Global NCAP

Safety is an important part of motoring. Drivers need to develop skills but also operate within their capability and view motoring as a serious responsibility. The cars themselves need to be engineered to be as safe as possible. 

The issue with safety in emerging nations is that to put full safety features in them pushes the price up and therefore out of reach for many buyers. Without emphasis placed on safety by the public nor government, a manufacturer could be tempted to cut back on engineering costs regarding safety to make a car more affordable. The new owner is happy to have a cheaper new car. 

A car needs to have good structural strength and sensible passive safety features too. I don't think a car needs say lane departure warning. Simplify the controls of a car and let the driver concentrate on where he is going, not wading through sub-menus on a touch screen. Madness. Getting back on track, are cars made for customers in less affluent countries safe enough or death traps for unwary buyers?

I quote: 'Global NCAP’s serves as a platform for co-operation among new car assessment programmes worldwide and promotes the universal adoption of the United nation’s most important motor vehicle safety standards worldwide'.

Latin NCAP: The new Hyundai HB20 and Ford Ka both received zero stars under what I assume are less stringent standards than western NCAP testing demands. Both cars are/were produced in Brazil and proved to be very popular. 

SafercarsForAfrica: The Great Wall Steed pick-up got a zero score. The Mahindra XUV300 SUV picked up the first 5-star rating under the Africa testing regime.

SaferCarsForIndia: the Mahindra Thar SUV got 4-stars for adult and child occupation. Meanwhile, the Suzuki S-Presso got a zero-star rating, a car that has featured on photographs on this site. I certainly do not endorse it as a suitable purchase. The Hyundai Grand i10 got 2-stars for adult and child protection. The Kia Seltos received 3 for adult protection and 2 for child protection. The TATA Tigor/Tiago was awarded 4 and 3 stars respectively.

The manufacturers may have something to say about the results. They would be welcome to write to me and I would add their comments. In the meantime, I'd say well done to Mahindra and TATA for putting more safety into your cars than Maruti/Suzuki seem to be doing. They may massively outsell you but you haven't dropped standards to meet the competition. 

The Great Wall Steed is no surprise as far as many Chinese carmakers go. Ford has pulled out of Brazilian manufacturing which may be just as well, while I expected a lot more of Hyundai.

I have taken this information from the globalncap website. You can view it by clicking on the name. 

Safety Stars

Picture source: NRMA Australia.

Car crash testing involves awarding starts depending on their safety in preset crash tests. Because all cars are assessed by the same standards, it's a case of comparing apples with apples. One can decide whether a car meets your safety standards compared to the competition before you make a purchase decision. 

What happens if the testing organisation moves the goal posts? Then it would depend on when the vehicle was tested. If they were tested years apart, then the comparison between models becomes irrelevant. A recent example in Australia highlights this.

A few years ago, a Renault van was tested and got three stars. Since then, many safety aids were added to the scoring system. Mitsubishi released the same van onto the market but rather than apply the same stars to it as the equivalent Renault model, the van was retested and the result was a zero star score. 

So here we have the same van sold under two nameplates, one can claim a three-star safety rating and the other zero. It exposed an anomaly where vehicles that have been on the Australian market for several years may have a five-star rating and a new model that may have more safety features could receive a lower rating, due to the much stricter scoring regime applied to it.

How is that supposed to help people make an informed choice? Until recently tests only covered structural safety and now technological safety features have been added. Surely for the time being a structural score and a technological score could both be issued. That would let people know that the structural integrity of the vehicle is sound but it lags in the technology side of safety. 

At some point in the future, the stars could be combined from both sides to create a higher star rating of say ten rather than just five. Giving a vehicle a zero score when it has some structural safety is misleading. The zero to five scoring system is too narrow to allow for this.    

The organisation ANCAP - which was behind this bizarre situation - introduced its tougher scoring system with good intent. As virtually all cars now get a five-star rating, the results weren't saying much. Adding new criteria was a way to highlight the safety features modern cars are expected to have. 

But to have a system that makes older vehicles look safer than new ones (when in some cases the opposite is the case) is working against helping people chose safer vehicles. For this, I give ANCAP zero stars.

