Lane Splitting Motorcyclists And AI Autonomous Cars

California is the only state to make lane-splitting by motorcycles a legal practice. (GETTY IMAGES)

By Lance Eliot, the AI Trends Insider

(Ed. Note: For reader’s interested in Dr. Eliot’s ongoing business analyses about the advent of self-driving cars, see his online Forbes column:

My recent column about the latest aspects of motorcycles and how AI is making them semi-autonomous and fully autonomous was quite well-received and garnered a lot of interest.

Here’s a link to that article on autonomous related motorcycles:

I thought it might be handy to consider another aspect about motorcycles, namely the somewhat controversial practice of motorcyclists that practice doing lane splitting.

Have you ever heard of lane splitting?

How about lane sharing, or lane white-lining, or even sometimes referred to as strip-riding?

Here in California we are quite familiar with this terminology since it refers to something we see every day, namely, motorcyclists that go between the lanes on our highways, freeways, and byways.

This is one of those aspects about California that one might be conflicted about.

We are the only state that specifically has made it a legal practice to do lane splitting.

Some states outlaw it outright, while most states are silent on the matter and tend to either allow it implicitly or kind of look the other way about it.

Is California right to have legalized the practice?

Are we ahead of everyone else?

Or, are we doing something unwise and unwarranted?

Controversy Surrounding Lane Splitting Practices

No matter where you are, I’d bet that there are rather divided opinions about the practice.

The notion is that motorcyclists do not necessarily need to make the same kinds of lane changes that cars need to do.

A motorcyclist is allowed under “lane splitting” to go between two cars and squeeze along forward.

Imagine you are driving your car on the freeway, doing so in the rightmost lane, and another car is to your immediate left in the fast lane that is adjacent to the rightmost lane. You and the other car are next to each other. You are both going say 50 miles per hour. You are both for the moment going the same speed and just a few feet from each other, really almost just inches at times. There is no chance of another car squeezing between you and the other car because you are both in unison and for the moment shoulder-to-shoulder of each other.

In California, a motorcyclist can try to squeeze between you and that other car.

Imagine that the motorcyclist is doing 55 miles per hour and comes up from behind you and the other car. The motorcyclist is blocked seemingly because you and the other car are occupying the two lanes. Other cars behind you would need to wait until somehow an opening develops, such as if you speed-up or the other car does, and the two of you are no longer going neck-and-neck. Or, maybe you exit off the freeway and open up the rightmost lane for other traffic to proceed.

In any case, in spite of the apparent momentary blockade of you and the other car, if a motorcyclist believes they can fit between you two, they are legally allowed to do so.

In fact, on most mornings as I drive on the congested bumper-to-bumper freeway to work, motorcyclists are streaming along and shimming between the cars. I can usually see them coming from behind me, snaking their way in and around cars, doing so while going perhaps 35 to 40 miles per hour, while the rest of the traffic is staggering along at maybe 15-20 miles per hour.

Presumably, the motorcyclists are polluting less than car drivers and so they are being rewarded by being able to snake their way through the traffic. I know several motorcyclists that laugh when I tell them that my daily morning commute takes an hour or more. For them, by using lane splitting, they can do the same distance in half that time. I’d say they relish the aspect that the rest of the traffic is either sitting still or moving at a turtle’s pace. They meanwhile are moving as fast as they can, albeit inhibited by the snarled car traffic.

You might be tempted to say that this practice seems reasonable.

Why not let the motorcyclists be able to do lane splitting if it makes their commute more efficient?

If you don’t allow lane splitting, you would be forcing those motorcyclists to act like cars and be forced to wait behind the cars that are stacked up on the road. It would certainly be frustrating to the motorcyclist. Furthermore, from a traffic perspective, and especially here in crowded traffic-mania Southern California, having the motorcyclists act like cars would make our traffic even worse. For each motorcycle that might have zipped along and snaked through traffic, they would instead be taking up the same space as a car and cause our traffic lines to get even longer and generate more traffic congestion. Some also suggest that if motorcycles were relegated to staying behind cars while in-traffic, there would be more injuries or deaths of motorcyclists by cars that rear-end into the motorcyclists (another point of some debate).

