|
Under control
|
01/03/2007
|
| |
The information load on a driver continues to rise as assistance and entertainment features proliferate. Jeff Daniels looks at the problems of transferring this data safely
There was a time when the car driver faced no real problems of information or of control. Information was a matter of frequent glances at the speedometer and hopefully in the rear-view mirror, and less frequent ones at the fuel gauge and the coolant thermometer. As for controls, apart from the wheel, pedals and gearchange, no more was required than switches for the lights and wipers.
It was the controls that multiplied first. There came controls for the heating and ventilation system and for the car radio, for direction indicators, for the heated rear window, the rearguard foglamp, the hazard warning system. More complex systems came with driver override selections – of sport or comfort mode for automatic transmissions, for example. Systems became available in which cruising speed and vehicle separation distance could be preset. It became increasingly clear, at least to those who must drive a variety of cars rather than just one, that there was no real standardisation of control layout, but that some solutions were better than others. Even so, mastery of the control layout of most modern cars, especially upmarket models, requires time spent in study of the handbook and practice in control location and operation.
Somewhat later than the multiplication of controls came the burgeoning of information. Today, the driver of an up-market car is likely to have available navigation and traffic information, indication of parking space clearance, a ‘blind spot’ lane-change warning (such as Audi’s Side Assist), a lane-keeping indication, an outside air temperature indication, and a tyre pressure monitoring system – to say nothing of modern audio and air conditioning systems with digital readouts, and the seemingly ubiquitous in-car telephone, and a growing number of warning lights which may relate to anything from low oil pressure to the condition of the ABS or the passenger side airbag. There is no lack of further proposals, for example the replacement of mirrors by a rearwards-looking television system, and the enhancement of night vision by systems based on the processing of infrared images.
There are growing indications that some drivers have difficulty assimilating all the information with which they are presented, and operating the car’s controls in response, while still maintaining their primary task of surveying and assessing the topography of the road ahead and the traffic conditions around them. Studies have been published to show that carrying on a telephone conversation can degrade driving standards as much as having the legal limit of blood alcohol. Fewer studies seem to have been carried out on the effect of other complexities such as (perhaps) trying to reprogram a navigation system on the move, or reconciling the warnings from a proximity-warning system with the visual evidence. What is certainly evident is that there are few reliable pathways via which the vehicle can interface with the driver and vice versa.
Conventionally, the preferred pathway is visual. However, the volume of information now being presented increases the risk of visual overload - that some possibly vital cues will simply be overlooked, or at least, noticed too late. That has led researchers to study the other two available pathways - aural and haptic, relying on hearing or feel rather than sight. Aural instructions already complement on-screen presentation in most navigation systems and there have been many attempts through the years, certainly since the 1980s, to replace or supplement warning-light indications with aural warnings. More recently some haptic inputs have been tried, for example Citroen’s lane-keeping warning system which vibrates the driver’s seat cushion on the “danger” side.
Compared with the visual route, there are drawbacks to both alternatives. Audio instructions or warnings take longer to deliver, and there is evidence that drivers “shut off” their audio channel at moments of intense visual activity, giving priority to the latter. Haptic information devices are at an earlier stage of development. They may impart a greater sense of urgency, but clearly they need careful development if their messages are to be unambiguous, and they are likely always to be relatively expensive, needing some form of force transducer built into the driver’s seat, the steering wheel, or the pedals, the only reliable “haptic information pathways”.
When looked at from the reverse perspective of the driver exercising control, on the other hand, the haptic pathway is the conventional one: the driver reaches out and operates controls manually. The aural and (perhaps) visual alternatives are only now being explored. Several cars already embody voice-activated control operation of some features, but difficulties remain. Voice recognition may have moved on from its earliest and most hesitant days, but current systems still mostly require the driver to learn a standard vocabulary of commands and to speak clearly. Even then, problems can arise if a different driver tries to use the system. The easiest voice control function, and apparently the most widely appreciated among users, is hands-free telephone dialling. Indeed, in this context, the original meaning of the interface (between vehicle and driver) is now often translated as the vehicle’s ability to accommodate and integrate a mobile telephone, a hand-held navigation device or a notebook computer or PDA. With all such devices, the vast majority of inputs are numerical, and number recognition is easier than word recognition. Further advances in the technology are inevitable. In 2006 the US specialist Nuance Communications announced a software package directly aimed at automotive use, with applications ranging through “voice destination entry for navigation systems to voice activated dialling and voice command of automotive controls”.
Looking further and perhaps rather hopefully towards the future, some scientists are looking to interpret not only voice but ‘expression’ inputs. In the USA, the Carnegie Institute of Engineering, its work backed by GM, has evolved a prototype suite of programs, the Driver Companion, which “combines a smart car environment and driver state monitoring, incorporating a wide range of input-output modalities (speech, gesture, face, and fingerprint recognition and speech synthesis), a display hierarchy featuring heads-up display for urgent information and a touch-screen display for less critical information, and automatic recognition of road signs.” Some time ago, BMW also revealed it had been working on systems which would respond to such parameters as the driver’s tone of voice or to specific hand gestures, although nothing has since been heard of this research strand.
