TRL's Tony Read discusses the use of onboard vehicle systems as an enforcement enabler
From an enforcement perspective at least,1686 Toyota's recent recalls over problems with accelerator pedal assemblies had a positive outcome in that for the first time a major motor manufacturer outside of the US acknowledged publicly what many have known or suspected for quite a while: that the capability exists within certain car companies to extract data from a vehicle onboard unit which can be used to help ascertain, if not prove outright, just what was happening in the vital seconds up to an accident or crash.
Previously in this magazine,650 TISPOL's Peter van de Beek has noted that too often it seems that those who are on the front line of traffic enforcement are also at the back of the queue when it comes to formulation of enforcement policy and, by extension, the development of the necessary technology. In particular, he expressed frustration at the continuing inability of police and enforcement officers to access onboard information which would make proving innocence or guilt a far simpler task; this has become far more difficult in recent years as ABS and ESP systems have removed much of the physical evidence which accident investigators once relied upon as vehicles are less likely to skid (see 'Intelligence-led?', pp.43-44, ITS International March-April 2010).
It is worth making some distinctions in terms of the technology being discussed here.
EDRs are a part of the vehicle itself and fitment requires little more than the addition of memory capacity to the airbag control module. This allows, typically, data to be retrieved on the five seconds up to airbag deployment as well as the 500 milliseconds after. Deployed systems vary but in the case of more modern vehicles, which have both front and side airbags, it is possible to derive both longitudinal and latitudinal movement information. There is in fact a much broader range of information that can be accessed but a key distinction is that with an EDR much of the data collected is derived from systems on the vehicle whose primary functions are elsewhere. An Accident Data Recorder (ADR) is an altogether more complex beast. It is the fabled 'black box' - a dedicated piece of equipment concerned only with providing a comprehensive record of 'What Went Wrong' and is a discrete system which is added to the vehicle.
491 TRL.
"Purpose-built ADRs are broadly similar. They incorporate a series of accelerometers and the ability to collect data from a vehicle's onboard systems.
"To give some idea of capability, ADRs such as the189 Siemens UDS and the RGS Voyager provide information on what was happening for around 30s before a trigger threshold, as well as the 15s and 100m of travel - which is actually time-independent - after that threshold has been reached. Such information includes longitudinal and latitudinal acceleration, vehicle speed, compass bearing/direction and distance travelled. Up to 10 'status signals' are also retained, such as whether the ignition was on, brakes light or indicators were operating, or whether side, dipped or main headlights were in use. Three of those status signals can be customer-defined, so might include whether blue lights and sirens were working on a police vehicle, for example. Some ambulances also include an alarm on the drugs cupboard.
"In all, such a system allows two to three events to be linked together to provide perhaps two to three minutes of overall data. Information is collected at 16Hz, though this increases to 500Hz within 120ms in the case of an incident. That's quite low compared to research systems but it's very good for real-life crash data.
"To put that in context, consider what happens with more traditional accident investigation techniques. ABS and ESP systems mean that today there are almost certainly no skidmarks at an accident scene. Top-end ABS systems are becoming so sophisticated that deceleration is being held so close to what was once the lock-up point so that we are starting to see an approximation of locked-wheel skidmarks on some surfaces, such as stone mastic asphalt, but this is still very rare.
"Skidmarks were the bread and butter of accident investigation. They allowed expert assessment of events over perhaps one to two seconds before an impact whereas an ADR provides up to 30 seconds of very high quality data."
Judgements can be made against preset parameters and drivers can be scored against others in a fleet. This has several advantages for a fleet operator. It facilitates monitoring and ranking of individuals' performance and decisions as to whether remedial action/training is necessary - or even whether someone needs to be removed from driving duties altogether. As with all organisations where driving is a part of the employees' work, it provides a measure of protection against corporate manslaughter charges, as it can prove or disprove culpability in the case of a collision.
GPS modules can also be added to facilitate resource allocation. From an enforcement point of view, this has both disadvantages and advantages, however. Read: "The main disadvantage of an ADR is that unless you know where a vehicle is, you can't geographically 'fix' the data it gives you. Not having GPS added discounts any 'Big Brother' concerns but I'd suggest that these are already addressed by the fact that data is only held for a very short timeframe, is only likely to be looked at once an incident has occurred, and then only by approved individuals; tachographs provide an existing example of something similar which works well. It's also the case that the necessary download equipment isn't cheap and you'd be unlikely to see all and sundry attempting to do this."
