I was looking through a dealers in SoCal's inventory today (in search of a specific GLE) and I stumbled across a 2024 EQS Sedan listed for sale. I didn't realize they had been released for sale yet, but low and behold they have and this particular one had some secrets! It's vin was slightly lower then I'd expect for a 2024 but sure enough it has the conditional autonomy option! It's at WI Simonson. I'm sure there's plenty more similar cars in CA and NV.


You can see the extra LIDAR in front


And the steering wheel buttons

This has the conditional autonomous driving. It's code DA9 for the package and then includes 200 as the actual code, P26 instead of P20 for the package code, and then PBH intilligent parking package with 27U, 503 and 507 which could be enabled one day for the autonomous parking garage parking. Don't think it has any other functions.




Here's what the printed window sticker looks like


This also has all the other 2024 updates inclujding the availability of the rear entertianment, the standard heated power rear seats with a slightly new design, the flex charging cable that can do level 2, the front motor disconnect and heat pump, etc.

 

That’s exciting! When my lease ends, I’ll be leasing a 2025 EQS580. It should have all the above and hopefully more options.

 

Crito's wallet is inoperative, visit workshop.

 

The lidar is hideous. Im sure its plenty expensive and of limited utility unless you plan to spend your day to day in 40mph bumper to bumper. On the other hand if MB didn't charge for it since in effect what you will be driving is a development mule I would consider it😉

 

$2500/year to get autonomous driving only in stop&go traffic below 40 mph on a limited number of highways in CA and NV...ROFL!

 

I knew the specs, however, its incredible that they expect anyone to pay to become a beta tester for their system. Crazyland.
And yet... we cant get the full LED headlights, instead we get a disappointing crippled version that spoils the low beams.

 

In their defense, that one is on the US DOT. Even with the new regulation now technically allowing the matrix headlights, the DOT still left outdated parts in the new rules that make the EU version of the matrix lights still illegal in the USA.

 

So - hypothetically - if i change a US car's software to allow for digital light (hardware already in the car) resulting in a better and safer lighting system, I would break the law?

 

Ironically yes. The US law for example requires lower light output than is allowed in Europe. It doesn't even take much to activate them. BenzNinja etc. can code it. However, the feature that folks are after only works outside of metropolitan areas. The system uses the navigation maps and won't allow for the "high beam" part to come on in cities etc. You have to be driving out in very dark rural areas for the system to really add any value. I just don't drive in areas at night where it would make a difference.

 

I do not see an advantage in cities. It works great in urban areas when (in Auto mode) the headlights switch automatically between low and high beam.

 

I agree with others, I don't know how many people will pay 2500$/year for the current implementation of Level 3. It is extremely limited and to be honest the current driver assistance systems we have in 2023 EQS is really good for highway at 40mph anyway.
The article I read earlier stated California law still would still prohibit drivers from using their phones even when level 3 is engaged and obviously you can't take a nap so I fail to see the point of it right now. It really should have been a free option to let people test it and let Mercedes collect data.

With the current limitations on weather as well, I don't really see it being enabled in Washington state anytime soon anyway

 

retrofit level 3 on 2022 and 2023 models possible?

 

I am pretty sure, it is not possible as it would require LIDAR (see the 2 sensors in front of the sedan in the photo). The car will also have blue exterior lights to indicate Level 3 being engaged.

 

I'm not sure why anyone would want to go through the trouble and cost to retrofit the hardware for a system that only works below 40 mph and can't be driven at night, in the rain etc. The LIDAR sensors would probably be the easiest to retrofit, but there are a number of other sensors and the high precision antenna array that is required for the more precise GPS needs a new roof with a hump. In addition to the sensors it requires all the computing hardware.

 

Level 3 requires lidar since laser can not only identify but also range an object simultaneously much much more precisely than radar can. However you essentially have to teach it what everytning it sees is! With 2d object recognition, AKA camera, you have to do the same thing but the ranging portion is separate (using radar) and thus much less precise. The fact has been and continues to be that machine sytems are superbly primitive compared to 3 billion years of evolution. The eyes can automatically, recognize miilions of objects, including millions of objects that have nothing to do with driving, at long distances and respond almost instantly to them, basically reflexively. No electromechanical contraption ever designed comes even close yet. For example take MB's video of their level 3 system at work, its all carefully choreographed, throw a kid chasing a ball or a granny falling in the street at the same time that its doing its maneuvering and there is zero chance of the system ever figuring out what to do with such a scenario.

