Controlling the car of the future

What would such a person make of air-bags, navigation, cruise control and rear-seat entertainment? The evolution which has led to these features is unlikely to slow down anytime soon. In fact, in certain areas it may be about to accelerate.

One of the areas which is going to exhibit rapid and radical evolution is the car interior, in particular the interfaces between the car systems and the occupants. The interior systems are evolving as the OEMs react to a number of factors acting on the basic idea of the car itself. These factors arise out of a number of external and internal influences: 

• Economics – cars need to be cheaper to produce and rich in features to sell well
• Energy issues – energy efficiency in the design and operation of cars is now a key selling point
• Consumer expectations on connectivity - these are being set by the mobile device market, not the automotive market
• Safety concerns – governments and insurers are making increasing demands to reduce accident and mortality rates associated with road use
• Congestion – city authorities require that road traffic is managed, potentially actively and intrusively, to allow cities to carry on functioning
• Market changes – non-western markets and associated cultures such as those in China and India are becoming the major target for growth in car sales.

For car occupants, the first and most visible impact of the influences described above will be the way information is presented and how the Human-Machine Interaction (HMI) occurs. As a result of these changing market and consumer demands, a number of new features, and systems to deliver them, will appear in the car interior in the next decade.

The challenge

Some of the features we will begin to see will include increasing connectivity to infrastructure to support security, traffic management and infotainment. As a result, commercial services for insurance, car maintenance and even advertising will appear in our cars. In-car media systems will offer multiple web and/or media streams to different parts of the car. Media will be available for playback in the car from mobile devices, in-car media support or from connectivity.

We should also expect to see systems in place to impose or coach eco-driving as standard. In the case of electric cars, this will range from systems for locating, scheduling and booking access, through to charging infrastructure. Such systems may include intelligent and connected agents which propose alternative modes of transport as a better way to travel under difficult traffic conditions.

But with such evolutions, comes a major challenge. As more information is brought into the car, with more diversity of style, priority and urgency, interfacing to these features within the constraints of the vehicle will require new and innovative methods of interaction, and a new set of paradigms for the design of in-car HMI.

The challenge here has multiple facets. Firstly, much of the information being presented is relevant to the driver who is occupied with a critical task, that of safely operating the car. Secondly, the safety constraints of vehicle design preclude certain styles of technology; it is not acceptable that a mouse or a keyboard can move freely in space as the result of a rapid deceleration, during an accident for example. Thirdly, any design approach has to support cultural localisation in a manner which is economically viable; the basic design paradigm should be cheaply customisable for Delhi and London alike.

OEMs and their prime suppliers (generally referred to as first tiers) have responded to this challenge with a number of concepts which are either on track for market introduction or currently under investigation. These concepts cover physical elements to be installed in the car and HMI design paradigms to be implemented on the physical support.

Physical supports for in-car human machine interface

When we talk about physical supports for HMI we mean devices that support output from the car's system to the user, and input from the user to the system. Currently an average car will have an interface based on the use of digital gauges, screens, speakers, buttons and rotaries. In more up-market cars we can also find touch screens, haptic joysticks, microphones and head-up displays.

The major goals driving the next generation of such devices are to provide more visual and audio real-estate, to make more use of haptic modes (force feedback), to avoid specific devices for specific functions and to be strongly culture neutral.

The most promising candidate for providing more visual real-estate is the technology known as 'under-the-skin'. This technology will transform the whole face of the dashboard, including the passenger side, into one display. The technology is based on the rear-projection of a display onto an appropriate material, which is actually the 'skin' of the dashboard itself; the user sees a display appearing through the plastic of the dashboard. This approach is much cheaper, and safer, than trying to integrate one very large or several smaller traditional displays. This technology may mean the end of the current gauges plus display design paradigm. The gauges will be represented appropriately 'under-the-skin'.

