The transition from the current situation of assisted-driving features to fully self-driving cars will require several steps along the way.

Transition to autonomous vehicles will go through several stages

Although plenty of ink has been spilled on the subject of the future self-driving automobile, we do not hear quite so much about what such vehicles will look like, what benefits they will bring, the challenges this new approach to road transport will throw up and the details of the necessary transition. It is clear that the switchover to 100% autonomous vehicles will not take place overnight.  We will probably have to go through various stages during which cars will become ever more automated and intelligent, but will still require the presence of a human driver at the controls.

In an opinion piece published on the online news site re/code, Matt Johnson, Vice-President and General Manager of the Automotive MCU Group at Austin, Texas-based Freescale Semiconductor, underlines a number of key caveats and makes a set of predictions for the transition.  He reckons that the advent of the 100% autonomous vehicle will take place over three phases, during which the Advanced Driver Assistance Systems (ADAS) designed to provide the driver with better road information, intelligent feedback and smarter physical aids will be progressively refined and eventually perfected.

The first stage, which Matt Johnson calls the ‘Assist Phase’, is already underway. New cars being rolled out today offer the driver a safety net – deploying cameras, sensors and alerts – against driving slips.  Today’s ADAS are able to cover ‘blind spots’, help to maintain a safe distance from other vehicles, alert the driver to imminent risks, etc. All these technological advances help to make driving safer but still require a human being at the wheel to implement the advice provided.

The second, ‘Automate Phase’ will be underway by around 2020. This period will see the advent of road vehicles capable of taking over all the driving tasks, with the human driver playing a co-pilot’s role, i.e. remaining alert enough to intervene where necessary. These vehicles will be equipped to collect – through sensors, cameras and radar systems – a huge array of data on the ambient environment and then process and act immediately upon the information.

The third and final stage, dubbed the ‘Autonomous Phase’, will not kick off until around 2030, bringing totally independent vehicles that do not require any human intervention at all and are capable of detecting and handling any potential threats even before a human brain could become aware of them.

Facial expression recognition to aid driving safety

At the recent fourth annual RE.WORK Connect Summit in San Francisco, two companies unveiled innovative systems designed to provide vehicles with greater autonomy and connectivity. Modar Alaoui, founder and CEO of Eyeris, a Palo Alto-based firm which is developing facial emotion recognition and analysis technology, explained the potential benefits of the Eyeris system for connected vehicles.

Embedded in a vehicle, the system is able to video a person’s face and then instantly analyse the expression on it – i.e. fear, anger, amusement, etc. Quite often, argued Modar Alaoui, we are simply not aware of having such emotions while we are at the wheel of our car.  With this type of information on the driver’s state of mind, a smart vehicle will be in a better position to assist with his/her driving performance.

Take for instance a case where the driver’s facial expression reveals inattention, suggesting to the smart vehicle that s/he is not 100% focused on the road. The car might react in a variety of different ways, such as sending a little impulse to the steering wheel so as to jolt the driver out of his/her torpor.  Alaoui told the Re.Work audience that “80% of all collisions are due to either anger or lack of attention,” also pointing out that a person who turns his/her eyes away from the road for five seconds to send a text message is basically travelling blind for the length of a football pitch.

The emotion recognition data collected by Eyeris could also be used by the public authorities to help improve road traffic circulation. For example, if it is shown that a large number of car drivers are getting angry at a particular junction at a certain time of day, the local authorities might need to think about taking steps to alleviate the congestion there.

Putting cars in touch with parking lots

Steve Banfield, Chief Marketing Officer at INRIX, talked about parking issues. Among other services, INRIX provides real-time road traffic data to companies working in the automotive and new ICTs fields. In September the Washington State-based SaaS and DaaS startup acquired ParkMe, a Santa Monica-based fledgling firm offering ‘smart parking’ services, which means that INRIX now has access to a full range of data that can be used to guide car drivers not only throughout their road journey but also when it comes to parking the vehicle at destination.

“At the present time, 14% of all cars sold are connected cars and by 2020 this ratio should rise to 55%,” predicted Banfield. Focusing specifically on parking issues, he told the audience that ”30% of all urban traffic is caused by drivers looking for a place to park. They take on average 20 minutes to find a parking spot – which works out at over 50 hours per year. And all this generates extra CO2 emissions.”

To combat this problem, INRIX has developed a system for directly informing the vehicle of the number of parking slots available at various parking garages in the vicinity. Parking garages are coded according to how full they are using three colours, ranging from green – a large number of slots available – through orange to red (parking lot full).  This information enables the driver to head straight to the least congested lot, which also makes for a better overall distribution of vehicles in the various city car parks.

The entire question of urban parking is today being taken very seriously. L’Atelier recently reported on Luxe, a startup that provides an app-based car valet service designed to avoid users having to drive around for hours at their destination looking for a place to park.

By Guillaume Renouard