I’d like to share with you a story about the sometimes-surprising nature of lane splitting.

A colleague recently came out here from the east coast and he had never seen lane splitting in action. He had heard of it and had seen pictures and videos about it, but not had the “pleasure” of experiencing it directly. I decide to have some fun about this, and so I asked him to drive us to the office on his first day here. I wanted him to get a driver’s view of the matter, rather than first being just a passenger in a car and experiencing the lane splitting from that seat.

For those of you that have not yet been driving when lane splitting occurs, it can be jarring when it first happens. My colleague was focused on the morning traffic mess and had forgotten my forewarning about lane splitting. He was chatting with me and pointing out the crazy drivers up ahead of us, and then suddenly, seemingly out-of-the-blue, a motorcycle went past us, doing so within inches of the driver side window. It happened so fast that my colleague was startled and not even sure what had just happened.

All he saw was the flash out of the corner of his eye.

He then turned his head in the direction of the motorcyclist, whom by now was already several cars ahead of us, having zipped past us and moving at a clip far above the speed of traffic. It was amazing to see the reaction of my east coast colleague. His mouth gaped open. He stammered that if he had opted to turn the wheel to the left and moved toward the left lane, he would have readily cut-off the motorcyclist and a dangerous incident might have occurred.

Welcome to Los Angeles, I said!

I mentioned that regrettably there are frequent such incidents of motorcyclist and cars that bop into each other.

I would guess that at least once per week I witness one of these incidents or the aftermath of these lane splitting incidents. Personally, I think that’s a frighteningly high frequency.

It has become commonplace for me to see a motorcyclist take a spill onto the freeway. Most of the time, thankfully, the motorcyclist gets back onto the motorcycle and continues on their way, apparently unhurt and undamaged. I see this repeatedly. On a few occasions, I’ve seen much worse, sadly. And, by listening to the traffic reports on the radio, every morning commute is filled with indications of motorcyclists downed here or there on our extensive freeway system.

In one sense, many seasoned drivers here take it for granted that there are lane splitting incidents. It is no different than expecting to see debris spilled onto the freeway by trucks that are overloaded or that have failed to cover-up their carrying loads. Each day, I see various debris such as fruits dropped onto the freeway, torn-up tires, old furniture, etc. I don’t think it surprising. If I didn’t see either flying or fallen debris on my daily commute, I’d be surprised. Likewise, if I didn’t see a lane splitting incident over the course of a week, I’d wonder what happened that week (I was either asleep at the wheel or the motorcyclists decided to take the week off).

Thus, this takes us to the core of the controversy about the lane splitting approach.

Some would say it is an overly dangerous practice and should be banned. Others say that it is up to the motorcyclists to decide what they want to do.

If a motorcyclist is willing to take the risk, they should have the freedom to choose whether to do lane splitting or not. Just because it is legal here to do so does not mean that the motorcyclists all have to do lane splitting. They can use their judgement as to when it is safe to undertake it.

But a counter-argument is that it is not just the motorcyclist that is involved in the lane splitting matter.

By and large, a lane splitting incident is going to involve a car. Therefore, car drivers are just as involved. The car driver has no ability to decide when a motorcyclist should be allowed to do lane splitting. Since lane splitting is legal here, the driver must just live with whenever and however a motorcyclist decides to do lane splitting. Even if a motorcyclist does something really stupid and tries a lane split that is obviously ill-timed and going to likely produce a crash, there is no means for a car driver to particularly prevent it from happening.

Sure, you might argue that if the motorcyclist was in the wrong, presumably the driver of the car will be found to be not at fault and it will instead land on the head of the motorcyclist. Though this might be true, you need to factor into this the hassle part of the equation. The lane splitting has led to a car incident that otherwise would not have presumably occurred (at least not legally; i.e., of course a motorcyclist in a state that bans lane splitting could nonetheless do lane splitting, but at least they would already be in the wrong in doing so). And, the driver of the car will need to make their case that it was the motorcyclist that led to an incident.

Motorcyclists here will tell you that much of the time it is the “foolish” car drivers that are at fault.