Self Driving Cars Reality

Add to this animals, children, cyclists, debris, road
works etc and it's all too much for technology

Regular readers of this blog will know that my take on the hype over autonomous cars has been less than enthusiastic. One article I wrote on the subject can be found here which contains a Part 2 link. Alternatively, click on the 'Safety' label to the right. True self-driving cars in real world conditions always seemed science fiction to me.

Not long back a top Toyota man tasked with this said it now doubted level 5 autonomous was achievable anytime soon, maybe never. Steve Wosniak (co-founder of Apple) was a strong believer and invested in this but now says he's basically given up on level 5. He said roads offer too many unpredictable scenarios for a self-driving car to handle. He went on to say that the public has been misled into believing autonomous driving could match what a human can do.

So why did companies spend billions on something that may never achieve what was promised? Didn't they have smart people advising them? Aren't clever minds beavering away on this to bring it to a successful conclusion? Because the impossible is just that. As for the advice they got, well they should have hired me instead (wry smile).

Why am I so sure about this, as I have no training in this field? I simply took a step back and looked at the whole scenario. First, I decided that while some humans abuse the privilege of driving, when acting responsibly people have remarkable capabilities to anticipate, evaluate and react to unexpected and varying conditions. Can artificial intelligence even get close to that? Secondly, I believe that the human mind was made by a super intelligent designer and that it has amazing qualities man can never recreate.

Parking NZ Style

When it comes to inconveniencing other people, the Kiwi male is very good. Parking across your access to and from your house, in drive ways that others use to get to their home and garage and footpaths being blocked. All of this to save them walking a few extra metres. If you ask them to move, the reply is 'I'll only be ten minutes". If you say I can't wait that long, then they reluctantly comply, distraught over the extra minuscule walking distance they now have to undertake. We're talking a few metres here by the way. NZ isn't a crowded country with street parking free and parking spaces rarely an issue.

Anyway, I was walking down this road recently and found there two vehicles parked facing the wrong way (illegal in NZ) and also denying pedestrians their footpath. It was closer to the front door no doubt but the minimal distance saved makes you wonder why break the law to gain so little. Not to mention inconveniencing elderly or invalids.

Human Nature & Driving

When a human is driving a car, they have their hands on the steering wheel, feeling the car react as the wheel is used. The pedals on the floor also cause fluctuations in velocity. The human is then connected to the job at hand and is fully involved. Of course distractions, over confidence and other factors can intervene, which can lead to accidents. However, a driver feeling the car respond to inputs they make keeps the person engaged.

We are seeing cars being released to the public with Autopilot. They work well on aeroplanes although used for the boring flying parts. I've seen videos showing a pilot asleep while the plane is in autopilot. So how do car drivers react to autopilot? I'm a human so I know.

Once the hands and feet disengage from the controls, the driver immediately starts to disengage from driving. It is impossible for a human to sit in a car, passively allowing the car to be driven for them and remain fully alert. They will not be able to react to a situation anywhere near as quickly and efficiently than they would if they were driving. Even the most alert passive driver will be affected, and the longer they are passive, the harder it will get.

Then there are those who won't remain passively alert at all. They will find other things to do. Read, text, snooze, turn to talk to other passengers, try to find something in their glove box, look at what's going on around them.

Amazingly, some people don't get this, or don't want to get it. They are the people behind making the car self driving. I've discussed some of the problems associated with that at other times (see under label "Safety') but this is focusing on Autopilot. Here is a recent case:

Reports state that a Tesla Model S that went through a red light and crashed into a stationary fire truck at high speed was in Autopilot mode. The driver told police she was looking at her phone prior to the collision and estimated her speed at 60 mph.

The company that makes the car warns drivers using this mode to remain alert and active and to be prepared to take action at any time. You couldn't make this up. No understanding of human nature with that instruction. How can you remain active while the car drives for you? Personally I wouldn't use Autopilot if my car had that feature. It doesn't work with a human driver. There are also too many variables with motoring. I like to keep control because I want to remain actively alert. I cannot do that while the car drives for me.

The 'Superior' Autonomous Vehicle

A woman pushing a bike across a road is tragically killed by an Uber autonomous vehicle. It didn't seem to see here, despite the fact that she had nearly got across the path of the vehicle when the collision occured. The driver was momentarily distracted, which is an irrelevant point as I shall explain.

