By on November 30, 2018

Successfully operating self-driving cars on crowded, complex roadways in sunny, dry locales like Phoenix, Arizona is already enough of a challenge, but researchers in the cold, tempestuous climes of Michigan have revealed what the latest and greatest autonomous technology is really up against.

Rain, sometimes hard rain. But also light rain. Also: cold temperatures, and trees with leaves that fall off in the winter. Given that so few places in the world boast such extreme weather and vegetation anomalies as Michigan, this won’t pose a problem for the widespread proliferation of driverless cars, will it?

While adverse climatic conditions can wreak havoc on cars, decades of testing and development have yielded batteries that crank harder, wipers that turn on at the first touch of rain, and an endless list of other stress-relieving advances. For driverless cars, it’s still the Wild West. Early days.

The proximity of Detroit Three HQs and development centers means Michigan State University finds itself at the center of things on this new frontier. There’s already a driverless car test facility, and school created the CANVAS program (Connected and Autonomous Networked Vehicles for Active Safety) to study how well the driving systems handle the tasks of avoiding things they shouldn’t hit. Part of that research included analyzing how weather impacts the sensor arrays and cameras used to guide the vehicles.

While MSU’s findings will soon appear in a study, researchers gave the skinny to Automotive News. Basically, autonomous car systems need a lot more work before companies can be assured of reliability outside the temperate Southwest.

The worst culprit when it comes to pedestrian and cyclist safety and driverless cars is rain. Raindrops confuse the algorithms used to detect such objects, claims Hayder Radha, the MSU professor of electrical and computer engineering who oversaw the study. As radar and lidar can’t do all the seeing, cameras are also used to guide a vehicle’s path. What the camera sees must then be interpreted.

“When we run these algorithms, we see very noticeable, tangible degradation in detection,” Radha said. “Even low-intensity rain can really create some serious problems, and as you increase the intensity, the performance of what we consider state-of-the-art mechanisms can almost become paralyzed.”

The presence of rain has stymied testing of production automobiles fitted with the latest driver-assist systems in the past. Camera-based systems used for lane-keeping and other functions are susceptible to a rain-caused algorithm confusion. To gauge the impairment, MSU tested a variety of conditions. Raindrop size varied, as did the concentration of raindrops. Wind was factored in, too.

From Automotive News:

Using a scale that ran from clear weather to a blinding rainstorm, they found algorithms failed to detect as many as 20 percent of objects when the rain intensity was 10 percent of the worst-case scenario. When rain intensity increased to 30 percent, as many as 40 percent of objects could no longer be detected.

Other problems cropped up during the changing of the seasons. High-resolution mapping used to help guide self-driving cars becomes less useful when the landscape changes in winter. The removal of foliage means landscape simply doesn’t look the same.

“You can imagine in environments where there are a lot of leaves on trees or on shrubs close to the road, they are an essential part of the map,” Radha said. “So summer and winter are completely different. When they fall down in winter, you have nothing to work with.”

Cold temperatures also wreaked havoc on the high-tech systems, with temperatures of 10F and below bringing about a marked uptick in the amount of “noise” returned by lidar sensors. To accurately map out the landscape in front of the car, lidar can’t send back hazy returns or false positives. Compensating for this could lead to a dumbed-down system that’s less safe and responsive than intended. Uber learned this lesson in Arizona in March.

Radha said suppliers are attempting to develop sensors that operate effectively in cold conditions, knowing that driverless cars won’t be able to catch on until weather no longer poses a challenge. As for snow, well, we’d be very interested to know how companies plan to contend with that challenge.

[Images: Ford]

Get the latest TTAC e-Newsletter!

Recommended

46 Comments on “Michigan Testing Shows Fairly Innocuous Weather Baffles Self-driving Car Systems...”


  • avatar
    Steve Biro

    And this is why driverless cars won’t be practical until vehicles speak to each other — and the infrastructure. And why all driver assistance technology should be able to be turned off — and remain off through key cycles. It simply isn’t ready for prime time — no matter how badly some people want it.

    • 0 avatar
      APaGttH

      Look at how hard this is in aircraft technology, where there are a lot fewer risks once airborne in a three-dimensional plane and where the infrastructure and aircraft do talk to each other.