To the layman onlooker, such research may seem too clever by half, and there are signs that other design authorities agree. While future developments may perhaps yield a car which responds faster if its driver shouts or waves a fist, a more seemingly promising approach may be wherever possible to use systems which react automatically to information, taking the driver out of auxiliary control loops, and providing him (or her) with the information necessary for completion of the primary task - which, lest we forget, is to travel between two given points at reasonable speed and with maximum safety.
A major debate in this area is the point at which driver “assistance” gives way to fully automated control. It could be argued for example that a blind-spot warning system could go beyond a visual or haptic warning of an adjacent vehicle, and could if necessary inhibit the steering input to prevent a manoeuvre that threatened to result in an accident. Does this act against the principle that the driver is always ultimately responsible? And if it does, how does that sit with the ABS that intervenes in the braking process, the ESP which acts to modify the car’s cornering trajectory, and (an imminent step in the evolution of “driver assistance”) a system which automatically parks the car by reversing into any gap of suitable size? In many cases no question of conflict arises, as in the case of systems - now becoming more widespread - to operate the headlights according to ambient light level, or the wipers according to the presence of rain on the windscreen. Nor do many people question the idea of automatically slowing the vehicle if it passes a warning sign at unreasonably high speed, a feature for which the technology has long existed. Yet there are other cases in which the debate about ultimate driver responsibility and the limitations of automatic override systems is likely to become heated.
Where enabling technology is concerned, there is no doubt that the evolution of cheap but reliable “high technology” sensors, whether CCTV cameras, or radar, infra-red or ultrasonic sensors, has made a huge difference to the kind of assistance that can be offered to the driver. They are enabling the long-mooted “cocoon” of safety space around the car to be created. As already noted, systems already available include intelligent cruise control (automatic control of safe separation), lane-change warning (blind spot cover) and rearward clearance measurement, now used for parking assistance but also, very soon, to form the basis for predictive rear-impact protection. Such information complements that supplied from satellite navigation systems, and from other in-car sensors which enable its dynamic behaviour to be monitored. The information now available from yaw rate sensors and lateral accelerometers can be put to a wide variety of uses.
Where the principal display of information is concerned, it has proved hard to wean drivers away from conventional dials, and these certainly have ergonomic advantages. However, electro-mechanical instruments have largely given way to electronic panel displays, nowadays so artfully crafted that few drivers are aware that the change has taken place. Alongside the instrument panel, the screen for displaying navigation information has itself become almost commonplace, and has grown in size and brightness with the perfection of LCD technology. Vehicle manufacturers have recognised that drivers are going to take their eyes off the road, at moments they judge safe, for long enough to read the display, and accordingly now engineer (and position) the display so that it can be easily read without any need to shift the point of view too far sideways and downwards.
The market appears now to be split between integral displays and those which can be removed and transferred. In either case, the list of features continues to grow, with real-time traffic information expected at least by up-market buyers.
Some manufacturers have moved to use a single screen able to switch displays, for example substituting a schematic of the HVAC or entertainment system for the navigation display. Beyond this there remains a division of opinion about the value of touch-screen technology, which is intuitive but still carries question marks about its long-term durability, and arguably requires the driver to spend more time with his eyes off the road. The alternative is to use some form of manoeuvre-and-select “joystick” such as that found in the BMW i-Drive system, which seems to be cordially hated by press road testers but better liked by long-term BMW owners. By common consent the i-Drive is not intuitive at first use, but its operation can be learned. Whether it represented a step too far in trying to integrate the control of a wide range of systems into a single control mounted on the centre console is a matter for further debate. Other manufacturers, especially in the volume segment, seem to prefer “stand alone” HVAC and entertainment systems with their own controls and their own miniature digital displays.
That other long cherished ideal, the head-up display, has been offered in some luxury-class cars (mainly from GM in the USA) since 2002, but still in relatively crude form. Image projection directly onto the windscreen has proved difficult (aviation HUDs generally use high quality, optically flat secondary screens) but some glass and coatings specialists now claim to have developed better solutions. Late last year, the US specialist Microvision announced that it had concluded an agreement with Visteon to develop “a commercial scanned-beam head-up display (HUD) product for automotive applications.” It remains to be seen whether the latest technology will enable the HUD to take on more of the functions of the conventional instrument panel, perhaps enabling the navigation/systems display screen to be moved even closer to the driver’s normal field of vision.
In summary, the technology of the driver interface is advancing, but it remains subject to the constraints of cost and of consumer conservatism. In some electronic areas, both sensors and displays, costs have fallen dramatically in the past few years, and the challenge now is to evolve systems which lighten, rather than increase, the load on the driver while at the same time enhancing the universal aim of greater safety.
|
| |
Author
Jeff Daniels
|
| |
| |
|
| Email this article |
| |
|
|
|
|