What result would the Ineos Grenadier achieve in the Euro NCAP or Global NCAP tests?

[MrMike]

The thing with safety.. sometimes it's the big things, but sometimes it's all the little details. It's very hard to guess deceleration rates from a video.

I disassemble vehicles that have been in accidents in a voluntary basis (Fire and Rescue) and from my own observations I think you would be reasonably safe in a Grenadier, the other object you'd come in contact with I'm not so sure as a 3T ladder framed object it would do some serious damage. I'd like a knee air bag but at a $25k option (MY24 model) I think I'll pass.

 

[255/85]

Just like at sea: gross tonnage wins out.

 

[parb]

Euro-NCAP and the US ncap is going to harmonize. Nhtsa has solicited comments about that. I guess the 5 star rating has lived it's useful life.

Now euro-NCAP is a very different beast, especially the current iteration.
For an ineos to score well on that you need a ton more sensors for the ncap scenarios. Things like dooring where a bicycle come from behind a parked car and the car is supposed to prevent the operator from opening the door so the bicyclist doesnt hit it will require both rear, corner and likely side sensing.

Or the scenario where a child is coming out from behind a parked car in front of the vehicle under test and the car has to stop before hitting the child. This scenario require the vehicle under test to see the child while obscured by the parked car or there is no space to stop. I believe most cars as currently configured will fail this test. To solve for this you need likely a really sophisticated corner radar and a wide angle camera mounted high. Yet the highest speeds in this scenario is unsolvable based on physics.

I've read the non public test data for several german and British vehicles and these tests are no walk in the park. Many requires several runs to get to an acceptable score. Surely with lots of tuning.

I think Ineos is wise to wait a bit with this.

 

[DaveB]

Euro-NCAP and the US ncap is going to harmonize. Nhtsa has solicited comments about that. I guess the 5 star rating has lived it's useful life.

Now euro-NCAP is a very different beast, especially the current iteration.
For an ineos to score well on that you need a ton more sensors for the ncap scenarios. Things like dooring where a bicycle come from behind a parked car and the car is supposed to prevent the operator from opening the door so the bicyclist doesnt hit it will require both rear, corner and likely side sensing.

Or the scenario where a child is coming out from behind a parked car in front of the vehicle under test and the car has to stop before hitting the child. This scenario require the vehicle under test to see the child while obscured by the parked car or there is no space to stop. I believe most cars as currently configured will fail this test. To solve for this you need likely a really sophisticated corner radar and a wide angle camera mounted high. Yet the highest speeds in this scenario is unsolvable based on physics.

I've read the non public test data for several german and British vehicles and these tests are no walk in the park. Many requires several runs to get to an acceptable score. Surely with lots of tuning.

I think Ineos is wise to wait a bit with this.

Wouldn't it be much safer, easier and cheaper if the cyclist used one of the senses built into his head to detect if the car was stopping and the door was opening.
What if the child was supervised by an adult so they didn't step out in front of a moving vehicle, or better yet use a pedestrian crossing.

 

[James]

Euro-NCAP and the US ncap is going to harmonize. Nhtsa has solicited comments about that. I guess the 5 star rating has lived it's useful life.

Now euro-NCAP is a very different beast, especially the current iteration.
For an ineos to score well on that you need a ton more sensors for the ncap scenarios. Things like dooring where a bicycle come from behind a parked car and the car is supposed to prevent the operator from opening the door so the bicyclist doesnt hit it will require both rear, corner and likely side sensing.

Or the scenario where a child is coming out from behind a parked car in front of the vehicle under test and the car has to stop before hitting the child. This scenario require the vehicle under test to see the child while obscured by the parked car or there is no space to stop. I believe most cars as currently configured will fail this test. To solve for this you need likely a really sophisticated corner radar and a wide angle camera mounted high. Yet the highest speeds in this scenario is unsolvable based on physics.

I've read the non public test data for several german and British vehicles and these tests are no walk in the park. Many requires several runs to get to an acceptable score. Surely with lots of tuning.

I think Ineos is wise to wait a bit with this.

Really interesting parb, thanks. Are you working in the space (to get non public data)?
While we;re all agreeing the grenadier is not the platform for much of this, it would be interesting if you can shed light on how the better performing cars are achieving their results. Is lidar essential? Does tesla, with more cameras onboard, do relatively well? Are you seeing anything counterintuitive in general?

 

[parb]

I am working in the automotive space.

It's a combination of sensors but I think more important is the overall system latency (eg how much time it takes from sensor processing and until you have computed a solution to the vehicle control system). There are times of systems in current shipping vehicles that rely on very early detection because system latencies are very high.

You need a couple of things, let's assume we just care about kinematics and ignore complicated actor models which try to predict behavior based on classifying what an object is.
You need to know angular vector to the other object. Cameras generally are good for that, mostly because they have great spatial resolution.