A motorcyclist trying to squeeze between two cars will often inadvertently get bashed by one of the two cars. A car driver might have swerved to the edge of their lane, doing so presumably without awareness of the presence of the motorcyclist. I’ve seen many side-view mirrors strike the motorcyclist, either damaging the mirror or tearing it apart from the car, and meanwhile it is usually enough of a blow that the motorcyclist loses control of their motorcycle and falls to the ground (by falling, I mean it is more akin to skidding along on the ground, since they are in motion at the time of the encounter).

The even worse encounters that I see involve a situation of two cars that are not actually abreast of each other, with one slightly ahead of the other, and for which one car suddenly decides to change lanes, and meanwhile the motorcyclist had mentally calculated that it was possible to squeeze between the two cars. The motorcyclist then rams into the rear of the car that opted to make a sudden lane change. This is worse than getting a glancing blow of a side-view mirror, and usually the motorcyclist goes down hard, often first having their body strike the back of the car.

Confounding Aspects Of Lane Splitting

There’s another factor to these incidents that you need to consider.

When a motorcyclist has an incident during lane splitting, it often inadvertently entangles other cars and car drivers into the incident.

Suppose you are driving along on the freeway at 35-40 miles per hour, and suddenly a lane splitting incident happens just ahead of you. Let’s assume you weren’t directly involved. But, you now have the chance of possibly running over the downed motorcyclist. Or, maybe you might hit their downed motorcycle. Or, maybe their motorcycle continues for a short distance, on its own, perhaps skidding, and strikes your car.  Or, the car that was involved opts to slam on their brakes and has unexpectedly halted just in front of you. Etc.

You can be an innocent bystander that gets enmeshed in the whole mess that ensues.

Guidelines About Lane Splitting

Last year, the California Highway Patrol (CHP) announced new guidelines about lane splitting in California (note that due to a legal controversy arising from the initial posting of the guidelines, the CHP subsequently took them down and a further review process is underway).

Let’s consider the initial draft guidelines that were posted.

First, the CHP recommended that only experienced motorcyclists try to do lane splitting.

This is certainly sage advice.

In reality, I’d suggest that most motorcyclists here on our freeways do lane splitting, regardless of their experience at riding a motorcycle. And, I’d also suggest that many of the newbie motorcyclists intentionally get a motorcycle partially because they are exasperated about waiting in traffic while driving a car. As such, they are determined to do lane splitting. It looks easy enough to do and likely these motorcyclists believe they can wiggle their way out of any trouble.

The CHP recommended that motorcyclists only do lane splitting when traveling at no more than 10 miles per hour faster on the motorcycle than the prevailing car traffic around them. Furthermore, the CHP recommended that lane splitting be only undertaken when the prevailing traffic is going below 30 miles per hour. More sage advice.

In my morning commute, I’d wager that the motorcyclists are often doing 20-30 miles per hour faster than the surrounding traffic when they lane split (rarely limiting themselves to just a 10-mph differential in speed).  Also, the lane splitting seems to happen at all speeds, including when the prevailing traffic is going above the speed limit, such as 70-75 miles per hour.

I’ve seen lane splitting take place at speeds that make me shudder and for which if something goes awry it would definitely lead to death for the rider. At such high speeds, there is little or no room for error. The slightest twitch can spell the difference between a motorcyclist upright and one that is careening onto the asphalt and likely going to get hit by one or more cars along the way. We have a motorcycle helmet law in California, but I tend to doubt that a motorcyclist flying off their motorcycle at 80-mph and into fast moving traffic is going to have much luck even while wearing their helmet.

The CHP recommended that lane splitting only occur between lanes #1 and #2.

This would be considered the fast lane and the lane to the right of the fast lane. We often have three, sometimes four, and even at times five lanes on our freeways in each direction. It certainly makes sense to suggest that the lane splitting happen only on the leftmost lanes. This keeps things a bit safer for the traffic in the slow lane and for when there are on-ramps and exits off the freeway. It would also provide a kind of consistency so that car drivers would know where to be on the watch for lane splitting.

Though this is again quite sage advice, I’d say that lane splitting seems to happen on any lane at any time.