 

It's not our eyes that do this, but our brain's ability to process the signals it gets from the eyes. The fascinating thing is that our eyes send so much detailed data, that our brain can't process it all and compresses the information it receives. This is evidenced by us often not noticing stuff around us that hasn't changed. Our brain is very much conditioned to process information that relates to things moving around us and fills in the static parts with old data actually. It's been shown that we see certain things time delayed, because the brain deems it irrelevant. This is why games like Find Waldo can be so difficult for us. I just read an article yesterday I think about a group of scientist that have grown a small human brain on a chip and it just recognized human voices. The current AI technology is really still rather crude compared to the human brain.

As far as LIDAR goes, LIDAR works much better in many different conditions. Cameras don't work very well at night and can be blinded by the sun during the day and basically result in a whiteout, and radar can have issues when it rains. This has led to some of the Tesla crashes after Musk decided they no longer need radar, even though his engineers told him that it was a BS decision, but he just wanted to cut cost. The radarless system famously caused one Tesla to crash into a crossing big rig, because the cameras didn't see it against the bright background. Teslas now have radars again, but Musk still refuses to admit that LIDAR is required if autonomous driving should ever be achieved.

Even so, we are so far away from full self-driving that all this hype is starting to get old. Not sure how Musk and company envision for self-driving to ever work in a Blizzard or snow covered roads in general. GM's subsidiary which was running autonomous taxis here in San Francisco is at the brink of folding since it turned out that their taxis weren't as self-driving as they claimed. A remote operator had to intervene every 2-5 miles, so they were basically remote controlled taxis.

 

Polarized glasses delete half the data and that makes it easier to see. And it's not even a selective deletion of what my brain thinks is irrelevant, it's just half the data. LOL

My problem with full self driving is I just don't trust the people making the equipment, especially the "telematic control unit" made in occupied parts of Palestine. I might enable it in my China-built Lotus Eletre though, when it becomes available.

 

It is still selective, though, because the data that is deleted is mostly noise such as reflections.

 

The brain was implied.... since the eyes is simply a a "sensor". Quick factoid, the eye has an estimated 576 megapixels in an area of less than half of the sensor of a 35mm DSLR. That's the crappy human eye, a bird of prey has many times the acuity. The cameras used in nav systems are lucky if they have a mega pixel or two. Those eyes are connected to a massively parallel electrochemical "computer" can control hundreds of semi autonomous "systems" while at the same time drive, talk, listen to music, and fantasize about 3 different things at the same time! BTW the cerebellum, by itself, the part that controls muscle movement by itself has billions of electrochemical "transistors" all of which again, work in a massively parallel fashion. When fools talk about AI as it were going to happen within any of our lifetimes I have a good laugh at the utter ignorance that computer "experts" have about biology.

 

In 1900, who would have thought we would fly and have dogfights in the first world war? Technological evolution is much faster than biological. It will only be limited by the data sets available. The first generation of level 3 will collect the data. The subsequent generations will utilize those data. Driving experience is not infinite because essentially, we are all driving in a closed environment . . . unless you are going off road to an unrecorded place. The obstacle to full self driving is the bean counters and law makers.

 

Sure, there's certainly truth to that, but we don't even understand how consciousness really works, so unless we create sentient AI by accident you kinda have to understand the problem first before you can recreate it. As for self-driving, Tesla Autopilot is gonna turn 10 years old next year and it still drives worse than a student driver overall, and can't handle many situations. They've been collecting data for 10 years and Musk keeps promising that FSD is only one year away. I think people who think that this gets solved in our lifetime just don't understand the complexity.

The engineers working on autonomous driving approach the problem purely visual and I mean visual in a broad sense including the infrared and ultrasonic spectrum. But anybody with driving skills that go beyond commuting to work in 40 mph stop&go traffic will tell you that what you hear and feel is just as important. That's why every car review these days is complaining about lack of steering feedback in modern cars. The feedback through the steering and the seat of your pants tells you a lot about what the car is doing and the condition of the road you are driving on. It tells you instantly if you are driving on a slippery surface for example before ESP has even a chance to pick up on it. For example, I suck at driving sims, because they lack the physical feedback unless I'm strapped into a full sim rig simulating all the forces and sensations in an actual car.

To bring back Cruise as an example of where this is obviously going wrong, they first started to really get into deep water after one of their cars hit a pedestrian. The initial hit wasn't the Cruise car's fault. The pedestrian got hit by another car and thrown into the path of the Cruise car. The Cruise car stopped, but that's where it started to go wrong. The pedestrian ended up under the car and outside the view of the sensors, so the car thought everything was clear and proceeded dragging the trapped pedestrian along with it. This is where other senses become very important.

 

Even if at some time self driving were to equal or surpass human ability (we all are very aware of human ineptitude ) this situation plainly shows the legal peril involved. Thats why MB manuals have more print devoted to warnings than to actually giving useful user information!