Q&A with TRW: Review of advanced driver assistance systems
TRW's portfolio of forward-looking camera options can be scaled to the needs of vehicle makers - from lane-keeping and guidance systems to functions such as object- and traffic sign recognition and pedestrian protection. Its cameras can also be oriented for rear-facing for back-over and rear object proximity warnings and park assistance systems. In this interview, Matthew Beecham talked to TRW's executives about the market for certain camera and sensor based ADAS systems.
Driver assistance technologies are evolving rapidly. Just looking back, say 3 - 4 years ago at what was predicted in terms of the application of camera-based technologies, how has that now played out in Europe and North America?
We would characterize the uptake of camera technologies today as behind the predictions of a few years ago, lagging more in North America than Europe initially; the DAS market was just beginning to take off as the recession started, which adversely impacted consumer demand for these technologies for a number of years.
However, these technologies are now taking off in a big way in both markets with the introduction of cameras in more mainstream vehicles. As larger, and in particular global platforms launch with cameras we will start to see the market growth predicted previously. This is being driven by changes in NCAP, again led by Europe but expected to become prominent in the USA as well, and by other influencers such as the USA's Insurance Institute for Highway Safety (IIHS), NHTSA and the National Transportation Safety Board (NTSB) - all who are calling for or taking a very serious look at potential legislation of camera-based technologies (which have been mandated for commercial vehicles in Europe for Lane Departure Warning [LDW] and Automatic Emergency Braking [AEB]).
Vehicle manufacturers have differing positions on what is "too much" in terms of helping the driver steer clear of trouble. Although most of us believe we are good drivers, the evidence shows that driver error is to blame in more than 90% of accidents. So, I guess the most effective active safety systems are those that can provide an autonomous response if the driver cannot avoid an accident. Would you agree?
This is a challenging aspect for Driver Assist applications. While in initial cases the legislation and other influencers mentioned may have tended toward warning systems like LDW, the statistics you mention clearly point toward systems that may need to do more to address the situations where drivers are unable to respond in time to avoid an accident.
In most cases a continuum of assistance may be the best answer, and studies suggest that systems which help drivers through some form of active assistance that coach the driver to more intuitively know what to do will be more effective and more likely to be accepted by the driver.
For example, if a lane departure warning system simply beeps at you once you approach or cross a lane marking without a turn signal indicated, drivers (and passengers) may get annoyed with the system and simply turn it off. However, if the vehicle has an electric steering system, the camera can be linked directly to the steering to provide lane assist that actively brings the vehicle back to toward the center of the lane by applying a steering torque - and the feedback is haptically felt in the steering wheel. In this case the driver simply follows the ADAS system "suggestions" rather than having to interpret the warning and decide how to respond in a potential danger situation. This has proved to be a more popular - active assist style system - with drivers, and of course if you have full AEB it can help prevent or mitigate an accident if the driver does not respond.
In general we find that drivers prefer (discreet) active intervention DAS systems (e.g. ACC, Lane Keep Assist) to warning-only systems (e.g. LDW, FCW) because the active systems reduce the stress and effort of driving where warning-only systems are felt to increase the stress of driving.
As we understand it, multi-function cameras are forming a core technology for advanced DAS. These cameras will be cheaper, more effective and easier to integrate than radar and infra-red systems. Is that correct?
In many cases yes - multi-function cameras are increasingly able to deliver multiple safety and convenience applications, and thus are being looked at as a single sensor option for functions including forward collision warning and automatic emergency braking - particularly at lower speeds - in addition to traditional lane departure warning or lane keeping. While these enhanced object recognition cameras may be more expensive than low-cost radar or Lidar sensors, their ability to offer vehicle braking and steering control in a single package makes them a cost-effective single-box solution for vehicle manufacturers.
It is important to note, however, that cameras have their limitations and that radar in particular still offer the long range sensing performance and resolution necessary for higher speed systems, and also provides a second source of sensing confirmation when making high level safety decisions such as in AEB systems.
In terms of vehicle integration, the task of installing a camera in a vehicle is similar to that of a radar a or laser; mechanical mounting via a bracket is usually required, and the software integration effort is similar.
What can these multi-function cameras offer?
At the simplest level, multifunction cameras can detect overall levels of ambient light and turn the vehicle headlights on or off when it's dark. They can also detect the (lack of) presence of other vehicles nearby and turn the vehicle's main beams on and off appropriately at night.
By increasing the camera's video processing capability, a camera can also detect lane markings and provide lane departure warning or lane keeping by comparing the vehicle's actual lane position to a calculated optimum.
Further increases in image processing capability enable certain objects to be detected and classified, including traffic signs, vehicles (e.g. cars, trucks and motorcycles) and ultimately pedestrians.
Traffic sign recognition is a clear benefit of camera-based technologies in the West. How do you see this market evolving in Europe and North America?
TSR is a desirable camera function that can assist drivers in properly obeying traffic laws such as speed limits, or in being aware that a dangerous curve or intersection is approaching.
There is stronger interest in TSR in Europe than North America currently, primarily because some European countries have variable speed limits that change at various times of the day or night, to improve safety at busy junctions during the day or to reduce noise near residential areas at night for example. This makes it more difficult for a driver to keep track of the actual speed limit in effect at any time.
In addition, EuroNCAP is beginning to reward vehicle manufacturers for the fitment of speed limiting systems to prevent vehicles exceeding the speed limit, and TSR can be used to provide this function automatically.
While we can see such multi-function cameras on the high-end and medium segment cars, do you see this technology permeating down to the low-end at all in Europe?
Multi-function cameras have already penetrated down to mass market C segment vehicles in Europe, and are expected to migrate further downwards in the next 5 years, potentially driven by EuroNCAP and future regulations. Obviously the A and B segments are the most price-sensitive, but at the same time smaller and lighter vehicles that are more vulnerable in a crash can also benefit the most from crash avoidance technologies, and OEMs could potentially use this as a safety differentiator in this segment.
What about rear facing cameras? What do you see happening in North America?
This topic depends on the functional requirements that the camera would be expected to perform. Rear-facing passive video cameras are becoming increasingly popular in the USA, particularly on large vehicles (e.g. SUVs, pick-up trucks and minivans), where they can assist drivers in reversing and also with hooking up a trailer. The market for rear-facing active cameras is currently smaller, but they are used for blind spot detection and rear collision warning for example.
NHTSA is expected to publish an updated regulation in the near future (FMVSS 111), which will require the fitment of technologies to improve driver rearwards visibility when manoeuvring, in order to reduce the number of deaths and injuries that take place each year in reversing accidents. Depending on the final performance requirements of the regulation, this may stimulate the fitment of passive and/ or active rear camera systems.
Could existing reverse parking sensors play a greater role in partial and full parking assistance systems?
Such sensors are a necessary part of the systems and perhaps could be enhanced to play a greater role - but this is not a sensor product line in which TRW is currently involved. We do supply electric power steering systems that play a role in park assist systems.
The remainder of this interview is available on just-auto's QUBE research service