Whatever the traffic situation dictates and whenever a motorcyclist wants to get ahead of the traffic, there are those motorcyclists that do so. The aspect that especially seems dangerous and borderline legal is when they use the HOV lane for lane splitting. Here our HOV lanes are usually bounded by a double yellow that is not to be crossed at all, and only when there is a designated break in the HOV lane are you allowed as a driver to enter into or out of the HOV lane. Some areas of the country allow entry and exit of an HOV at any time. We generally do not.

The lane splitting motorcyclists will often slide back-and-forth into and out of the HOV lane. This often seems to work out well for those daring motorcyclists in terms of wanting to get ahead in the traffic. You might say that it maybe makes their efforts somewhat “safer” since the cars in the HOV are not supposed to be crossing the double yellow and likewise cars wanting into the HOV are not supposed to be crossing the double yellow. All I can say is that by my observations the lane splitting using the HOV seems to catch drivers especially off-guard and appears to be as dangerous, or more so than with lane splitting on the other lanes. Just an observation.

I’ll cover just a few more of the CHP recommendations and not get to all of them.

One recommendation stated that a motorcyclist should not lane split near large vehicles such as buses and trucks.

I agree wholeheartedly.

Doing lane splitting in those situations is dicey since the driver of the oversized vehicle often cannot see the motorcyclist and also since other traffic can lose site of the motorcyclist due to the large vehicle’s physical size too.  As you might guess, I’ve observed lane splitting even when there are large vehicles nearby.

The CHP recommended that motorcyclists doing lane splitting should be wearing brightly colored clothing and gear.

I’d say that most of the motorcyclists that I see are usually wearing traditional oriented motorcyclist clothing consisting of black or brown leather jackets, and rarely do they have any kind of especially high-visible clothing or gear on them. Until or if the motorcyclist culture somehow changes toward brightly colored attire, I’d say that this CHP recommendation is unlikely to be widely adopted.

There are even recommendations for car drivers.

The CHP indicated that car drivers should not try to impede a lane splitting activity.

Allow me to explain.

There are some car drivers that don’t like the lane splitting.

As such, they will at times intentionally position their car to block a motorcyclist that is trying to perform a lane splitting action. This has included getting so close to another car that they are almost willing to have their car scrape against another car. Usually, these car drivers that are anti-lane splitting will be watching for an upcoming motorcyclist and then shift in the lane to the side of the motorcyclist, trying to sneakily narrow any space and thus discourage the lane splitting action.

This can be a quite dangerous cat-and-mouse game.

The motorcyclists that are seasoned at lane splitting know that some cars are trying to intentionally impede the lane splitting. As soon as a motorcyclist detects this possibility, they will try to outwit the car driver by faking to one side and going to the other side. I’ve witnessed a motorcyclist that nearly got jammed up by a car driver that seemed to be intent on preventing the lane splitting, and after the motorcyclist managed to burst just past the car, the motorcyclist then slammed his fist down on the hood of the car and sped away. It gets that crazy on the roads here.

The final piece of CHP sage advice that I’ll mention herein is the aspect that per the CHP recommended practices that car drivers are to try and aid or enable lane splitting by shifting in their respective lane at the time of the lane splitting action.

This to me seems the most questionable suggestion of the various points made.

It is one thing to tell car drivers to not impede lane splitting, it’s another idea altogether to have car drivers try to make it easier to do.

Why do I seemingly object to this notion of car drivers helping lane splitting actions?

Here’s why.

Though on the surface of things it seems like a good idea to have cooperative drivers, and in fact many drivers do shift in their lanes to help provide more room for a lane splitting action, it can also get out-of-hand.

I’ve seen some car drivers that in their desire to be generous to the motorcyclist, shifted so far in their lane that it scared other nearby cars. Those other car drivers did not realize what the intent of the shifting car driver was, and instead thought that the car was weaving or perhaps going to make a sudden move into the other lane (maybe the driver is drunk, maybe the driver has lost control of their car, etc.).

Because other drivers don’t actually necessarily know that you are trying to help a lane split, it can create other adverse consequences.

Those other cars can begin to move or swerve due to wanting to get away from your actions. This can then create a cascading series of such moves. It can be disruptive to traffic overall. It can make for tension and dangerous situations. I realize that the drivers doing this are trying to be good citizens, but unfortunately it has repercussions that I think even they at times are oblivious to.