I was on vacation in Australia driving a large MPV. Approaching me was a person pushing a bike on the shoulder of the road, which was quite wide enough to do so. Just as I was nearly up to the person, they suddenly turned their bike onto the road in front of me. In a flash I did a frantic S movement away from the pedestrian/bike and missed them.

I looked back in the rear view mirror to see that the young man had pulled back off the road and was standing still. He looked disoriented as he took in the near miss scenario. The point is would an autonomous car have reacted a quickly as I did? In the case of the Uber vehicle mentioned at the outset, it had more reaction time. If the car I was driving was in autonomous mode and it didn't react, could I have intervened and avoided the impact? No! A human is at their sharpest while fully in control of their vehicle.

So what of Uber's autonomous vehicles? Here is a quote: Uber's self-driving vehicles are equipped not only with cameras, but with radar and lidar, which works like radar but uses lasers to detect objects on and off the roadway. 

Uber and other companies working to develop self-driving cars tout the safety of their systems not only because the vehicles won't lose focus on the road, like human drivers, but because they have superior sensing capabilities. 

Last fall, Uber officials showing off their vehicles in Tempe said their radar and lidar were able to detect objects, including jaywalkers, as far as 100 yards away and avoid collisions.

Source of quote: USA Today.

I question that these vehicles have superior sensing capabilities after seeing the photographic evidence. I cannot understand their jaywalker claim either, after this incident. One cannot expect new technology to be perfected without real world testing. The sad reality is people have, and will continue to die while this testing continues.

I would prefer it if they weren't overselling the current capabilities of the technology. I guess if they were candid about the present state of things, then they may not get real world testing permits. However, putting these vehicles out while not ready is dangerous. It seems to me the whole process is being rushed and unnecessary deaths are occurring as a result.

My condolences to the Herzberg family and friends of the family.

Autonomous Cars : Part 2

Those behind the autonomous car industry are a defensive bunch it seems to me. They have been putting cars on the road in real world situations to learn how to deal with it. Some of the results are less than impressive.

A while back a man driving a Tesla car was killed when it slammed into a truck while using its autopilot function. It was claimed that the driver was watching the movie at the time (this was strongly denied by his lawyer) and that he didn't have his hands on the steering wheel. Therefore Tesla seemed to take the stance it was driver error.

However, a safety agency report said Tesla’s Autopilot design also contributed to the accident. The 'autopilot' function was not designed to stop if a truck was crossing in front of it. The agency also recommended that vehicle manufacturers and federal regulators take steps to ensure that more advanced automation isn’t used in situations it’s not designed for.

Calling it autopilot is just plain stupid and dangerous if the driver cannot take his hands off the wheel at any time. That isn't autopilot. The system's inability to notice the dangerous situation unfolding surely means it wasn't ready to be put to use in a car. OK, the driver was trusting the system too much but to have to have your hands constantly on the steering wheel makes the autonomous system pointless.

In another more recent case, a small self driving bus had been put into operation and almost immediately was involved in an accident. A tech writer on board stated the truck driver couldn't see the small bus as he reversed into the space he required (see pic above). The accident was quickly described as human error.

However, what surprised me was how this machine was allowed to operate outside its play pen. There was plenty of space behind the pod to reverse into to avoid the slow motion scrape, but no it just stopped. Like a rabbit caught in headlights. It could have sounded its horn to warn the truck driver of its presence, but such a basic feature wasn't in its defensive arsenal.

My take: The move from basic to advanced autonomy is a long road, perhaps an impossible journey for human technology to ever pull off while unpredictable scenarios are part of the equation. True, failures are part of any learning curve. However, the speed to deflect any criticism away from the technology is disturbing. It may be to avoid negative publicity but to do so also come across as arrogance. Its a belief that technology is our future, a god we must trust in to take us to a better world. Yeah, right. All I can say is dreams are free.

Part one can be found by clicking here.

Takata

One song says love changes everything, and so does money. Corruption is rife in the world today and money is so often the motivation. Even human life can be a secondary consideration in the quest to gain or protect money.



