Print 33 comment(s) - last by jconan.. on Jan 1 at 1:37 AM

PC technology filters down to the automotive market

BMW is no stranger to high-tech computer wizardry in its vehicles. The German company befuddled BMW owners around the world with the introduction of the Windows CE-based iDrive driver information center in the current 7-Series. iDrive, which has been panned by most enthusiasts and auto journalists, later filtered down to the 5-Series, 3-Series, 6-Series, 1-Series and X5.

Since the introduction of iDrive way back in 2002, many manufacturers have introduced their own control schemes with varying amount of control knobs and buttons to control everything from basic vehicle functions to the intensity of interior lighting -- most of which are far more intuitive than iDrive.

BMW is looking to make another leap forward with the introduction of Internet Protocol (IP) networking in future automobiles. The company is using off-the-shelf Ethernet components to replace the vast array of networking systems included in today's automobiles (CAN, LIN, MOST, Flexray, etc.).

It should come as no surprise that by settling on standard PC networking technology, greater cost efficiencies can be utilized as well as greater interoperability between car manufacturers. A wider array of standardized, off-the-shelf communications gear could lead to faster development cycles for vehicles and lower overall development costs.

BMW used IP networking to connect engine control units (ECUs), Dynamic Stability Control (DSC) and the dashboard head unit among other things. BMW's current testing uses the IPv4 standard, but the company hopes to move to IPv6 in future testing.

"One of our research goals was to verify the real-time capabilities of IP for safety-critical applications," said IP project manager Richard Bogenberger. "In order to guarantee the short response times required, we used features such as QoS and traffic shaping. Our experiments with prototypes demonstrated, that the real-time behavior far exceeded the requirements -- even when we ran multimedia applications across the same network."

BMW has no set time frame for when the first "IP Car" would be made available to the public, but chances are that we could see production versions within the next ten years.

Comments     Threshold

This article is over a month old, voting and posting comments is disabled

The other benefit
By djc208 on 12/3/2007 6:39:48 PM , Rating: 2
The one that popped into my head first is the aftermarket and add on possibilities.

How much easier would it be to sell new electronics that can integrate into the car with this. No more proprietary buses and communication setups. The radio in my car is linked to the bus to determine power, brightness, interaction with the steering wheel controls and factory amp, as well as hands-free setup and rear-seat entertainment. If you want to go aftermarket you need various adapters and modules to let the two talk. This could be as simple as plugging in the power and network connection and installing the manufacturer specific software to let the the radio understand the car.

Even the car manufacturers can sell add-ons and just leave open ports in the car. Remote start, buy the module and plug it into the hub. Backup camera, there's an empty port behind the bumper.

RE: The other benefit
By exdeath on 12/4/2007 12:29:49 AM , Rating: 1
Never going to happen.

Proprietary buses and hardwired peripherals = mega mega markup and profit margins for the dealer installed options.

Why would they let you install any $100 backup cam when they can sell you the "premium" BMW backup camera package for $8,000.

RE: The other benefit
By porkpie on 12/4/2007 1:20:27 AM , Rating: 2
Why would they let you? For the same reason every PC maker today "lets you" install generic hardware in their computers. Because if they don't let you, their competitors will, and you won't buy their cars any more!

RE: The other benefit
By stephenrapaport on 12/4/2007 4:22:26 AM , Rating: 2
Certainly this has some risk to it. But it is not impossible to design a safe Ethernet system. Also, the benefits and potential uses of this tech are huge. Besides possible car 2 car communication for potential automated driving, and weather or other emergency information. There could be intergration into the car's navigation system for directions anywhere anytime-- yeah GPS has this sort of, but not if you're looking for X's Chinese place on 22nd. Also, there could be real-time monitoring of saftey issues (kind of like on-star does now but less lame). Lastly, and possibly coolest (also least likely to happen while we are alive) imagine being able to control a Ethernet equipped car over your internet connection so it could come and pick you up, navigating itself using that same Ethernet tech. Now, granted it is a long long way off BUT how can anyone argue against the obvious potential and inevitability of having ours cars be connected to the internet. Are you REALLY surprised?

