General Motors acquired Cruise, an autonomous car start-up, last year, and has been testing its cars ever since in San Francisco and other places. Why not? “Our vehicles encounter challenging (and often absurd) situations up to 46 times more often than other places self-driving cars are tested,” according to Cruise CEO Kyle Vogt in this article at arstechnica.com.

In this “test” involving a little inner-city lane sharing on a three-lane one-way street, the car glanced off the side of the bike (or did the motorcycle graze the car?), causing the bike to wobble and fall over from a speed of 17 mph; the car reportedly was travelling 12 mph. The rider walked away with a sore shoulder and the blame for the accident, according to the SFPD.

Here’s the accident report:

That’s not even close to being the first autonomous collision in SF, but it might be the first involving a motorcycle. The arstechnica reporter included this handy link to the Department of Motor Vehicles list of accidents involving autonomous vehicles. I read all 27 of the 2017 entries, and in none of them was the autonomous vehicle at fault. Every one I read involves the autonomous vehicle getting rear-ended, sideswiped or clipped by a non-autonomous vehicle, often driven by a person described as “distracted.” In two reports, the driverless car was hit after it slowed or stopped for a scooter in one case, and a pedestrian in the other. In none of the reports did the autonomous vehicle turn left in front of, or pull out in front of an oncoming vehicle – our main cause of concern as motorcyclists.

Meanwhile in my neck of SoCal, the Orange County Register reports on yet another fatal motorcycle collision yesterday morning, which judging from this photo, looks like a clear example of a car turning left in front of an oncoming motorcycle.

MO’s deepest and sincere condolences to the family of the unnamed 68-year old Harley rider  the 24-year-old driver of this car killed. Maybe autonomous vehicles will be better for us than we realize, and the sooner the better?

  • Mark D

    As a long-time San Francisco rider, I’m not surprise, and agree the fault was on the rider. This is a perfectly typical move for a driver, and a good reason why you should always split lanes only between the left-most last and the one next to it. Let’s also remember how many meatbag-operated cars hit motorcycles!

    Glad nobody is hurt, and I remain optimistic that more autonomous cars will reduce crashes overall.

  • Peter Swinton

    So an autonomous car changed lanes and decided to change back when the car in front slowed, hitting a motorcycle that had moved into the space it was vacating. This leads to a number of questions:

    Did the car sense that the motorcycle was there? This should be in the data log.

    If not, why not.

    If it did, why did the protocols choose hitting the motorcycle as the preferred option compared to potentially hitting the braking car in front or potentially being hit from the car behind?

    Did the motorcycle honk? I have a loud dual-tone horn and have found it effective in preventing a lane change where the driver didn’t see I was there. Even if the rider did honk, can an autonomous car hear, and if so, how would it be programmed to respond?

    I don’t know the rules around filtering in California. Was the rider found at fault because he was filtering, or would he also have been at fault if he was in the right lane, then moved into the space vacated by the autonomous car?

    I don’t think this is quite as clear cut as the police report implies. And it warrants further investigation into the sensing abilities and protocols of autonomous cars. I would think that knocking down a motorcycle, bicycle or pedestrian should never be a preferred protocal compared to denting another car.

    • Juliet Bravo

      Many interesting and thoughtful questions but the bike was doing 5 mph more than traffic, or 40% faster. That alone rests blame on the bike.

      • Jeff S. Wiebe

        If an AV is not capable of adjusting to the unexpected, it is not ready for the roads. The % difference is not the measurement to look at here, but the absolute number of 5 mph. That very small variation in speed has to be not only allowed for by human drivers and law enforcement, but an AV should have a far greater capacity to deal with such anomalies, or it is not yet ready to be on public roads.

        • john burns

          and if you’re not ready for a car to move over on you, you maybe shouldn’t be lane-sharing…

          • Peter Swinton

            Yes a motorcycle should be ready for anything, especially when filtering. But should the protocols allow an A/V to move over on a motorcycle?

          • Don Silvernail

            You are right on the money, John. There are lots of people in Calif. who shouldn’t be land-sharing, splitting or whatever we’re calling it these days.When the AV’s are out there in droves these people will be revealed.