Cars are convenient but also dangerous. Safety features are important to consumers and modern cars are full of  such life saving features. I find some such as lane departure warning as quite laughable but most are welcome additions. I must say as one who has always used seat belts I thought airbags were for Americans too lazy to buckle up (no offense intended). However, they do add to occupant safety so I've been won over.

Which leads to the point of the article. Japanese firm Takata changed the main ingredient for its airbags from tetrazole to ammonium nitrate in 2001. Ammonium nitrate provides more bang in a smaller volume than other propellants, which allows the company to offer a more compact device to manufacturers. It was also cheaper to make. However, they have in a few cases caused death by exploding and sending pieces of the device into the car's occupants.

Employee emails show that there were “dire warnings about safety and quality lapses years before Takata Corp would fully acknowledge the threat posed by its defective airbags,” according to a report by the Senate Commerce Committee. Senator Bill Nelson, a Floridan Democrat, says it now appears that Takata was aware of “serious safety and quality control lapses” in its factories as early as 2001.

Company officials debated the data regarding the product and testing of the inflator was manipulated to cover the issues. Nissan, BMW of North America and Ford have accused their airbag supplier Takata of fraud in a lawsuit to cover losses they incur because of its faulty inflators that have been linked to at least 17 deaths worldwide. Court documents filed in United States claim five automakers were aware of potential defects but continued using Takata’s airbags because they were less expensive. The allegations involve Ford, Honda, Nissan, Toyota and BMW.

All in all a sad tale of a business trying to get a competitive edge but getting it wrong and trying to cover it over. Did they really think they would get away with it? The cost of owning up would have been huge. It got worse by not coming clean. Money changes everything.

Autonomous Cars : Part 1

Is this sort of image creating an unrealistic expectation?

There is much hype around cars that can drive themselves. Those who are funding the development of them say they will create a safer environment on the road. One publication was so upset that a car on autopilot recently led to the death of the driver, it lashed out at "the extreme fallibility of human drivers" as if to minimise the artificial intelligence's failure. So what can we make of it all?

1) Does human driving reflect extreme fallibility? The fact that humans of all sorts take to the roads, then fatalities per km are remarkably low. If it is such a concern, drivers should be retested at sometime in their lives beyond the first getting of a licence. I have only had one driving test all my life.

2) Are autonomous cars going to reduce fatalities? By taking the driving off the driver while expecting him/her to keep as alert as a driving human is unreasonable. Added to that a person who spends little time driving will lose skill and sharpness. In addition, can artificial intelligence ever have the ability to read and anticipate every scenario?

3) So what is autonomous driving's raison d'etre? I see it best in motorway situations, taking over in heavy traffic and relieving stress. It would also reduce breaking ripples. There is not that much for the computer to do and the driver will be kept fresher.

In summary, I wonder if some are overstating how far technology has gone and can go in the foreseeable future. Funding may be reduced if the progress per dollar spent was more realistically costed. The recent death mentioned at the outset could be because a human was relying too much on the auto pilot. Perhaps the rhetoric coming out of the media about autonomous control of cars has mislead people to expect too much.

The picture accompanying this article shows how far some may think autonomous driving has gone or can go. In reality, the scenario pictured could take many years to happen or may never be achieved. Trying to get technology (created by humans anyway) to mimic human abilities is impossible. Computers are good at number crunching and taking on tedious jobs but a high level of anticipation, understanding and reaction is best left to the human driver.

Part 2 can be found by clicking here.

The Driverless Car - A Layman's View

The driverless - or automated - car is considered a future reality. I have asked people about it and everyone has said "No way. I want to be in control". It may in fact show that people enjoy driving anyway, and they prefer that independence. They also may lack confidence in technology to deliver them safely to their destination.

Some benefits: 1) Fewer traffic accidents in basic driving conditions such as motorways. 2) A relaxed journey for the driver. 3) Age, disabilities and intoxication would not be issues in using cars. 4) Reduced traffic police and less insurance. 5) An automonous car could come to pick you up or drive to a garage for servicing without any passengers.