RE: The other benefit
By Chaser on 12/4/2007 9:45:21 AM , Rating: 2
Absolutely not. I am glad BMW is "thinking outside the box" but really its not. More like inevitable. Despite the naysayers here. :)

RE: The other benefit
By TomZ on 12/4/2007 11:38:13 AM , Rating: 2
A lot of what you are describing are benefits that could be achieved just by having a simple Ethernet gateway. The use case of connecting to external devices, other cars, or information networks absolutely does not require the entire in-vehicle network to be switched over to IP.

No, in order for this idea to make sense, it has to reduce node cost, reduce design time/effort/cost/risk, or bring new capabilities to the vehicle that cannot be easily achieved with other network types.

Remember, the BMW report was just a really basic feasibility study focusing on a couple of technical aspects like responsiveness. It did not attempt to make a technical or business case for actual adoption of IP to replace existing vehicle networks. That's still a larger question that remains to be answered.

RE: The other benefit
By InsaneGain on 12/4/2007 12:36:42 PM , Rating: 3
I'm starting to think of the potential for trojan viruses and hackers to take control of vehicles connected through a standardized network to the internet.

RE: The other benefit
By jconan on 1/1/2008 1:37:31 AM , Rating: 2
I wouldn't be surprised about being able to control my car from the internet. Rather I wouldn't want my car to be controlled from the internet similar to why galactica doesn't have networked systems. Since hackers already are capable of doing anything it wouldn't surprise to see where they are moving onto next. Probably no one would want their car controlled without their permission.

(sci-fi and reality are starting to blend. yesterdays readers today's inventors. lol...)

RE: The other benefit
By theapparition on 12/4/2007 7:14:43 AM , Rating: 2
This is already happening to some extent, although still very proprietary.

I'm having an alarm/remote start installed on one of my wife's vehicules. I've wired up alarm systems before, and everywhere you had to run wires. Want to control the door locks, you had to run wires to the doors, etc.

Now, you just interface with the computer bus. You want to unlock doors, send a signal to the computer to unlock the doors. You want to start the car, send the proper signal. Getting much simpler now.

RE: The other benefit
By Guigsy on 12/4/2007 8:46:35 AM , Rating: 2
Just because the network protocol is a common standard, it doesn't mean that any of the traffic on it will be readable or you'll be able to interact with other systems.

This is pretty similar to what's happening with mobile CPUs. You can make propriatory and specialist components/protocols for specific jobs such as mobile processors or car networks, but as soon as the common or garden technology that's the big stinking gorilla of an industry standard elsewhere can fit into the same space, it'll replace the specialists. It doesn't matter if it's ARM and XScale processors are used in PDAs because they are small and low power, or that another protocol has been designed for car safety, the industry standard is well understood and is easier and cheaper to develop for. The ancient architecture of the totally non-specialist 386 derived processors related to what you have in your desktop machine are being squeezed into phones, and PDAs. It means millions of pieces of software and development kit is instantly easy to port. IP is making its way into cars because the hardware is cheap and the faults are well documented and workarounds have been used for years.

However, I understand that all the BMW systems will all connect and play nice. The load balancing and QoS will prevent safety critical systems from overriding trivial systems. What happens if you plug in your new sat-nav? How will the system decide where it fits in the priority stack? Should have a higher priority than the stereo? What about the turning signals?

What would rock!
By afkrotch on 12/3/2007 5:04:27 PM , Rating: 3
Imagine if all BMWs were wireless and could be used as a wireless repeater. Then the move went to all cars. Would be a great way to throw broadband all over the country. Course nothing would suck more than saying,"I have to wait for my neighbor to get back. My net is down."

RE: What would rock!
By SiliconAddict on 12/3/2007 7:52:31 PM , Rating: 1
This is quite possibly the worst idea in the history of ideas.
Dear god the concept GAH.

RE: What would rock!
By spluurfg on 12/3/2007 11:34:17 PM , Rating: 2
Gee I don't know... the range on a small wireless repeater in a car would have to be pretty small, and traffic can only move so quickly with a very high number of wireless hops, so internet service would still be tied to a fairly close radius around a fixed link to a major backbone... (IE you can't hop across thousands of wireless repeaters across the country and get anything close to a decent service, and each one probably couldn't move much data... how would you design a routing system?). Plus, at the speeds these things could be traveling, one could lose signal within seconds and have to hop to another repeater/car.

Though this kind of thing does have its industrial application, I don't think we're at the point where you could really turn all cars into a gigantic dust/mesh network...