        • Don Silvernail

          Twelve mph is so slow I can’t even visualize it. If two vehicles try to occupy the same space at the same time….? Well, I’d like to hear the biker’s story too. If the authorities are going to side with the AV as a matter of course, we aren’t going to learn much about this new technology – except that human driven motorcycles will not be allowed to share the road with AVs in the future!

          • Jason

            The AV has data to back up it’s position on the road and what it did.

      • Campi the Bat

        Still well under the posted speed limit, though, and speed differentials when splitting are not legally defined for enforcement in California. Whether or not going seventeen around cars going twelve is particularly wise doesn’t matter when we’re strictly discussing legal fault.

    • Jeff S. Wiebe

      Agreed, especially re: last paragraph. I am not reassured by commenters here who are giving benefit of doubt to an AV over a human being. AVs only justifications come from their being unambiguously *better* than humans, at safety, efficiency, etc. An AV should not be considered ready to test on public roads until it is very difficult indeed to get into an accident with an AV. This isn’t a computer learning chess, but a computer learning to drive, where injuries and fatalities, not to mention loss of property, time, money, are real possibilities.

  • Paul Barwick

    “Maybe autonomous vehicles will be better for us than we realize, and the sooner the better?”

    Darn tootin’ I’ve been saying for years now that as a motorcycle rider I can hardly wait to have these dumb, drunk, distracted, and otherwise incompetent fools replaced by machines that don’t drink and drive or text or eat or read while they are on the road.

    For the record, I was a daily motorcycle rider in San Francisco for 30 plus years until I moved out a couple of years ago. Biggest thing I miss about California is the ability to legally lane split.

    • Mister X

      Don’t I know it, every time I go through SF on 19th Avenue, because if I couldn’t “filter”, it would take an hour.
      Since I live North, I just filtered across the “new” Bay Bridge last week for the first time, cool, but oddly smaller than it looked in the press photos.

      • Michelle

        Google is paying 97$ per hour,with weekly payouts.You can also avail this.
        On tuesday I got a brand new Land Rover Range Rover from having earned $11752 this last four weeks..with-out any doubt it’s the most-comfortable job I have ever done .. It Sounds unbelievable but you wont forgive yourself if you don’t check it
        !gu163d:
        ➽➽
        ➽➽;➽➽ http://GoogleOnlineMixEasyTechJobsOpportunities/easy/jobs ★✫★★✫★✫★★✫★✫★★✫★✫★★✫★✫★★✫★✫★★✫★✫★★✫★✫★★✫★✫★★✫★✫★★✫★✫:::::!gu163lhhh

      • Lavon

        Google is paying 97$ per hour,with weekly payouts.You can also avail this.
        On tuesday I got a great new Land Rover Range Rover from having earned $11752 this last four weeks..with-out any doubt it’s the most-comfortable job I have ever done .. It sounds unbelievable but you wont forgive yourself if you don’t check it
        ➽➽
        ➽➽;➽➽ http://GoogleCashGoldCareerPartTimeJobs/get/hourly✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱.✱✱✱✱✱✱✱✱

      • Evelyn

        Google is paying 97$ per hour,with weekly payouts.You can also avail this.
        On tuesday I got a brand new Land Rover Range Rover from having earned $11752 this last four weeks..with-out any doubt it’s the most-comfortable job I have ever done .. It sounds unbelievable but you wont forgive yourself if you don’t check it
        !gi120d:
        ➽➽
        ➽➽;➽➽ http://GoogleLegitimateRichJobsFromHomeJobs/get/hourly ★✫★★✫★✫★★✫★✫★★✫★✫★★✫★✫★★✫★✫★★✫★✫★★✫★✫★★✫★✫★★✫★✫★★✫★✫:::::!gi120lhhhhh

      • Christina

        Google is paying 97$ per hour,with weekly payouts.You can also avail this.
        On tuesday I got a great new Land Rover Range Rover from having earned $11752 this last four weeks..with-out any doubt it’s the most-comfortable job I have ever done .. It sounds unbelievable but you wont forgive yourself if you don’t check it
        !da147:
        ➽➽
        ➽➽;➽➽ http://GoogleNetJobsGalleryWorkFromHome/find/jobs ★✫★★✫★✫★★✫★✫★★✫★✫★★✫★✫★★✫★✫★★✫★✫★★✫★✫★★✫★✫★★✫★✫★★✫★✫:::::!da147luuuuud