You need to know it's motion vector. This is where radar shines, it tells you for each sample which direction an object is traveling at and which velocity.

You need a distance to the object. Radar works okay at far distances but aren't that great at close distances. Radar isn't the easiest to fuse with camera data.
Lidar is pretty good at distances, especially at closer distances. Stereo scopic cameras work but the math is computationally challenging at higher pixel resolutions.

Note I skipped how to find the vulnerable road user. That is another answer just in itself and not the easiest to answer. In general its some form of classifier for the various sensor modalities.

The challenge with some of the ncap scenarios in current standard is that some scenarios are really hard because so much of what needs to be detected are partially obscured. Finding something that is the size of an adult torso is straight forward. Finding the upper 1/3 of a child when behind a parked car next to the curb as you drive towards it is hard. Like really hard.

Oems can't have too many false positives where the car brakes for nothing or people will turn off those functions. Ncap requires testing to be done at the middle sensitivity setting. There is a ton of tuning that goes into getting a ncap score that is good but doesn't result in shipping a car that consumers hate. Therein lies the problem, too sensitive and you get to many false alarms, not sensitive enough and you won't pass ncap.

It's not as simple as just a different or more sensors, it's the overall system design with classifiers, sensor fusion and avoiding false positives while not missing real positives that makes this challenging.

In the tests I often see 10-20 test runs with consultants who typically help out with the test (I think my dataset was a nhtsa dataset from a consultant but I don't remember). And the OEM probably did 100s of tests before they got to the official facility, with the tier 1 doing thousands of tests before that.

I'm just scratching the surface but hopefully you get an idea.

 

[Tinerfeño]

I see some soft points in this discussion. mainly on the collision testing, not electronic assists (which I'm not interested)

1. Collision testing is made according to a standard. It is based on actual accident data but due to need for standardisation this setup is simplified. In actual traffic situation the test results are giving an idea how the car behaves. 5 stars are not a guarantee that you survive all accidents.
2. Test speeds relate to city traffic. In higway speeds there are no means protecting occupants because of high decelerations. This is a problem also in rally cars with safety cages and 5-point safety belts that keep to cabin space intact and occupants in seats but a sudden stop will kill because human internals can't stand it.
3. In straight forward collision modern cars have designed structures to collapse in a controlled way to increase time to stop and this way reducing the deceleration. Cars with strong chassis stop too rapidly and don't get good results in tests.
4. Additional problem in most 4x4 cars in real collisions is that especially in partial collision front tyre climbs to the bonnet of a modern passenger car with low and soft front. This causes vertical accelerations that are as harmful as longitudinal.
5. Pedestral safety has a practical limit at 40 km/h. Below that survival rate is high, above bad. Bull bars don't help pedestrians because they are too hign and too hard. I have winch bumpers with A-bars because it is practical for auxiliary lights and on collisions with elks and deers it saves car structures and possibly human lives in case of an elk. In my home town in Finland there is on average 1 deer accident every week an 1 elk accident per 1-2 months. I've been in one deer accident at dark conditions and 80 km/h and only my auxiliary lights were lost. White-tailed deer, about 100 kg, was killed instantly.
6. In some countries cycling may be for kids only but in most Northern European areas it is really common and practical means of travel. During me early career I used a bike lot, almost 10 years daily when dintance to work was just 3,5 km.

Summary: Drive reasonably concerning speed, distance to car in front of you, foresee what's going to happen, and note driving conditions. The sturdy chassis and body of IG will help you in case of some type of accidents, like multiple-vehicle collision, where the cabin space will survive, but in some cases not.

I have over 600000 driven kilometers and only that one deer accident. Couple of mild ditchings in slow speed. In this case snow plough had gone over the ditch and there was no road when I evaded oncoming car

 

[James]

I am working in the automotive space.

It's a combination of sensors but I think more important is the overall system latency (eg how much time it takes from sensor processing and until you have computed a solution to the vehicle control system). There are times of systems in current shipping vehicles that rely on very early detection because system latencies are very high.

Thank you, it’s fascinating. Not surprised by complexity, especially attempting to draw on and integrate different sensor and resulting data types, and under huge time pressure. Presumably machine learning around object analysis computationally intensive too.

 

[Tom109]

Complexity and oversight will continue to get more complex until the only solution will be self-driving vehicles.

 

[parb]

I'm not very skilled in structural engineering. That's a whole discipline by itself and euro-NCAP has a decent amount of testing of the shape of the front of vehicles and the impact those shapes has if hitting a human. It's an area I've only done cursory glances at and this is not my skill.

Sensors, computation and algorithms are making rapid advances. I think we are in the early days of pretty sophisticated systems to assist drivers with the dynamic driving task. I think the most complicated driving environment is the inner urban city. This is where robi taxis like Google waymo is aiming. But for consumer owned vehicles highways and suburbia is where the innovation is happening from the auto makers.