RE: What would rock!
By Moishe on 12/4/2007 8:14:17 AM , Rating: 2
We're not at that point, but we could be in the future.

If we had broadband wireless on the cell towers and repeaters on the cars you'd have the security of knowing which cars everything is routing through. That's as good of a "paper"-trail as you'd get.

You'd have network access in most places and cars could then begin warning of accidents, hard braking, sleepy/drunk drivers, etc.

I would want options for disabling external network on the fly. I'd want a DMZ or some other method for allowing net traffic to be routed through my car without it being able to touch my internal network. Basically, you set each car up like a corporate LAN. If there was intelligence making sure that switching to the shortest hop to the closest landline was fast, you'd have broadband almost anywhere there was a car and a cell tower.

RE: What would rock!
By spluurfg on 12/4/2007 11:01:36 AM , Rating: 2
Yes, but considering the current 802.11 wireless implementation, I think the range limitations given the speeds vehicles are travelling at would make it extremely difficult to secure even one stable wireless connection for any meaningful length of time. Plus, the short range of say, 300 feet would mean that being just a mile away from a network anchor would require dozens of hops over -wireless- networks. I don't think that it can route significant traffic, even if one could stay connected to a single AP for more than a few seconds, or devise a way of routing traffic that will adapt to a different route through dozens or possibly hundreds of hops many times per minute.

Besides, why bother having every car as a repeater? That implies that we are relying on a very short range wireless technology if we are aiming for such saturation. Wouldn't it be easier to say, simply build an effective 3g or 4g wireless broadband infrastructure? Which, by the way, pretty much going to happen already?

Using IP huh?
By SiliconAddict on 12/3/2007 7:47:09 PM , Rating: 2
Guaranteed to get hacked now. Why can't company's stick with CANbus networks?

RE: Using IP huh?
By TomZ on 12/3/2007 8:53:03 PM , Rating: 2
CAN is too slow.

RE: Using IP huh?
By Deceiver on 12/3/2007 11:42:48 PM , Rating: 2
CAN is too slow.

Which is what FlexRay is for. It goes up to 10Mbps, has dual channels for redundancy, has static slots for deterministic control data, and dynamic slots for other info.

I know BMW is one of the first to use FlexRay, but it seems to be a much better physical layer than ethernet. That is just scary to think about, and I'm surprised they are already moving away from it.

RE: Using IP huh?
By yawnbox on 12/4/2007 1:59:57 AM , Rating: 2
I am guessing that FlexRay is more than just a protocol, whereas it is built to integrate many of the respective OSI model layers. I am also guessing that FlexRay is a better choice because it's primary purpose is to transmit data within the car; however, this is a big difference when compared to Ethernet and the use of IPv6. I think that Ethernet would do a much better job when having to network with external nodes, also allowing a great deal more expandability. I would imagine that a great deal of BMWs think tank for automotive IT has to do with the integration of both kinds of technology, and/or replacing one or the other.

RE: Using IP huh?
By TomZ on 12/4/2007 8:44:07 AM , Rating: 2
Sure, FlexRay is faster than CAN, and its adoption is on the rise.

Regarding Ethernet, I also agree that it is probably a poor choice for the autmotive environment. A couple of things come to mind quickly. First is wiring - Ethernet requires more wires than CAN or FlexRay which adds weight to the car. This is opposite current design trends which are to reduce weight by simplifying wire harnesses. Also the RJ45 connector would not be suitable for automotive, although it is a simple task to integrate Ethernet into existing automotive-grade connectors.

Second, cost per node of Ethernet is probably higher since it requires a physical layer driver which can be implemented on-chip, but that takes a lot of silicon. Plus Ethernet requires off-chip magnetics. This drives the cost higher than CAN or FlexRay. Node cost is a very important consideration.

My guess is that if IP is used in cars, that the industry might develop an alternative physical layer for it. This gives them most of the benefits of sharing the same protocol as PCs and the Internet use, but keeping cheap nodes and wiring.

By cjc1103 on 12/3/2007 5:07:29 PM , Rating: 1
Our experiments with prototypes demonstrated, that the real-time behavior far exceeded the requirements -- even when we ran multimedia applications across the same network."