  • Jeff S. Wiebe

    My first thought is that in any accident involving an autonomous vehicle (AV), the burden for safety rests on the artificial intelligence (AI) in the AV. If AI is not more capable than an ideal human in an ideal state of alertness and awareness, then it should not be in an AV on the roads. Any accident with an AV should be assumed to be the AV’s fault unless there is unambiguous evidence that the human involved was severely reckless or deliberately crashed. As with insurance fraud, some pedestrians will feign being hit, but in general, when a vehicle hits a pedestrian, we assign fault to the vehicle.

    • Wonko_T_S

      “If AI is not more capable than an ideal human in an ideal state of
      alertness and awareness, then it should not be in an AV on the roads”
      I disagree. I’ll be happy if AVs are better than the average driver. Who these days seems to only have 10% of his/her attention to spare for the task of operating his vehicle.
      As long as AI is better than the average driver then the average will be improved.

      • Douglas

        Cars (and vans, pu’s, etc) have gotten too easy to drive….too much is seemingly done for the operator so they labor under the illusion they can “multi-task” (as in texting, puttin’ on makeup, eatin’ their meal, making chin music on the phn or with others in the vehicle, fartin’ with the cd player, et al) while motoring down the road. “Oh, if I get too close to another car, I’ll be warned by the bell or buzzer…..” or somesuch.

        Vehicles have also, IMHO, been produced & sold that are capable of performance that only a few folks have the necessary discipline to operate them as a ……good, law-abiding citizen would. I know that is an anathema to those who have the overarching desire to “demonstrate their prowess” or “get their adrenaline fix” on the public roads….there are track days and offroad areas for them to “show their stuff”.

        • Preach it brother. I’m of the opinion all automated driver aids be banned until fully autonomous capabilities are available because they will only breed complacent, oblivious driving, IMO.

    • ‘When a vehicle hits a pedestrian, we assign fault to the vehicle.’

      Incorrect. ‘We only assign fault in the case that the pedestrian was in a crosswalk and had the signal. In all other cases, the entirety of the circumstances come into play, and usually favor the driver unless they were grossly negligent. For example, a person steps into the road from between two parked cars as a vehicle is driving down the street at a reasonable speed and is hit by the vehicle. Not only is that not the fault of the driver, the driver can sue and almost certainly win damages to fix their vehicle, like it or not.

  • Campi the Bat

    Worthy of note is that the Accident Details Description available to us was written by a Cruise AV representative, rather than the responding cop or any other ostensibly neutral party. It’s no more an authoritative statement of fact than the plaintiff or defendant’s opening statement in a courtroom.

    • john burns

      that link takes you right to all the Police Reports and their descriptions of each accident, they were not written by Cruise reps.

      • Campi the Bat

        The Ars Technica article ultimately leads us to the DMV’s traffic accident report, which is not a police report. The actual police report would be included with Traffic Citation #170989746, which is not provided to us and is merely summarised by the Cruise AV statement. We can even see under Section 6 – CERTIFICATION of the DMV traffic accident report that the description provided was drafted under the direct authority of Cruise’s Associate Director of AV Engineering, Kevin Chu.

      • Ferris Argyle

        As someone who favorably mentioned the NRA’s advocacy, you may wish to take heed of Gandhi’s “be the change that you wish to see”. The lack of motorcyclist.com advocacy on behalf of motorcyclists during the last California go-round on lane-splitting was in stark contrast to Gabe’s article of years past, as is this article.

