Ioniq 5 beats Tesla M3 in safety tests

Tesla removed the radar in the spring from the Model Y and A but legacy cars still have the radars. At an AI conference (don't have the link handy,) the unreliability of radar was shown with graphs. BTW, the early Toyota radars had a similar problem with fantom braking especially around bridge joints.

Bob Wilson

 

Offhand, it would seem that camera and perhaps radar based systems would depend upon driving conditions, e.g., road spray, rain, fog. How do independent testers evaluate performance over such a broad range of conditions, especially since I'd think it would be hard to mimic the full dynamic range in a controlled way. (Maybe I should do some reading before posting here.

 

AFAIK, Tesla does no advertising at all. Where's the marketing hype? In the feature name?
(FWIW, I don't use autopilot or cruise control, but YMMV.)

 

You might review the AI day presentation supporting Full Self Driving (FSD) and following comments by Tesla advocates. For example, at one point they tested their FSD coming upon a future cyber truck. Adding weather and light effects should be possible. As for IIHS/NHTSA/NTSB testing, they need to bring big bags of cash for their future test track(s).

Bob Wilson

 

They have so many more bicycles on the roads in Asia than the USA. Regardless, it would be an ‘over the air’ update.

It only makes sense there would be regional variations in the relative risks. For example, large diesel pickup trucks with red capped drivers are Alabama’s biggest road risks.

Recently Elon pointed out that FSD has a distinct San Francisco bias but then he and the developers are in the Fremont suburb. You don’t suppose the IONIQ 5 might have a Korean bias making it deadly in the USA?

Bob Wilson

 

That IONIQ 5 almost tagged that bicyclist.

 

Source: https://koreajoongangdaily.joins.com/2021/07/12/business/industry/tesla-electric-vehicle/20210712185706913.html

Apparently there has been no measured impact on Korean sales of Teslas. Just a trolling thread about a car not sold in the USA based upon a regional, Korean test. A test that thanks to over the air updates and FSD might be soon forgotten … outdated.

LATE THOUGHT:
Four legged animals are also common in some parts of the world. I wonder how both do with a donkey, horse, mule, and elephant.

Bob Wilson

 

I already subscribed to the IONIQ forum since they claimed to be better than the Prius I owned. I did a test drive and felt car sick. So I have insights others may not understand.

Lacking my background, I don’t fault their absence of either Prius or Tesla experience. Rather I bemoan their chauvinism about their ‘ride’ with no fingerprints on steering wheels of the cars they throw shade at. Deliberate ignorance does not inspire trust or confidence.

Bob Wilson

 

Maybe a car to go along with the 1997 Mercedes-Benz A class when undergoing the "Moose Test"?

 

Hyundai Motor Group EVs all have a reputation for meeting or exceeding their rated efficiency. Everyone I know with a Hyundai Kona, a Kia e-Niro, or a Kia Soul EV gets better range than advertised. This is not something that can be said of Tesla, except for perhaps some hypermiling outliers. At one point every single model of Hyundai/Kia EVs were in the top-10 list of efficiency, as recently as 2019 if memory serves. So I wouldn't be so quick to judge efficiency based on that metric.

But, in fairness, we don't know if E-GMP will continue the high efficiency trend that South Korean cars are known for. Let's see what Bjorn Nyland finds in more extensive testing of the two.

 

Having spent over a decade of Prius ownership and dealing with EPA, and 3d party reviewers, I'll go with the EPA. One of my favorite forums is "EcoModder.com," another set of home-brew engineers who seek maximum performance from their rides. I prefer engineering over wishes.

Of course, others are free to select whatever source meets their expectation. Just their assertions seem to lack reproducibility, physics, and basic math found in the EPA sources.

Bob Wilson

ps. Thank you! I went back to a favorite EPA data source and found significant problems with the Hyundai and Kia entries. Their reported MPG in the Test Car Database is significantly off. This is not the first time I've found manufacturers have screwed up their data report to the EPA.

 

I mentioned checking my source, the EPA Test Car Database, and finding a problem in the RND_ADJ_FE:

Not limited to just Hyandai and Kia, Ford also has a test car result with a bogus MPG. So what can you trust?

My experience is the vehicle gross weight and the "Target Coef A, B, and C" are usable. In particular, there is a formula that gives the horse power (746 W) as a function of speed:

drag HP = v * (A + B*v + C*(v*v)) / 375​

Most of this data comes from the manufacturer who typically outsource the testing. But the EPA reserves the right to replicate the testing and have fined manufacturers who submitted bad data. So who fines a journalist, publication, or random internet poster sharing inaccurate claims?

Regardless, here is a good place to start understanding the data the EPA gathers and shares:
https://www.fueleconomy.gov/feg/download.shtml

Bob Wilson

​

 

We have a Subaru Forester with the EyeSight Driver-Assist System (stereo lens sys). It’s a 2017 and the system worked well back then.

I also have a 2020 Hyundai Kona EV with one forward looking camera and it’s collision avoidance + lane keeping seems even better in every way. This for 2 vehicles not even pretending to aim at autonomous driving, just safety.


Sent from my iPhone using Inside EVs

 

Late thought, the fines and loss of credibility from exceeding the EPA claims makes 'sandbagging' the low risk option. The WPA values have historically been more optimistic. But with a methodology, I would be inclined to take their values over a layman's report.

Bob Wilson

 

Actually they exceed the 63-64 mph speed that in a Prius returns the EPA numbers. So there are two flaws: (1) driving too fast, and; (2) cautious under reporting of the manufacturers to avoid an EPA fine and black eye.

Why don't you cut and paste a dozen or a hundred times? It won't reverse the two problems mentioned earlier. Now I did appreciate their 3-day, 'smackdown' series several years ago:

• 3-5 cars driven on one trip to Las Vegas and back
• Mix of country, urban, and superhighway
• Swap drivers and share notes

It is called 'thread drift' and happens often. I don't sweat the small stuff. As for 'safety', we'll see what happens when they undergo USA testing ... someday.

Bob Wilson