      Lion Air pilots fought a faulty safety system twenty-six times, never realizing that turning off the safety system would stop the aircraft from pitching nose down for no reason. An entire 737-MAX full of passengers died because the system failed and the pilots didn’t know how or weren’t aware enough to turn it off.

      • 0 avatar
        MBella

        Usually in Aircraft, when the computer can’t process what’s going on, the auto pilot deactivates, and it’s up to the pilots to take over. I was on a flight where some sudden windshear hit the plane close to the runway. Pilots are trained for this, and have their hands on the throttles, ready to initiate a go around. Imagine when the average crappy driver, who now has had their driving skills fade because they haven’t driven all year. Now, the car tells them to take over, right as things get really bad. How long is it going to take for them to regain situational awareness? Couple that with the lack of driving skills, it might get real bad.

        The 737 MAX situation is weird because this is the first time I know of where the computer is programed to accept a potentially faulty sensor reading. Angle if attack disagreed, and the system accepts the higher angle as accurate. The MCAS now starts automatically trimming the stabilizer up to push the nose down. I find it amazing how much Boeing dropped the ball on this. They introduced a new system where the horizontal stabilizer is automatically adjusted even when AoA sensors disagree. They didn’t train pilots about the presence of this system. Then, they made the Angle of Attack display (which is the only display that shows the AoA disagree message) optional.

        • 0 avatar
          krhodes1

          I disagree – the PILOTS dropped the ball here. Effectively, what happened to the Lion Air flight was a runaway trim system. WHAT caused the runaway is irrelevant – the pilots should have been able to recognize that issue and deal with it as they are trained to. And as the pilots of the previous flights did successfully… Whether it was this new system, a short, an autopilot malfunction or any of the other various things that can cause a trim runaway is irrelevant. This is not some invisible thing – the manual trim wheels would have been moving and very obvious that it was what was causing the nose over.

          The absolute majority of the blame here lies on Lion Air’s maintenance and pilots. The plane never should have been dispatched in the first place, and the pilots completely missed what was going on.

          I have no doubt that Boeing will change the programming of this system, but ultimately there was little need for the pilots to know the root cause here – they just needed to recognize a trim system runaway and deal with it. It’s not possible to train pilots for every possible scenario of what can go wrong with the aircraft, only the common things.

          • 0 avatar
            sgeffe

            Unusual to see something like this from Boeing!

            Although you’re right about seeing things like actuators and controls MOVING should have provided reference. (Unlike AF 442, where the co-pilot’s stick being used to aim the plane at the moon couldn’t be seen in the left seat!)

            Does Boeing utilize the concept of “normal law” and “alternate law” when describing flight-envelope protection, like Airbus?

          • 0 avatar
            krhodes1

            The 737 is not a fly-by-wire aircraft. MCAS can only move the stabilizer trim, not the flight control surfaces.

            But even on their fly-by-wire aircraft (777 and 787), I don’t believe it works like in an Airbus. Boeing gives the pilot a lot more authority. Where Airbus has had trouble in the past is in not clearly communicating to the pilots what mode they are in, and what the computer will and won’t do at that point. AF447 being the poster child for that – the pilots didn’t believe they COULD stall the plane! And being a pitch dark night they couldn’t see what attitude they were in. if they could have seen the horizon out the windows, the crash never would have happened. Of course, in a FBW Boeing aircraft the controls still move and are linked, so it would be unlikely to happen due to that.

            Interestingly, I have read that the reason for this system is that in order to fit big enough engines to the 737-MAX, Boeing had to move the engines farther forward than is optimal. This can cause a very strong pitch up under certain circumstances. MCAS is meant to counteract that, with additional stall protection as a side benefit.

            Unfortunately, most aviation knowledge is paid for in blood. I have no doubt there will be more indications added to the cockpit as a result of this.

