Several tech companies and car manufacturers, including Google, Uber, Tesla and Toyota, have invested tremendously in the realisation of self-driving cars. We have witnessed impressive progress in recent years following the release of partially autonomous vehicles and the successful testing of fully autonomous prototypes.

Self-driving cars promise to make the roads safer and more efficient. The idea of fully autonomous vehicles is set to revolutionise both personal mobility and commercial transport. The question is, why is it taking so long to put driverless cars into mass production? As it turns out, there are some serious obstacles keeping fleets of self-driving cars off the roads.

Reliability of current sensor technology

Self-driving cars rely on a system of sophisticated sensors such as lasers, cameras, LiDars and radars to collect information and navigate. These sensors are not 100% reliable on poor roads and adverse weather conditions. The sensors struggle in navigating accurately through heavy rain, snow, thick fog, and poorly marked routes.

The pioneers of autonomous vehicles admit that this is a difficult challenge to overcome. The current solution is to have a human driver take over control of the vehicle once the sensors begin to fail.

Interpretation of various traffic rules and signals

Road signs and traffic rules are complicated enough for human drivers as it is. What’s more, every country has a unique traffic system. Self-driving cars have to learn and apply these laws and signs before they can go on the road. Currently, the machine learning algorithms in autonomous vehicles have not had enough time to interpret various signs, especially verbal and non-verbal signals.

This problem also brings up the issue of legal liability in cases of accidents involving self-driving cars. Judicial systems are still working on fair legal resolutions for such cases.

Ethical decisions

Decision making is a complex problem when it comes to automated technology. Self-driving cars will have to be decisive when presented with multiple bad choices. For instance, if a pedestrian crosses the road in front of a speeding car, does it swerve into oncoming traffic, crash into the curb, or try to stop? These actions affect the safety of passengers and other road users as well. It’s difficult even to imagine the complexity of an artificial brain capable of ethically-acceptable split-second decisions involving life and death situations.

Current trends in autonomous vehicles give us a glimpse into the future of self-driving cars. But, fully autonomous cars still have sizeable milestones ahead of them before the public can legally and socially accept their usage.