You might also find of interest that motorcyclists sometimes thank car drivers that shift over in the lane to accommodate a lane splitting action.

The most common form of thanking involves the motorcyclist waving their hand in a friendly gesture to the car driver, doing so as they get past the car. One small aside about the gesture. Since sometimes the motorcyclists use one finger to make a pissed-off gesture to cars that cut them off or don’t help the lane splitting, I’ve seen some helpful drivers that get utterly confused to get a hand wave and mistakenly think the motorcyclist is irked at them, when in fact the motorcyclist is trying to show appreciation.

What a world we live in!

AI Autonomous Cars And Lane Splitting Handling

What does this have to do with AI self-driving driverless autonomous cars?

At the Cybernetic AI Self-Driving Car Institute, we are developing AI software for self-driving cars. In our view, this also includes the ability of self-driving cars and the AI to be able to contend with lane splitting.

For my article about AI self-driving cars contending with motorcycles overall, see:

For many of the automakers and tech firms that are making AI self-driving cars, the notion of dealing with lane splitting is quite low on their priority list. Indeed, they would tend to say it is an “edge” problem. An edge problem is one that is not at the core of the overall problem that you are trying to solve. You assume that an edge problem can be dealt with at a later time, after having first solved the core. For automakers and tech firms, the core involves getting an AI self-driving car to work on our roadways in a rather bland and typical manner, after which they will deal with exceptions and so-called edge aspects.

For my indication about edge problems in AI self-driving cars, see:

For my overall framework about AI self-driving cars, see my article:

When I’ve had discussions with them about this matter, they tend to point out that only California has a law that legally allows lane splitting.

Why worry about a law that is only pertinent to one state?

They are trying to make AI self-driving cars that work for anywhere in the United States and so it is “obviously” a rather limited concern when it is only lawful in California.

I refute this idea that it is only limited to California. As mentioned earlier, many states allow it by not explicitly banning it. Plus, I assert that motorcyclists all across the country at times will do lane splitting, even in places where it is banned (I’d bet that most motorcyclists think it is a pretty low chance they would get nabbed for doing an illegal lane split, unless they did so brazenly and stupidly in front of a police car).

Lane splitting is actually quite popular in parts of Europe and in many Asian countries.

If you are making an AI self-driving car, I’d suggest you ought to be considering how the self-driving car and the AI will cope in countries besides just the United States.

For my article about internationalizing AI self-driving cars, see:

So, I tend to reject the idea that dealing with lane splitting is a rather narrow topic of only concern to California driving.

To further pursue this notion by some that lane splitting is a rarity and to be neglected for now, I’d like to first introduce the notion that there are varying levels of AI self-driving cars.

The topmost level is considered Level 5. A Level 5 self-driving car is one that is being driven by the AI and there is no human driver involved. For the design of Level 5 self-driving cars, the automakers are even removing the gas pedal, the brake pedal, and steering wheel, since those are contraptions used by human drivers. The Level 5 self-driving car is not being driven by a human and nor is there an expectation that a human driver will be present in the self-driving car. It’s all on the shoulders of the AI to drive the car.

For self-driving cars less than a Level 5, there must be a human driver present in the car. The human driver is currently considered the responsible party for the acts of the car. The AI and the human driver are co-sharing the driving task. In spite of this co-sharing, the human is supposed to remain fully immersed into the driving task and be ready at all times to perform the driving task. I’ve repeatedly warned about the dangers of this co-sharing arrangement and predicted it will produce many untoward results.

For the levels of self-driving cars, see my article:

For why AI Level 5 self-driving cars are like a moonshot, see my article:

For the dangers of co-sharing the driving task, see my article:

Let’s focus herein on the true Level 5 self-driving car. Much of the comments apply to the less than Level 5 self-driving cars too, but the fully autonomous AI self-driving car will receive the most attention in this discussion.

Here’s the usual steps involved in the AI driving task: 

  • Sensor data collection and interpretation 
  • Sensor fusion
  • Virtual world model updating 
  • AI action planning
  • Car controls command issuance

There are some that suggest that even if lane splitting is worthy of consideration, you can delay worrying about it by instead just letting lane splitting happen for the time being.

