Helping the police with their inquiries Software developed for designing car safety systems is finding a new role in forensic engineering, writes Ben Hargreaves
When the accident happened, one of the occupants of the car was thrown through the passenger window with such force that he died. The other man in the vehicle – the dead man’s brother – survived. He had not been thrown from the car, but remained inside.
When he came round, he insisted to sceptical police officers that he was not the driver. The police, wearily accustomed to such denials, began to prepare to prosecute him for causing death by dangerous driving. His story seemed implausible. Was there any way to prove or disprove it?
The scene of the incident yielded some clues. Tyre marks, debris from the vehicle – the forensic details that police note down as they try to build up a picture of what happened. But something else was needed to construct the case. It was time to enlist the help of engineers.
Advanced Simtech has been working with police forces for the past four years, using simulation software originally designed to help automotive companies reconstruct accidents to improve car safety. In this case, staff at the company ran several hundred simulations of what could have occurred, carefully varying parameters, until they matched the real-world evidence.
Time = 3.500000
Time = 4.400000
Time = 4.900000
The police, on this occasion, were wrong. The simulation demonstrated that the man was telling the truth. The driver had been thrown from the car. The case never went to court, and the dead man’s family were grateful that the surviving son had been absolved of blame. The simulation was even presented to them as proof.
“The files were shown to the mother,” says Mike Brown, managing director of Advanced Simtech. “She was able to see what happened and understand that her son wasn’t at fault. For her, it was part of closure.”
Time = 3.100000
Time = 4.200000
Time = 4.550000
Brown’s company markets a product called Madymo – the acronym stands for Mathematical Dynamic Modelling – in the UK, Scandinavia and the Benelux region. Typically used by automotive designers, more recently Madymo has been employed by police to determine what happened in car accidents.
“It’s become an accepted tool in the legal system because the capabilities of our software are very well suited to this kind of work,” says Brown. “We specialise in the dynamics of what happens to the occupants of the vehicle, the bio-mechanics. We tend to get involved in crashes that involve single vehicles with multiple occupants.”
This area of “forensic engineering” goes far beyond mere visualisation to encompass the physics of a crash, explains Brown. “Our main strength is in bio-mechanics – there are a lot of packages out there that allow you to quickly visualise a crash. What we do is a very detailed analysis which specialises in focusing on the people inside the car. We can recreate the accident so that it correlates with the physical evidence at the scene.”
Madymo’s use by police forces comes as there is a greater drive to treat car accidents in which there are deaths as murder, meaning that determining who was driving becomes of prime importance. “They are dedicating more resources to deaths caused by dangerous driving and, whereas before they might have to put a halt to the investigation, they can now move it forward with the software.”
Data from the crash scene are handed to engineers at Advanced Simtech, who put them into a Madymo mathematical model. All the possible permutations of driver behaviour, passenger position, speed, braking, suspension and road conditions are considered through a number – often hundreds – of experimental simulation runs.
“Most people who carry out a basic simulation will start with deterministic values,” says Brown. “They will say: ‘the road friction is 0.7’, and ‘this person was sitting in this position’. But, in reality, we don’t know any of this. That’s making assumptions, and making assumptions too early means you are pushing yourself toward a certain position. We take all the possible parameters and give each of them a tolerance – say +/-10%. We know that, in terms of seating position, a seat has a certain range, but, because we don’t know whether its occupant was leaning forward or sitting back, we have to put a little bit of tolerance in there and let our predictive software tell us what would happen.
“Then you can say – OK, if you have the seat in a certain position, if you have the brakes applied at a certain percentage, and a particular steering input, then these factors will give you the answer you’re looking for, and match the evidence from the scene of the crash. We spend a lot of time just getting the motion of the vehicle right before we think about the occupants.”
One advantage of the software, compared with, say, a finite element analysis package, is that it can run quickly on a good spec PC, allowing users to do lots of simulations fast. Commercial director Bipin Patel says: “We might run 5,000 experiments for one case, and several hundred in an evening. Other packages might take 15 hours to do one. We can create a simulation in 20-25 minutes.”
As the software has been increasingly used by the police, it is now regarded with caution by lawyers for the defence, says Brown. “You can demonstrate the ‘what if?’ questions,” he says. “People say, well what if the car was going 5mph faster? What if the steering wheel turned here? What if the driver was leaning over to switch on the radio? We have simulations for all those eventualities and we can demonstrate what would have happened. It’s a powerful tool that doesn’t leave room for hypothesis.”
Patel adds: “There’s been cases where the defence admitted guilt before the case starts, and it’s all thanks to the simulation.”
PC Simon Edwards of Gloucestershire Police says he has been able to secure convictions thanks to Madymo. “It’s a fantastic tool. It allows us to achieve results that would otherwise be impossible, and see what version of events is correct,” he says.
-------INFO BOX
Bio-mechanics boosts passenger safety Madymo was developed by Dutch firm TNO Automotive Safety Solutions. According to Advanced Simtech, which had a seven-year history as TNO Automotive UK but underwent a recent rebranding, the software is used by 95% of vehicle manufacturers, as well as tier-one and two suppliers.
“It was developed to help the industry design restraint systems – seatbelts and airbags primarily,” says Bipin Patel. Madymo is used in the design cycle to simulate at low cost the way in which safety systems will interact in a crash.
Advanced Simtech is working on accident reconstruction with seven police forces, and its accurate bio-mechanics is also being used in other fields. A client called BioKinetics, for instance, has been using Madymo to work with soccer’s governing body FIFA to model the impact on the brain of heading a football.
Madymo was used by ITV London to simulate the effect of a commuter train crash on passengers, where it is commonly believed that, the more people there are, the safer it is. Mike Brown says: “We specialise in how people move, and how people interact with any kind of environment, so there are uses in terms of health and safety that have massive potential. We can look at the dynamics of a group of people in a small area in a crash, and see how they behave.”
The software has also been used to mimic the spinal loads experienced by a pilot when ejecting from a fighter jet. And defence manufacturers have been employing Madymo to assess the impact of mine blasts on the occupants of military vehicles. Brown says: “We look at what the likelihood is of personnel hitting rigid objects in the event of a blast, and what the injuries that might result are.”
(Hargreaves, Ben. "Helping the police with their inquires."
Professional Engineer 20.11 (2007): 22 - 23.)
-------This is an interesting development. Of course, the old adage, "Garbage in, garbage out," still applies. E.g. the case of the ill-fated
Mars Climate Orbiter probe, where Lockheed Martin engineers sent navigational data to NASA in English units instead of metric units. NASA engineers, on the other hand, had been expecting metric units. The mix up resulted in the probe dipping 60 miles (96 km) closer to the surface of Mars than planned. The craft was either torn apart by atmospheric forces or bounced off the Martian atmosphere, never to be seen again.