          • 0 avatar
            MBella

            An air accident is almost always a chain of problems. Yes, the crew could/should have recognized the stabilizer runaway. The system operating on sensor data that directly conflicts with another like sensor is a design flaw. That is 100% on Boeing. I am not aware of any previous time when either a Boeing or Airbus automated system would make a correction while receiving conflicting data. At that point on other systems, they would get an alarm, and it would be up to the crew to decide. It’s hard to blame to crew. They were receiving the stock shaker since nose up, alerts for Altitude disagree and airspeed disagree. The Angle of Attack disagree message wasn’t present, because the Angle of Attack display wasn’t fitted. Obviously they could have stopped the runaway stabilizer because the previous crew did, but humans are imperfect. I don’t understand why that error message isn’t present somewhere else if the AoA display isn’t fitted. Southwest just ordered all their MAXs to be fitted with them for this reason. As you said, Boeing is going to have to prevent the MCAS from operating when it received conflicting data. The AoA display will likely also become standard.

  • avatar
    2drsedanman

    Somewhere there are attorneys laying the groundwork for personal injury and wrongful death suits for future use when/if this tech becomes widely available.
    “Have you or a loved one been a victim of an autonomous vehicle accident……………”

    • 0 avatar
      jpolicke

      Well eventually all the asbestos, 9/11, faulty medical device victims will have been paid off, and something has to pay for The People’ Court and Gunsmoke reruns.

  • avatar
    R Henry

    ….yet Detroit and all the others throw money down this dead end street in a great Autonomous Vehicle pissing match.

  • avatar
    d4rksabre

    “Cars that drive themselves only just as bad as cars driven by human beings” seems like progress to be honest.

  • avatar
    Rnaboz

    LOL! Wait, with global warming, the snow will be gone!

  • avatar

    Anyone who thinks these technologies are ready now for prime time is dreaming.

  • avatar
    dividebytube

    A few years ago I had to drive my parents back to their home located 70 miles away on the coast of Lake Michigan. The night before we got lambasted with snow; maybe one of the snowiest Michigan winters I remember since 1978. Luckily I had the T100 4X4 with snow tires. It was a tight fit for everyone but I got them back in one piece. I didn’t stay long and headed back home.

    As I was driving I thought about self-driving cars. How would they handle the conditions I was in? I mean you couldn’t even see the road, or even the sides of the road or the lane markings. They were just covered in a foot plus of the white stuff. It was just a tunnel of trees on the side and slippery white stuff everywhere. I knew where I was going because I had driven that same stretch of highway hundreds of times.

    I obviously don’t know the technical limitations of LIDAR, nor the overall accuracy of GPS systems. But imagining a system that could handle this – along with the traction concerns – made me wonder how it could be done with (near?) 100% safety. I guess we’re going to find out.

    Or maybe the self-driving car will start flashing “Environment Overload” and drive gently off onto the shoulder. ;)

    • 0 avatar
      krhodes1

      I think almost certainly the car is going to just pull over and stop if it can’t figure out what is going on. I assume it will be like the various automated “people mover” things at airports – in an emergency, there are manual controls that can be deployed to run the vehicle at low speed. Of course, if this happens in a blizzard and you freeze to death, well, technically it wasn’t the CAR that killed you…

  • avatar

    Rubbish technology that will never see the light of day. Don’t even bother covering it anymore.

  • avatar
    Mike Beranek

    HAHAHAHAHAHAHAHAHA!
    Yeah let’s see how AVs work when the road isn’t even visible because of whiteout snow conditions. A human driver can detect the faintest remnants of the tire tracks of the last vehicle and follow them. Sometimes you can’t see them and you have to “wing it”. I think we are many years, or even decades, away from technology that can do that.
    So the dry, sunny southwestern climes will be filled with AVs, but up here in the Snow Belt there won’t be any. How’s that going to work?

    • 0 avatar
      forward_look

      We have an air base and the first snow of the season (which was 3 weeks ago) the newly transferred airmen from the south are in the ditches. The ones with SUVs that think they can drive through anything are over the guardrails.

  • avatar
    Master Baiter

    My BMW’s rain-sensing wipers work pretty well, but sometimes they get fooled. And that’s just trying to control the wipers–a task that’s a million times simpler than driving.

    Self driving car? Not gonna happen without dedicated roads and infrastructure, which we can’t afford because we’re broke and corrupt. Maybe China or South Korea can pull it off.