Maybe you prefer a driverless pod

Some drawbacks:
1) People wanting to be in control as mentioned above. 2) A driver losing driver sharpness when constantly being driven around then having to suddenly take the wheel when necessary.
3) Litigation against the car maker if the system fails to stop an accident. 4) Cyber security and communication issues.
5) When the system breaks down, then being able to take over and drive' for example a disabled person. 6) Having the ability to negotiate more difficult road conditions. In other words, computers cannot think. How will they handle sudden, unpredictable situations? 7) With a road obstruction, knowing whether to wait or reverse and take another route.

In summary: On motorways and town roads, I could see the automated car coping. I don't on open roads such as we have in New Zealand that are challenging. The ability for the human brain to suddenly make a decisions for something that could not be predicted is something a computer cannot do. It has to be preprogrammed or it won't know what to do.

Frankly, I see this as a possibility in certain situations but for all scenarios a long time away, perhaps never. There is no doubt more to this than I know, but even a layman's overview shows up the complexity of the subject, both technically and emotionally.

But what about twisty back roads and sudden challenges?

Toyota Trial Mitigate Impact System (M.I.S.)

Toyota is trialing the Mitigate Impact System, or M.I.S. for short. Of course it is better if the other car does MISS you, but if there is a crash this system mitigates - or lessens - the impact. Some say visually it isn't attractive, but as the manufacturer says "Do you want to protect your precious paintwork or not?"

If the trial is successful - and I cannot see how it wouldn't work - then they will gradually start appearing on new models in 2014. I applaud the initiative, but question whether the tyres should be colour matched to the car. Apparently au naturel means impacts won't require any touch up to the tyre. Smart thinking.

Cynical Car Makers?

Many car makers in Brazil apparently don't make vehicles that are very crash worthy. Legally they do what is necessary to pass the laws and that is all that matters to some of them. I ask the question "is it cynical" as I wonder who is really responsible.

Are the authorities to blame for not having higher standards? Can the manufacturers be blamed when if made better - hence more expensive - they could lose customers to cheaper, less safe models? Should the consumer find out which ones are safer and then support them? In other words are customers indifferent?

There is an article that analyses the situation that can be found here. It is quite long, but seems well researched. As for whether car makers could do better, I wrote an article (click here) about cars tested in more affluent markets. It doesn't cover many brands and those omitted cannot be judged; but for those mentioned, it's concerning.

Car safety is important. Manufacturers who go beyond what is required by law to protect their customers are to be applauded. For those do the minimum to maximise profit, I'll let you decide.

Also, the driver must  take responsibility to be careful for the safety of all road users. Drive to the conditions.

Check your vehicle keep it well maintained. Make sure each journey is as safe as you can make it.

The Limitations Of Car Crash Tests

New cars are tested for crash worthiness as we know. They are usually rated from zero to five stars and people are greatly influenced by the results. Potential buyers may shy away from anything less than a five star rating.

My personal view is that too much is read into the system. Of course, a newly released car getting two or three stars today would be unacceptable but the difference between a four and five star rating may not be as people think. The reason is that it is very difficult to replicate all the nuances of 'real world' accidents when testing cars in controlled environment.

Now a new overlap test that puts only 25% of the front of a car into an object (at 40 kph) as opposed to say 40% has been carried out. It was done by the The Insurance Institute for Highway Safety, and the results were sobering. It seems that many manufacturers that trumpet safety, are simply very good at making their cars pass existing tests. This new test showed how few cars are made to stand up to a smaller object impact, such as would happen with a tree of power pole.

Poor: Mercedes C Class, Lexus IS and ES350 and Audi A4.

Marginal: Acura TSX, BMW 3 Series, Lincoln MKZ and VW CC.

Good: Acura TL and Volvo S60.

The above models are premium, the next batch of tests will be every day models, and it won't be pretty I suggest. Yet even some reputable models as listed above had unacceptably poor results. The IS and C-Class had sufficient structural deformation to trap the dummies feet while the door of the VW CC came off!

It seems that five stars can be achieved if you study standard tests and engineer cars to meet those. A common accident like hitting a power pole is for some reason not in existing tests. So car makers generally don't worry about making them safe for that scenario. That is cynical to me. One of the brands tested builds cars not just to get five stars but to make safe cars, and that is Volvo.

So if someone says they won't buy a four star car, well maybe it failed a technicality on the existing test, whereas a five star car might in fact be worse in certain accidents. If you are very concerned about safety, buy a Volvo.