"The1690 European Commission's VERONICA project identified that, yes, ADRs would be a good idea - a decision that was hardly a surprise," Read continues. "However, the VERONICA 2 project looked at the practical aspects of implementation - what data to collect, and how. The main disadvantage of the North American EDR system was deemed to be that it didn't deal with accidents involving vulnerable road users - a collision with a pedestrian or two-wheeler was unlikely to cause airbag initiation and therefore provide incident data. This was seen as a major failing.
"An ADR can have its parameters set - a police vehicle, for instance, which could well be driven a lot more harshly than a standard car, might have higher activation thresholds. However, while the assumption is that an EDR is set at the airbag activation/seatbelt pre-tensioning threshold, the reality is that many have what is effectively 'near deployment' functionality. So, if a vehicle has a low-energy impact the airbag system might be 'woken up' but then decide subsequently that it doesn't need to deploy.
"At present deployment in vehicles manufactured for the European market is effectively stalled. There is no ongoing project looking at this and in the current political and financial climate I don't see that changing any time soon. That's a pity because something like the North American EDR solution is not especially difficult or expensive to implement, whereas a purpose-built ADR is a different prospect even if economies of scale are taken into account."
And yet, notes Read, there have been instances where police forces have been able to gain data from airbag control modules and 'freeze-frame' data from vehicles' diagnostic codes.
"There seems to be reluctance on the part of some OEMs to make any retained data available. One issue is that they themselves may not know the capabilities of the onboard modules in question and it might be that such queries are better aimed at the modules' manufacturers instead; they supply modules to meet a given requirement and it may be that the module has greater capabilities than the OEM requested or are aware of.
"There are certainly cost implications and I think perceptions of product liability which mean that at the moment some vehicle manufacturers don't see an advantage to providing such a service or facility. That's a real shame because in many cases EDRs and ADRs would provide a robust defence against liability claims. There are, for example, lots of potential commercial opportunities, such as tie-ups with insurance companies where drivers benefit from lower premiums if they have such a system fitted.
"In some respects, one could argue that the manufacturers of purpose-built systems must take some responsibility for the current situation. They have tended to concentrate on the police and emergency services and almost ignored the fleet market and the general public. But when you get down to it, it's difficult to see any disadvantages - the evidence is independent of any slant which a witness might put on it; in civil claims, there is more likely to be an equitable settlement; and in criminal cases, true blame can be apportioned. From the vehicle manufacturers' perspective, there's the opportunity to more readily rebuff some of the more spurious claims made against them and remove themselves from the picture at a very early stage.
"In terms of who might take this forward, I think there are so many stakeholders that identifying a lead or a champion is difficult. Perhaps it's fairer to say that all of those concerned need to reconsider the current situation and look at how we take adoption forward."
From an enforcement perspective at least,
Previously in this magazine,
EDRs and ADRs
In the US, Event Data Recorder (EDRs) are now common on new vehicles. Along with these have come efforts to provide legal clarification of what sort of information is collected and recoverable, as well as who should be allowed to recover what information and under what circumstances. (NHTSA 49 CFR Part 563 stipulates if a vehicle retains one of a number of parameters then it must retain all the parameters listed. If a vehicle doesn't retain any of the listed parameters it doesn't have to retain any at all and therefore doesn't have to comply. However, NHTSA reckons that about 97 per cent of vehicles will have to comply.) Recent legal cases in the Netherlands and the UK, where 'black box' data has been used to secure driving convictions for the first time, are a step forward in this respect but show how painfully slow progress is in some parts of the world.It is worth making some distinctions in terms of the technology being discussed here.
EDRs are a part of the vehicle itself and fitment requires little more than the addition of memory capacity to the airbag control module. This allows, typically, data to be retrieved on the five seconds up to airbag deployment as well as the 500 milliseconds after. Deployed systems vary but in the case of more modern vehicles, which have both front and side airbags, it is possible to derive both longitudinal and latitudinal movement information. There is in fact a much broader range of information that can be accessed but a key distinction is that with an EDR much of the data collected is derived from systems on the vehicle whose primary functions are elsewhere. An Accident Data Recorder (ADR) is an altogether more complex beast. It is the fabled 'black box' - a dedicated piece of equipment concerned only with providing a comprehensive record of 'What Went Wrong' and is a discrete system which is added to the vehicle.
A defining solution
The ADR is in many respects the Holy Grail of accident investigation and research, according to Tony Read, Principal Consultant - Incident Investigation and Reconstruction at"Purpose-built ADRs are broadly similar. They incorporate a series of accelerometers and the ability to collect data from a vehicle's onboard systems.
"To give some idea of capability, ADRs such as the
"In all, such a system allows two to three events to be linked together to provide perhaps two to three minutes of overall data. Information is collected at 16Hz, though this increases to 500Hz within 120ms in the case of an incident. That's quite low compared to research systems but it's very good for real-life crash data.