        This report is even more troubling if it’s written by the police. The report finds that:
        – The self-driving car had “returned fully to the center lane” AND that it was “re-centering itself in the lane”
        – The self-driving vehicle “was traveling in the center of three one-way lanes. Identifying a space between two vehicles (a minivan in front…) in the left lane, [it] began to merge into that lane. …the minivan decelerated…the [self-driving car] stopped making its lane change and returned fully to the center lane”, ie. the self-driving car passed to the right of the slowing minivan
        – “A motorcycle that that had just lane-split between two vehicles in the center and right lanes moved into the center lane”, ie. the motorcycle passed to the left.
        – The motorcycle was 5mph faster than the car and drove into the side of the car, ie. a more vulnerable vehicle came from behind and drove into a car with full visibility; not the car drove into a small, hard-to-detect vehicle which may have been in the car’s blind spot (depending on whether its sensor placements are similar to the mirror placements).
        – “The motorcyclist was determined to be at fault for attempting to overtake and pass another vehicle on the right under conditions that did not permit that movement in safety”; ie. if a vehicle passes another vehicle on the right and a collision ensues, de facto the conditions didn’t permit that movement safely, and de jura the passing vehicle is at fault. The key here is that this only applies if you’re passing on the right, so should have been applied to the self-driving car, not the motorcycle (https://law.justia.com/codes/california/2011/veh/division-11/21750-21759/21755/)

        The police report made no mention of whether the self-driving car signaled its intent to return to the center lane, nor of whether the motorcycle signaled its intent to move into that lane.

        Takeaways:
        – Be extra-careful around self-driving cars for technical and balance of power reasons
        – Give preference to lane-splitting over a space becoming available in the lane to the right; from a legal perspective the same does not apply to a space in the lane to the left, though it may from a safety perspective.
        – Take accident fault findings to court if you can afford the time

  • michael32853hutson@yahoo.com

    how sad-i think everyone who has ever ridden a bike out on the open road know this can happen,bikers are usually the most defensive drivers because of it,every year it seems there are these stories-but i STILL don’t trust the Auto-cars! kind of like hearing airline travel is the safest form,blah blah,and you are much more likely to be killed driving on the nation’s highways, but still opting to drive rather than fly

  • Rapier51

    I always try to scan for where the driver is looking to help predict where they may go. I think that I always look at the driver, or try to, when I look at a car, I mean if there are several other things ahead, then maybe not. As opposed to looking at 4000 pounds of metal and plastic.

    With no drivers to cue off of I suppose we will learn about the typical logic these cars use and about their sensors and can get the same result as watching a driver, and you see that they see you.

    In the future everyone will be a controls engineer so we can serve the robots.

    • Mister X

      I agree, at least most humans check their mirrors before changing lanes, AV’s do not, that alone should absolve the rider.
      From the details I’ve seen reported, he had no idea the infernal contraption was going to jump back into his lane.

      • That’s where you’re going astray… It’s not ‘his lane’ until the ‘infernal contraption’ is fully in the next lane.

      • Jason

        AVs have radars checking for objects next to the vehicle and behind it at all times. So no, they don’t “check” their mirrors. They can simultaneously see in front, beside, and behind them at the same time.

  • Steve Clark

    If an autonomous vehicle has to make a choice of hitting a large or small vehicle or pedestrian it will chose the smallest one to minimize injuring a larger group of people, an example given was hit a bus or pedestrian guess who gets run over. This is a thinking machine and in all the sci-fi movies I have seen this never comes out good.

    • I don’t know who told you that, but both you and them are wrong. If the choice is nothing but ‘hit something’ then it’ll hit the thing with the most survivability. That’s not the pedestrian in your example.

  • Mister X

    Basically, the Cruise AV was proceeding too fast for conditions when it initiated it’s lane change, and instead of braking to adjust it’s speed in order to merge, it darted back into the lane it was exiting.
    That algorithm alone makes it not ready for public use.

    • Until the vehicle fully leaves the lane they are merging from, they have right-of-way in that lane and can return to it at their discretion. I overtook a minivan merging into the next lane when suddenly the driver came back and clipped me. Guess who got cited? Also, passing on the right is generally against the law in most jurisdictions, lane- splitting notwithstanding.

  • QuestionMark666

    The motorcycle and our human operated cars will be banned in 20 years according to Bob Lutz, former GM exec in an article last month in Automotive News. The reason will be that governments determine the human is too unsafe to be on the roads.