Let me get this straight. They are running DVD's over the network that also controls things like speed control and anti-lock brakes?? I know Junior wants to see Pirates of The Carribean in the back seat, but shouldn't the essential car functions be on a separate network? Let's hope they're not running Windows CE.

RE: Multimedia?
By TomZ on 12/3/2007 5:13:48 PM , Rating: 2
Automotive networks are already segmented to solve this problem with today's physical layers (CAN, etc.). I assume that Ethernet/IP networks would also be segmented in the same way. Automakers cannot take the chance that the entertainment systems messes with the powertrain communications.

RE: Multimedia?
By ZaethDekar on 12/3/2007 5:18:16 PM , Rating: 2
They already have settings for IP traffic shapping.

Doing a test at a training I was in for HP ProCurve, we had a video streaming from the main computer, and then we were supposted to figure out how to make it slow. What the instructor didn't tell us is that he had it at the highest priority.... so even when people were maxing out the network the video never stuttered, well until we set the priority up on the traffic generator.

I am sure they will make a setup where the cars main components will have the highest priority and they will lock out other programs from accessing it. Otherwise that would be a serious security problem.

RE: Multimedia?
By Clauzii on 12/3/2007 5:53:26 PM , Rating: 2
Like a rocket-system where vital functions ALWAYS come first, even if the DVD or the aircondition screw up?

RE: Multimedia?
By Moishe on 12/4/2007 8:18:02 AM , Rating: 2
ideally, the car's critical systems would not have to compete for bandwidth/processing time. I think the user controlled stuff should be on a separate node.

I really hope..
By Clauzii on 12/3/2007 5:05:16 PM , Rating: 2
.. that when going WiFi on cars, they know exactly how to keep security top notch!

RE: I really hope..
By exanimas on 12/3/2007 8:25:42 PM , Rating: 3
Just imagine it now, "O NOES, MAI BREAKS HAB BEEN HAXED!"

How about this
By AlexWade on 12/3/2007 9:20:24 PM , Rating: 2
Imagine if there was short-range wireless car-to-car communication. The car would have GPS and wirelessly broadcast its location, whether it is braking, and some other information. For privacy, the data is never stored. Then, a computer can read all the cars around it and sound an audible alarm if a car is doing a hard brake and in the future be used to automate driving without complex programs that use cameras to study the road. This would, of course, require a non-Windows based system.

Of, the car is connected to a secure network. If severe weather or some other emergency is happening, the alert is pushed through the internet to the cars in the affected area. You will be able to get traffic updates too. Passengers will be able to get net access. Again, Windows would be a no-go for this application.

Whatever is used, there must be strict privacy requirements.

RE: How about this
By TomZ on 12/4/2007 8:36:27 AM , Rating: 2
Windows CE is ideally suited for automotive applications, especially those with GUIs. It's a real-time operating system with deterministic response time and has been used a lot already within the industry. So I'm not sure why you're excluding from these types of applications.

So much for customer feedback...
By jskirwin on 12/3/2007 6:17:54 PM , Rating: 2
The German company befuddled BMW owners around the world with the introduction of the Windows CE-based iDrive driver information center in the current 7-Series. iDrive, which has been panned by most enthusiasts and auto journalists, later filtered down...

Customers don't like it? So what! We'll make it standard across all vehicles.

Don't you mean.
By exdeath on 12/4/2007 12:26:36 AM , Rating: 2
"BMW includes a Linksys IP router in automobile" rather than "BMW develops"?

Or did BMW also invent the internet?

By kalimystic on 12/13/2007 5:43:15 PM , Rating: 2
Gosh- I thought of ALL companies in the world BMW would have been "engineered" to be better then using a Windows based ported System V - TCP/IP stack. But then again, I can't wait to hotwire my neighbors car and run it through the garage (over and over and over again).

Forget personal computers-- I just hack-jacked a 7' series..

I am outta here.....

By bcalcote on 12/3/2007 10:29:56 PM , Rating: 1
WWWWWhat... WWWWWhat... WWWWWhat... YEAAAAAA!!!! - Lil Jon

"It seems as though my state-funded math degree has failed me. Let the lashings commence." -- DailyTech Editor-in-Chief Kristopher Kubicki

Copyright 2016 DailyTech LLC. - RSS Feed | Advertise | About Us | Ethics | FAQ | Terms, Conditions & Privacy Information | Kristopher Kubicki