22 February 2019

Autonomous introduction

Developments in machine learning, AI systems and sensor technology are driving forward the widespread testing of autonomous vehicles. A number of automated safety systems have been integrated into our cars for some years now, whether that’s ABS braking, active parking assist, active cruise control, lane assist – or even the level of automated driving that a Tesla offers. There are a number of dashcam videos available across the internet, showing miraculous escapes from impending serious accidents that Tesla’s have avoided – showing their autonomous safety systems working to protect their occupants and save lives. There is however a lot more footage available showing Tesla owners having to intervene when the autonomous system has missed something – obviously there is still a lot more work to be done.

The California DMV releases a yearly report on the use of the States’ autonomous vehicles, recording how many miles have been driven, if any incidents occur and also how many times every single vehicle’s autonomous mode is disengaged by a test driver for safety, also stating for what reason it was disengaged. Every single operator in the State must submit this data. Should we then find it surprising that Tesla have submitted ‘Zero miles’ for the year? So, I don’t want to focus on Tesla, but it is interesting that the testing they are conducting is a little different to everyone else, as they conduct what they call ‘Shadow Testing’ - namely that they draw data from their fleet of privately owned vehicles already on the roads around the world, whilst also utilising both simulation data and testing they conduct at test tracks.

The value of this type of testing I feel cannot be overstated – real world autonomous testing, with a vehicle that has integrated with all other road users and for all intents and purposes looks like a normal car - by that I mean that it hasn’t got a visible roof rack of sensors and cameras and hasn’t got brand logo’s plastered all over it letting other road users know that it is an autonomous test vehicle and they should treat it a little differently to all other vehicles.

Autonomous vehicles have a number of big challenges ahead, and the reason I think this kind of testing is so important is that outside of convincing the public to place their trust in an autonomous car, one of the biggest challenges will be the safe navigation of the road systems whilst sharing them with regular drivers, who drive in very different ways and will likely drive differently around different road users. I am surprised we don’t see a lot more OEM’s using this form of data collection, however things are set to change with the introduction of 5G, smart infrastructure and an advanced connected mobility on the horizon.

Publisher: Shaun Hunter

Global News Editor: Trisha Chowdhury

Legal Affairs Editor: Julian Wilkins

Chief Executive: Peter Wooding

All rights reserved. No part of this publication may be reproduced or stored in a retrieval system without the written permission of the publishers. Whilst every care has been taken in compiling this publication, the publisher cannot accept responsibility for any inaccuracies or changes since going to press, or for consequential loss arising for such changes or inaccuracies, or for any other loss direct or consequential arising in connection with the information in this publication. The views expressed by the contributors are not necessarily also those of the publisher. E&EO

Main Switchboard: +44 (0) 203 325 4414