"To put that in context, consider what happens with more traditional accident investigation techniques. ABS and ESP systems mean that today there are almost certainly no skidmarks at an accident scene. Top-end ABS systems are becoming so sophisticated that deceleration is being held so close to what was once the lock-up point so that we are starting to see an approximation of locked-wheel skidmarks on some surfaces, such as stone mastic asphalt, but this is still very rare.
"Skidmarks were the bread and butter of accident investigation. They allowed expert assessment of events over perhaps one to two seconds before an impact whereas an ADR provides up to 30 seconds of very high quality data."
ADR evolution
The basic ADR technology is well established, having been around for over a decade. The major developments tend to be in remote/real-time downloading and driver profiling. The latter allows peer comparison and can be used to distinguish between, for example, members of the public and police/emergency services drivers, and transponder technology exists which can distinguish exactly which of several occupants in a vehicle is driving.Judgements can be made against preset parameters and drivers can be scored against others in a fleet. This has several advantages for a fleet operator. It facilitates monitoring and ranking of individuals' performance and decisions as to whether remedial action/training is necessary - or even whether someone needs to be removed from driving duties altogether. As with all organisations where driving is a part of the employees' work, it provides a measure of protection against corporate manslaughter charges, as it can prove or disprove culpability in the case of a collision.
GPS modules can also be added to facilitate resource allocation. From an enforcement point of view, this has both disadvantages and advantages, however. Read: "The main disadvantage of an ADR is that unless you know where a vehicle is, you can't geographically 'fix' the data it gives you. Not having GPS added discounts any 'Big Brother' concerns but I'd suggest that these are already addressed by the fact that data is only held for a very short timeframe, is only likely to be looked at once an incident has occurred, and then only by approved individuals; tachographs provide an existing example of something similar which works well. It's also the case that the necessary download equipment isn't cheap and you'd be unlikely to see all and sundry attempting to do this."
Adoption obstacles
Privacy, unsurprisingly, continues to be an issue. Toyota's travails spotlighted the developments in the US but away from the North American market the automotive OEMs do not appear to be placing too much emphasis on providing vehicles with any form of EDR/ADR capability."The
"An ADR can have its parameters set - a police vehicle, for instance, which could well be driven a lot more harshly than a standard car, might have higher activation thresholds. However, while the assumption is that an EDR is set at the airbag activation/seatbelt pre-tensioning threshold, the reality is that many have what is effectively 'near deployment' functionality. So, if a vehicle has a low-energy impact the airbag system might be 'woken up' but then decide subsequently that it doesn't need to deploy.
"At present deployment in vehicles manufactured for the European market is effectively stalled. There is no ongoing project looking at this and in the current political and financial climate I don't see that changing any time soon. That's a pity because something like the North American EDR solution is not especially difficult or expensive to implement, whereas a purpose-built ADR is a different prospect even if economies of scale are taken into account."
System potential
Historically, most vehicle manufacturers have been very reluctant to get involved in accident investigations and the provision of evidence. The intimation is that the systems from which data can be derived were not designed for that purpose and that has led the judiciary in some cases to dismiss or discount such evidence.And yet, notes Read, there have been instances where police forces have been able to gain data from airbag control modules and 'freeze-frame' data from vehicles' diagnostic codes.
"There seems to be reluctance on the part of some OEMs to make any retained data available. One issue is that they themselves may not know the capabilities of the onboard modules in question and it might be that such queries are better aimed at the modules' manufacturers instead; they supply modules to meet a given requirement and it may be that the module has greater capabilities than the OEM requested or are aware of.
"There are certainly cost implications and I think perceptions of product liability which mean that at the moment some vehicle manufacturers don't see an advantage to providing such a service or facility. That's a real shame because in many cases EDRs and ADRs would provide a robust defence against liability claims. There are, for example, lots of potential commercial opportunities, such as tie-ups with insurance companies where drivers benefit from lower premiums if they have such a system fitted.
"In some respects, one could argue that the manufacturers of purpose-built systems must take some responsibility for the current situation. They have tended to concentrate on the police and emergency services and almost ignored the fleet market and the general public. But when you get down to it, it's difficult to see any disadvantages - the evidence is independent of any slant which a witness might put on it; in civil claims, there is more likely to be an equitable settlement; and in criminal cases, true blame can be apportioned. From the vehicle manufacturers' perspective, there's the opportunity to more readily rebuff some of the more spurious claims made against them and remove themselves from the picture at a very early stage.
"In terms of who might take this forward, I think there are so many stakeholders that identifying a lead or a champion is difficult. Perhaps it's fairer to say that all of those concerned need to reconsider the current situation and look at how we take adoption forward."