  • avatar
    APaGttH

    If we extrapolate.

    The GM EV-1 was released in 1996. Outside of idyllic climates, the first iteration was almost useless in cold climates due the impact of temperature on battery technology of the era. It took about 20 years for technology to catch up (roughly) to create a viable electric vehicle that could seat at least 4, provide enough utility for daily use, and go more than 200 miles for under $35K.

    So technology is likely around…20 years away from being close to Level V autonomy. Huh, and that is what a lot of experts are also saying.

    • 0 avatar
      28-Cars-Later

      Electric car technology had already existed prior to EV-1, autonomy did not exist prior to recently. Therefore there is a whole other stage of discovery which wasn’t necessary for EV-1 and subsequent iterations which is necessary in the latter case.

      • 0 avatar
        APaGttH

        The idea of autonomous vehicles and potential solutions to the challenge have been around since the 1930s.

        Yes, electric cars existed over a century ago, and were non-viable. The EV-1 was non-viable out of the gate because the technology behind lead-acid and Ni-Cad batteries made it untenable. Not to mention the lack of existence of safe, reliable fast DC charging that could easily be installed in a home, the computer and IC systems, composites, and other space/weight saving technologies that didn’t compromise safety in the name of reducing weight.

        I mean shoot – if you want to get all E. B. White Symanticsdome on me, Davinci had working concepts of aircraft in the 1500s – but the powered flight wasn’t viable for almost 400 years later because…wait for it…the technology to make it viable didn’t exist.

  • avatar
    indi500fan

    “Given that so few places in the world boast such extreme weather and vegetation anomalies as Michigan”

    Michigan may have more of it and more often, but pretty much everywhere else gets it sometimes. Plenty of places in eastern California and northern Arizona get lots of real winter.

  • avatar
    Jeff Weimer

    TBF, neither can many humans, if what I’ve heard about the rains in LA and San Diego yesterday are true.

  • avatar
    namesakeone

    As long as big business licks their chops at the thought of shipping merchandise cross country without having to pay a truck driver’s salary, the push will be on for autonomous vehicles, and damn the risk.

    • 0 avatar
      Jagboi

      They already do this. It’s a trivial exercise to put 300 containers on a train with 2 crewmembers. Oddly enough, trains have also managed to navigate through snow and rain. It’s a miracle!

      • 0 avatar
        MBella

        This is where the real answer to many things are. Instead of focusing on stupid passenger trains that Boone wants to use, improve the cargo infrastructure and get the Long haul trucks off the roads. Use trucks to only delivery from depots.

      • 0 avatar
        Wheatridger

        I like the way you think. So let’s make it easy for the car manufacturers. Let’s embed a steel rail (or two) in every lane of every roadway. Then the cars can follow that, magnetically.

        • 0 avatar
          jpolicke

          While you’re at it, why not run power through the rails, which the cars can access through pickups that stick out the bottom of the car. Then to make it foolproof, cut a slot between the rails and put a guide blade between the car’s pickups.

          Tesla, meet Aurora.

          • 0 avatar
            MBella

            That wouldn’t be the dumbest idea. You also wouldn’t be stuck moving the weight of a large battery around.

  • avatar
    Spike_in_Brisbane

    I am looking forward to video of the first incident of road rage between a driver and his own car. Oh wait, John Cleese already did that!

    • 0 avatar
      thegamper

      That will totally be me yelling at my car. “Dammit, quit driving like an effing clown!”. “What the H is wrong with you?!!”. “pick up the pace, get over already, plan ahead for Christ sake!”.

      I’ve got issues.

  • avatar
    SCE to AUX

    And to think Ford and GM are betting the farm on this stuff.

  • avatar
    Superdessucke

    They’re going to need to program it to do the famous Midwest Rock in the snow.

  • avatar
    thegamper

    If only there were this amazing piece of engineering that could distinguish between different types of objects, rain or shine. Something that could function in the cold. Something that could control a car and make split second decisions about the best course of action. Quickly process changes in environment. Something that could even adapt to the situation, use executive decision making to buck rules in the name of safety. Something that will prioritize human life. Gosh. If only.

