audio sensor would probably be pointless.

2 Comments

  1. Who has clear stats on cars being the cause of motorcycle crashes and fatalities?
    The scholarly works get turgid, eg
    Crash characteristics and causal factors of motorcycle
    fatalities in Australia, M.R.Bambach1, R.H.Grzebieta1, R.Tebecis1 and R.Friswell1
    Transport and Road Safety (TARS) Research, University of New South Wales, Australia.
    The definitive stats of drivers pulling in front of bikes 44% is useful but what of the rest. And by state and area perhaps

Leave a Reply

Your email address will not be published. Required fields are marked *

 

This site uses Akismet to reduce spam. Learn how your comment data is processed.