 

[Lord Ripon USA]

Good stuff guys.
I worked on steering systems when fresh out of college. 1980’s.
We put the radio controls in the horn pad.
1984 Pontiac set
Collapsible columns, intermediate shafts, steering shafts, etc.
It’s pretty amazing how far we’ve come.
Detecting roadside deer alone would say billions in dollars in MI alone.
Keep getting it done.

 

[ADVAW8S]

Looks like there was a crash in Germany this week with a Grenadier and a Van. I hope everyone is okay. Here is the link to the article and sounds like they had to use the jaws of life to get the person in the van out. The Grenadier looks like it survived and I hope it wasn't @emax As the color looks Donny Grey.

VERKEHRSUNFALL MIT EINGEKLEMMTER PERSON

Um 15:37 Uhr wurde die Feuerwehr über Sirenenalarm zu einem schweren Verkehrsunfall mit eingeklemmter Person auf die B311 alarmiert. Ein Transporter und ein SUV waren aus ungeklärter Ursache im Bereich "Russenbühel" frontal zusammengestoßen. Der Lenker des Transporters wurde dabei im Fahrzeug...

www.ff-stjohann.at

 

[anand]

Looks like there was a crash in Germany this week with a Grenadier and a Van. I hope everyone is okay. Here is the link to the article and sounds like they had to use the jaws of life to get the person in the van out. The Grenadier looks like it survived and I hope it wasn't @emax As the color looks Donny Grey.

VERKEHRSUNFALL MIT EINGEKLEMMTER PERSON

Um 15:37 Uhr wurde die Feuerwehr über Sirenenalarm zu einem schweren Verkehrsunfall mit eingeklemmter Person auf die B311 alarmiert. Ein Transporter und ein SUV waren aus ungeklärter Ursache im Bereich "Russenbühel" frontal zusammengestoßen. Der Lenker des Transporters wurde dabei im Fahrzeug...

www.ff-stjohann.at

Looks like absolutely no deformity to the passenger compartment of the Grenadier, so that's a good sign

 

[Pat]

Looks like there was a crash in Germany this week with a Grenadier and a Van. I hope everyone is okay. Here is the link to the article and sounds like they had to use the jaws of life to get the person in the van out. The Grenadier looks like it survived and I hope it wasn't @emax As the color looks Donny Grey.

VERKEHRSUNFALL MIT EINGEKLEMMTER PERSON

Um 15:37 Uhr wurde die Feuerwehr über Sirenenalarm zu einem schweren Verkehrsunfall mit eingeklemmter Person auf die B311 alarmiert. Ein Transporter und ein SUV waren aus ungeklärter Ursache im Bereich "Russenbühel" frontal zusammengestoßen. Der Lenker des Transporters wurde dabei im Fahrzeug...

www.ff-stjohann.at

Holy moly.
It was in Austria in the backyards of Salzburg. Yes, the driver of the Daily (lveco) is not in good shape what I can read from the article. Still unclear the rationality behind the crash.

 

[255/85]

Complexity and oversight will continue to get more complex until the only solution will be self-driving vehicles.

Consumer demand for ever increasing performance without regard for commensurate operator knowledge, skill, and judgement begets increased manufacturer/governmental oversight and vehicle complexity with the end result being fully automated (i.e. - self driving) vehicles.

It's a game of leapfrog. The great bulk of the car buying public is slowly screwing themselves out of the autonomous driving experience and travel freedom... and whistling past the graveyard oblivious to their peril.

 

[Bryan]

They send me this:

Dear *****,



Thank you for reaching out to us!



Unfortunately, it is not possible for Euro NCAP to test every car on the market. Euro NCAP is a non-profit organisation, faced with budget and time constraints, we select the best-selling European models and brands so as to reach as many consumers as possible. Therefore, we have no plans of testing IEOS Grenadier yet.



Hope that helps answer your query!



Health and cheers,

 

[parb]

They send me this:

Dear *****,



Thank you for reaching out to us!



Unfortunately, it is not possible for Euro NCAP to test every car on the market. Euro NCAP is a non-profit organisation, faced with budget and time constraints, we select the best-selling European models and brands so as to reach as many consumers as possible. Therefore, we have no plans of testing IEOS Grenadier yet.



Hope that helps answer your query!



Health and cheers,

If you pay them they will happily test... That's how that organization works, it's a non profit but it's not really community testing in that sense.

 

[[ Adam ]]

100% safer than a 1993 Defender 110

 

[Bryan]

If you pay them they will happily test... That's how that organization works, it's a non profit but it's not really community testing in that sense.

I think if more owners ask they will test the Grenadier.

 

[Bryan]

100% safer than a 1993 Defender 110

Even a modern Mahindra is much safer than an old Defender or any other old vehicle.