Google has just published a set of statistics on the tests it has been carrying out on its self-driving cars. The report points to areas where improvements are still required.
Nowadays many people are becoming quite keen on the idea of autonomous road vehicles. However, the reliability of self-driving cars (SDCs), will need to be fully proven before the general public are allowed to use them routinely on open roads and city streets. For over a year Google has been testing its SDCs – always with a specially trained driver in readiness to take over the controls if a problem arises – on public roads in the United States, in order to find out just how efficient the technology is. The Mountain View digital giant has now published a report entitled ‘Google Self-Driving Car Testing Report on Disengagements of Autonomous Mode’ – disengagements meaning either ‟when a failure of the autonomous technology is detected”, or ‟when the safe operation of the vehicle requires that the test driver disengage the autonomous mode and take immediate manual control of the vehicle.”
The report, which covers the period from 24 September 2014 to 30 November 2015, reveals that over the period Google’s SDC fleet experienced 272 disengagements due to technology ‘failures’. These incidents could be for example a communication failure between system and backup or an anomaly in sensor readings.
During the same period, there were 69 reported disengagements where it was deemed that safe operation of the vehicle actually required the driver to take immediate control, 13 of which were to avoid an imminent collision. In total Google has now operated its self-driving cars in autonomous mode for more than 1.3 million miles, 424,331 of them running on public roads in California.
The statistics show that at the moment self-driving car systems are not infallible, but Chris Urmson, who heads up Google's driverless car programme, points out that the number of incidents detected has decreased significantly over the test period. He underlines moreover that the advantage of SDCs is that the problems become quantifiable and solvable, whereas human performance is less predictable and human drivers will probably always be prone to making mistakes.
Self-driving cars were in the spotlight during the recent Consumer Electronics Show (CES) 2016, with inter alia the announcement of a partnership between Lyft and General Motors to build driverless taxis. However, the debate is still ongoing as to exactly what role autonomous road vehicles will play in our daily lives. Will everyone have his/her own self-driving car and use it in the traditional way to get from A to B, leaving it parked and unused at all other time, or will these cars be part of an efficient fleet of vehicles shared by all the residents in a particular area? This second option certainly seems to be the more sustainable approach for large urban conglomerations.