In other words, if a motorcyclist wants to lane split, let them go for it. The AI of the self-driving car presumably could care less that the motorcyclist is doing the lane splitting. No need to contend with the matter. Just let it happen. It’s all on the shoulders of the motorcyclist.

If your view of a self-driving car is that it is sufficient for it to drive like a novice driver, I suppose there is some merit to this point about ignoring the lane splitting. I’ve seen novice drivers that are so overwhelmed with the driving task that the last thing they notice or care about involves motorcyclists that are lane splitting. The novice tends to assume that if the motorcyclist is doing lane splitting, the motorcyclist knows what they are doing. It would be as though a bumble bee has flown around your car. Let it do so.

I cringe at this belief.

Are we really expecting that true Level 5 self-driving cars are to be driving on our roadways in the same manner as a novice driver?

I hope not. If so, we’ll all be in a lot of trouble.

Autonomous Cars Dealing With Lane Splitting

As I’ve mentioned many times, a seasoned human driver knows how to drive a car in both proactive and defensive ways. They are on the watch for patterns of driving situations that alert them to take in-advance action. They manage to avoid accidents that might otherwise have occurred. I’m not saying that humans are flawless.

I am just saying that we are more than just lucky that as a society we do not have more car accidents than we already have. I’d assert that human driving skills are an amazing aspect that generally keeps us relatively safe on our roads.

It is a marvel to every day do my daily commute and I don’t encounter accident upon accident and upon accident.

For the human foibles of driving, see:

For the defensive practices for AI self-driving cars, see:

For insights about car accidents, see my article:

For the egocentric designs of AI self-driving cars, see my article:

In short, I reject the idea that lane splitting can be “ignored” and also that even if being considered that it should somehow be placed at the back of the bus, as it were.

Here’s a few reasons why being aware of lane splitting is an essential car driving skill for AI autonomous driving systems.

First, be aware that many of the AI self-driving cars are initially going to be very rudimentary in their driving practices. They will tend to be skittish drivers. We’ve already seen that some AI self-driving car systems will only abide strictly by the speed limits and are quite civil in their behavior towards other cars. This has led to situations wherein an AI self-driving car kept waiting for other cars to go first, or frustrated other cars into making rash moves that then potentially led to an incident or potential incident.

I realize that some pundits of AI self-driving cars will say that there is nothing wrong with the AI being skittish and that the real problem is those pesky human drivers. Outlaw human drivers. Allow only AI self-driving cars. Problem solved. Those pundits conjure a Utopian world in which there are only AI self-driving cars on public roads.

Let’s talk about reality.

Currently there are about 250+ million conventional cars in the United States alone, and those cars are not going to magically disappear or become true Level 5 AI self-driving cars overnight.  Indeed, the use of human driven cars will last for many years, likely many decades, and the advent of AI self-driving cars will occur while there are still human driven cars on the roads.

This is a crucial point since this means that the AI of self-driving cars needs to be able to contend with not just other AI self-driving cars, but also contend with human driven cars. It is easy to envision a simplistic and rather unrealistic world in which all AI self-driving cars are politely interacting with each other and being civil about roadway interactions. That’s not what is going to be happening for the foreseeable future. AI self-driving cars and human driven cars will need to be able to cope with each other. Period.

We can also add to the reality list the aspect of motorcyclists. For those pundits that want to eliminate human driven cars, they would undoubtedly be aghast at the idea of still allowing human driven motorcycles. If the human drivers are gone, so would the human driven motorcyclists.

If we did indeed wave a magic wand and had only AI self-driving cars, and if there weren’t any motorcycles at all, the lane splitting topic pretty much becomes irrelevant.

If we allowed for AI driven motorcycles, which is an area of ongoing research, you’d likely need to contend with lane splitting. But, in that case, you’d presumably be able to have cooperative behavior between the AI’s of the self-driving cars and the self-driving motorcycles.

They would electronically communicate via V2V (vehicle to vehicle communication), and agreeably allow for lane splitting (in theory).