Alongside the 'under-the-skin' display there is the appearance of 2D gesture recognition as an input device. This implies that the system can detect a gesture made by the end user in a 2D plane. This detection is achieved by a combination of an infra-red illumination and an infra-red camera. The natural configuration here is to position the 2D recognition plane on the surface of an 'under-the-skin' display. In this way, OEMs can obtain a versatile touch sensitive surface at a reduced cost and with a very large surface. This approach is likely to become a standard for car interfaces. 

Haptic devices, such as the iDrive found in BMWs, offer the possibility to combine input and certain elements of output in one device, with the additional advantage of not requiring visual attention. For the driver this would appear to be a good interaction device. We can expect to see more of these kinds of devices in various formats (rotary, joystick).

In addition to touch interfaces, manufacturers are behind to explore the use of audio for both input and output in middle to high-end cars. Satellite navigation typically provides audio guidance, alongside icon and map display. In future models, this approach will be enhanced by the use of spatially localised audio fields in the car. This means that a user will hear a different audio stream depending on where he or she is positioned in the car. This will support the infotainment system but will also allow a driver to have an audio stream customised to the task of driving, which will not interfere with the audio streams of other car occupants who may wish to listen to other content, like the radio.

The input side of audio is speech recognition; the ability to issue commands to the system by voice. This has long been an area of intensive investigation and product development in the automotive industry. On the face of it, this seems to be a technology of choice for the driver and perhaps the other car occupants.  However, experience and research indicates that the idea is not yet proven. In order for an end user to feel comfortable with speech recognition, particularly when driving, it is essential that the system is highly accurate with recognition rates of between 80 or 90 per cent.

More fundamentally, there is a question mark over the cognitive load for a human using speech commands. For the driver, the effort required to interact with the in-car systems should be minimised and speech interaction is yet to be proven to do this. Despite this, speech input for cars will continue to develop and will improve as more processing power is applied and the technology matures. It should be expected that speech will be used in a constrained manner and perhaps in conjunction with other modes, such as 2D gesture recognition.

Human-machine interface design paradigms

So, given the appearance of new physical supports, what can we expect to see for in-car HMI in the near future?

The appearance of the large 'under-the-skin' display coupled with 2D gesture recognition will probably become standard in the era of adaptive HMI. This approach recognises that the amount of information to be presented to the driver during a trip cannot be achieved through static gauges and display paradigm. The HMI will reconfigure based on the nature and priority of the data that the driver needs. For instance, when a car approaches a junction the display may displace or remove the gauges and present a map display with hazard indications and speed advisories. In some designs, such high priority information may be accompanied by an audio stream within the driver's local audio field. On leaving the junction the display would reconfigure to show the conventional gauges.

Navigation may begin to function in a similar manner, with guidance being presented to the driver only when active. With the use of speech, the driver will be able to set up the address using direct speech as opposed to the somewhat cumbersome keypads currently in vogue.

The elements of the in-vehicle display not associated with the driver will present a wide range of windows, icons and stylised widgets to the front seat passenger. This display will be fully configurable by gestures that will allow widgets to be opened, re-positioned and closed at will. In order to cope with the vibration associated with in-car surfaces, it is likely that the gestures to be used in manipulating the interface will be large, involving wide sweeping movements and long ranges.  It is also likely that interactions involving text entry, such as entering a web search term or writing an SMS, will be available through character recognition on the display surface. For the rear-seat passengers, local displays or a version of the 'under-the-skin' display will provide a similar level of functionality.

The audio environment, customised to each passenger, will support the streaming of media such as music or film soundtracks. Equally, text-to-speech will allow passengers to read emails and SMS. Local audio fields will allow several different streams to be run simultaneously with little or no audio interference across them.

All of these evolutions are just the beginning of another round of change to the car as we know it. We may even see changes to the driving controls themselves; joysticks and fly-by-wire could soon make their appearance. Like the 1960s Ford Anglia owner, the BMW Series 5 owner of today will look at cars of 2020 and wonder where the ignition key is.

Further information:

www.altran-praxis.com

Related forum discussions

	To start a discussion topic about this article, please log in or register.