  • avatar
    tonyquart

    I think this is why I would not risk my family to ride inside those cars, not until they reach at least 95% safe and perfect. I have just read at https://www.lemberglaw.com/self-driving-autonomous-car-accident-injury-lawyers-attorneys/ about this topic. I personally think that it still needs at least 10-15 years from now for the automakers to make a safe self-driving cars. until then, I hope that they don’t test these cars on public roads. There are too many risks.

  • avatar
    dusterdude

    From the challenges that still have to be solved with autonomous vehicles, I hope it is at least 15 – 20 years before it is widespread. Simple things like a plastic bag blowing across the road just in front of an autonomous car can be a challenge, not to mention snow etc). I’m stating that I HOPE that it is at least 15 years away, as I know business wants autonomous vehicles asap ($). Hopefully our governments evaluate safety concerns carefully before approving… but when there is big money involved I wouldn’t be so sure..

    • 0 avatar
      sgeffe

      They were resurfacing the main street into downtown Toledo and leveling manhole covers, using the big steel plates to place over the top. Several mornings in a row, my radar-based forward collision-warning went into full alert approaching and driving over the top of one! I can only imagine if my car was a 2016-7 Accord with the auto-brake, and that hopefully I’d have the presence of mind to floor the throttle, overriding the system, before being punted into low-Earth orbit by the mouth-breeding soccer mom who wasn’t paying attention in the SuperGigrundoMobile behind me!

      And as others have said on here, one raindrop can do a number on a backup camera!

      From November-April, roads in most of the Midwest vary from clear and dry to two feet of snow or glare ice, and some variation of black ice and/or a slushy, sloppy mess anywhere in between! Imagine my SHOCK at seeing that these GoogleBezosAllYourTravelAreBelongToUsPods are actually going to have problems on account of these weather conditions, not to mention there’s occasional earthquakes and other hazards in the bone-chilling cold, as in Anchorage! (And it’s DARK there six months out of the year!)

      Human stupidity knows no bounds!

  • avatar
    krhodes1

    Ultimately, I as I have said on here many times, I am of two minds when it comes to autonomous vehicles. On the one hand, humans SUUUCK at driving. It’s really a minor miracle ONLY 30K people die a year in the US in motor vehicle accidents. On the other hand, I am in IT and I know full well that computers suck too – they just suck in different ways. So will we end up simply trading one set of dilemmas for another set of dilemmas? We COULD, if we had the will, make people better drivers… I am not so sure we can make computers suck less. :-) And for sure, leaving the computers out of the picture is a WHOLE lot cheaper.

    But that said, I would be perfectly happy with some limited autonomy that can handle just boring Interstate highway driving in good weather. I feel like THAT should be initial focus, not the “do-everything” autonomous car that really seems like a fool’s errand at this point. It shouldn’t take THAT much additional infrastructure to allow my car to run up and down I-75 (perhaps in a dedicated lane) while I take a nap – I will drive to and from the highway. I do think that the halfway solutions like Tesla’s Autopilot are just stupid. They work just well enough to get you killed when they suddenly don’t.

  • avatar
    Robbie

    Too much rain: car parks itself in a safe spot, and tells owner it has to wait until the rain subsides. How big a problem can this be?

Read all comments

Back to TopLeave a Reply

You must be logged in to post a comment.

Recent Comments

  • ToddAtlasF1: Do I want a command economy? Do I think the government knows better what people need than they do? Do I...
  • ToddAtlasF1: I don’t mind when Tesla kills its customers. They’re volunteers and probably wish other...
  • CKNSLS Sierra SLT: highdesertcat- I don’t know how a Titan compares MSRP-wise. But you can get at least that...
  • Tele Vision: @ watersketch My now 76(?) year-old Mom had an RDX for a while. She got rid of it so as to get an Audi...
  • Zipster: T.Atlas With your track record of political commentary, that you would accuse the other side of fascism...

New Car Research

Get a Free Dealer Quote

Staff

  • Contributors

  • Timothy Cain, Canada
  • Matthew Guy, Canada
  • Ronnie Schreiber, United States
  • Bozi Tatarevic, United States
  • Chris Tonn, United States
  • Corey Lewis, United States
  • Mark Baruth, United States