For my article on the advent of AI systems for motorcycle driving, see:

For my article about the grand convergence that has led us to this moment in time, see:

See my article about the ethical dilemmas facing AI self-driving cars:

For potential regulations about AI self-driving cars, see my article:

For my predictions about AI self-driving cars for the 2020s, 2030s, and 2040s, see my article:

I’d say that realistically there will be lane splitting and it will occur during the advent of the emergence of AI self-driving cars.

Furthermore, it will occur during the era of AI self-driving cars that are skittish.

Machine Learning and Lane Splitting Aspects

How does skittishness play a role?

Remember that I mentioned the story of my colleague that was surprised when a lane splitting action occurred? I had told him beforehand about lane splitting, and yet when it happened, he was surprised. Luckily, he did not do anything rash.

Suppose a skittish AI self-driving car suddenly has a lane splitting motorcycle that darts within inches of the self-driving car. What will the self-driving car do?

The AI might detect the motorcycle and assume that the motorcycle is on a path to hit the self-driving car. Perhaps the AI directs the self-driving car to make a rapid lane change to avoid the motorcycle. Or, takes some other action that is not quite expected or anticipated by other drivers and nor the lane splitting motorcyclist. This could spell trouble for the AI self-driving car, and the nearby traffic, and for the motorcyclist.

From a Machine Learning (ML) perspective, suppose the AI encounters this lane splitting a multitude of times, but has no context for why it is occurring.

What should the ML learn from it?

If each time it happens the ML opts to take a sudden evasive maneuver, it could be that the ML gradually accepts that this is the right way to deal with the matter. What the ML has “learned” is not necessarily the proper kind of driving action to take.

For more about ML and AI self-driving cars, see my article:

For AI self-driving car driving styles, see my article:

On the topic of tit-for-tat and AI self-driving cars, see my article:

Overall, I’m suggesting that if AI self-driving cars are skittish or novice style drivers and cannot explicitly deal with lane splitting, they might get themselves into trouble, including possibly making rash driving moves that could lead to harm for any human occupants in the self-driving car, along with harm to other humans in nearby cars and for the motorcyclist too.

Furthermore, for those that suggest that AI self-driving cars will just learn how to deal with lane splitting, doing so after some number of driving hours and experiencing lane splitting, I’m not convinced that what the AI learns will really be the most prudent driving practices related to lane splitting.

If we also embrace the recommendation of the CHP that car drivers should aid the lane splitting action, it seems especially unlikely that the AI is going to figure that out on its own. More likely is that the AI would figure out how to defend against it, rather than to try and enable it.

For my article about Darwinian learning related methods, see:

In general, I’d vote that the AI be explicitly programmed or trained in dealing with lane splitting.

This also must include the realities of how motorcyclists really do lane splitting.

If you were to try and setup the AI to believe that lane splitting will only happen on lanes #1 and #2, and only when the motorcycle is just 10 miles per hour faster than the prevailing traffic, and so on, you’d be creating a rather limited and potentially confused AI system when it had to deal with the real-world lane splitting activities.

The AI needs to also be able to contend with the reactions of human drivers. As already mentioned, some human drivers will try to aid the lane splitting, while others will try to disrupt or prevent it. The AI self-driving car needs to be aware of those potential actions and be ready to deal with those other car drivers.

Let’s pile on about the circumstances of lane splitting.

It could happen during daylight when visibility is perfectly clear. Or, it can happen at night when it is dark and hard to see the motorcyclist. It can happen during dry weather when the roads are readily driven on, or it can happen in the pouring rain and the roads are slick and slippery.

There could be just one motorcyclist trying to do a lane split, or there could be a multitude of motorcyclists doing so, all at once (I’ve seen this happen many times during my daily commute, namely motorcyclists riding together as a pack or team).


Lane splitting – is it a boon to our driving world, or is it a contemptible practice that should be curtailed?

I’m sure the debate will be going on for a long time about the merits of lane splitting.

Meanwhile, it exists, and it happens. AI self-driving cars need to be ready for it. None of us want AI self-driving cars that by intent or by happenstance ram into a motorcyclist doing lane splitting or get involved in the aftermath of a lane split that has occurred to some other driver.

Let’s make sure that the AI is savvy about lane splitting.

I’m not willing to split hairs on that.

Copyright 2019 Dr. Lance Eliot

This content is originally posted on AI Trends.