ACC sometimes activates brake lights unnecessarily

I recently discovered that (in my opinion) Clarity ACC is activating the brake lights more often than it should. Maybe some people think that it is working just fine, or that Clarity is not the only car that works this way so who cares. And some people don't like ACC for other reasons and don't use it. But for those who use ACC regularly and don't like the idea that their brake lights are coming on seemingly willy-nilly, or who have no idea when they come on and are curious when they do, here is what I have found. I will state the results first, then afterwards present the evidence that I have so far of why I have come to this conclusion.

First to be clear I don't think there is an actual defect in brake light activation in the Clarity. I am sure that Honda is following all of the regulations. The problem I think is more likely a result of less than ideal ACC software implementation. I have tried changing ACC settings, and using different driving modes, but that doesn't seem to solve it.

There are times when I have thought another driver was intoxicated because they keep pressing the brakes for seemingly no reason. In the old days that was one way to spot a drunk driver. But I now realize that nowadays most of these people are probably driving cars that have a similar shortcoming in cruise control implementation (although I suppose some of these drivers could actually be drunk).

Some of this is subjective, but in general I feel that ACC works relatively smoothly, no real complaints other than its known quirks (like slow acceleration). My only real issue is with the brake lights.

Here are two scenarios where I see the problem when driving my Clarity:

1. The car in front of me slows down a little. Even though there is still plenty of room and all that is needed is to slightly back off the throttle, instead ACC seems to panic and momentarily initiates somewhat forceful slowing. However a fraction of a second later it realizes that the situation is not as dire as it thought it was and it immediately lightens up. Since these "panic attacks" last only a small fraction of a second I hardly notice them because in the end there is very little actual change in speed. The problem is that even though the panic attack lasts only a fraction of a second, it causes the brake lights to come on. Which means that if I am following someone who is slightly uneven with their speed my brake lights are coming on quite often. From other drivers viewpoints the car in front of me does not seem to be changing speed, or at least not by that much, and there is still plenty of room in front of me, and thus they are likely puzzled why I keep stepping on the brakes.

2. I reach a slight downgrade, with no cars in front of me. In that case ACC seems to be slow to react to the downslope. Apparently it cannot sense that the car is moving slightly downwards (doesn't it have accelerometers?) and so it blunders along for a few seconds longer at the same level of throttle, which of course causes the speed to begin to increase. Then it notices then the speed is increasing and again this seems to cause a momentary panic attack, because instead of just slight backing off the throttle which is usually all that is needed, instead it quickly begins reducing speed, activating the brake lights. Then as with the other situation, it only takes a fraction of a second for it to realize that it has overreacted and it backs off. But once again you have someone behind you who is completely perplexed why you pressed the brakes. Now I realize that on a steeper downhill some heavier regen or braking is required, but I am talking about very slight downhills where you just need to back off the throttle a little. ACC seems to realize this also as that is what it ultimately does, but only after first going through a half-second of panic attack and activating the brake lights.

The downside of all of this is that drivers behind you will likely think that you are drunk or a total idiot. I supposed this could actually be viewed an an upside, since if other drivers think that you are drunk they will move away from you. I am guessing that the police know that some cars work this way and would look for other signs of impairment or distraction before pulling someone over.

Okay so what makes me think that it works this way, since generally speaking we don't know when or if ACC is activating the brake lights. Well there are a couple of ways to tell. One is that you can sometimes see a reflection from your taillights on a car behind you or other objects. But I recently discovered that believe it or not it is possible to hear when the brakes activate. It's a faint click sound which seems to be emanating from the driver footwell.

Here is a test that you can do in your garage, or any place where you can easily see the reflection from your brake lights. Without turning the car on, very lightly press the brake pedal. The brake lights will illuminate, although no sound is heard. Now press just a tiny bit more and if it's like my car you will hear a faint click. The click occurs even though there is still no significant pressure on the pedal, so I am guessing that it is a relay of some type which is separate from the brake light activation switch. Perhaps this relay activation can be read using the OBD-2 devices that are being discussed in other threads, which might be a way to determine when the friction brakes are applied (which is a separate topic). After hearing the click, press the pedal a little more and you will now hear the normal squeak type of sound that occurs when pressing any brake pedal, which I assume is the brake pedal springs compressing, although I suppose there could be other sounds contributing.

If you can't hear the click, then either I have bionic hearing, or else my car Clarity makes a louder click when pressing the brakes than others. I have a 2018 by the way.

I began to realize that this click is what I am often hearing when ACC is running. I know this because during those times when I can see the reflection of my brake lights behind, it always corresponds with that click. Then again I cannot easily test this extensively since the reflection scenario is somewhat unpredictable. In another thread @insightman has talked about rigging up a temporary foil reflector, I just tried a mockup of something in my garage and it seems like it will work, by covering a small part of the third brake light with foil I can get a reflection through the trunk window that I can see in my rear view mirror. I plan to try this on a future test drive just to confirm my prior experience with the clicks.

Another reason that I plan to do the foil test is because there is a possible situation where the brake lights will come on without making the click. In other words the brake lights may be coming on even more often then I think! That would be situations when regen is slowing the car enough that by law they have to activate the brake lights, but it is not using the friction brakes. The rule for electric vehicle braking is that the brake lights must be activated whenever speed is decreasing at a rate of more than 3 mph per second. However the same law says that the brake lights must not be activated when speed is decreasing at a rate less than 1.5 mph per second (the law is expressed in meters per second, I converted it to miles per hour). Manufacturers can set their system to activate the brake lights anywhere within this range. Although it seems like for consistency among vehicles there should be a fixed standard, but that's yet another separate topic.

Once I realized that I could audibly determine when the brakes are activated, that's when I began to realize that ACC activates the brake lights more often than I thought it was. I previously really liked ACC, now I am more hesitant to use it.

Maybe this is a limitation with ACC due to reaction time, and Honda is doing what they think is the most safe thing by starting heavy slowing during the fraction of a second that it needs to evaluate the situation, just to be safe. But that is giving a lot of benefit of the doubt. Personally I suspect that with a little more effort they could make ACC work a little smoother in these situations and still be safe, which would reduce the number of unnecessary brake light activations.

 

As I mentioned I realize that some people are not happy with ACC for various reasons, I have seen many comments and posts about that. That's a separate topic with a lot of variables, such as differences in expectation (like I said that is subjective) but I also suspect that some Clarities work differently due to calibration, driving conditions etc.

What I am bringing up is a specific topic that I have not seen mentioned before about overactive brake light activation when using ACC, even when ACC is otherwise operating relatively satisfactorily. Probably hasn't been brought up because we can't easily see when the brake lights are on, and ignorance is bliss sometimes. I am hoping that the recent breakthroughs in OBD-2 information that MrFixit and others are working on will help more people be able to determine how often the brakes are being activated in ACC. And maybe a few people will do the clunky foil test, which as goofy as it sounds could provide some interesting data for those who are interested (I realize that's not everyone).

 

If you got the impression that I don't care if ACC activates the brake lights unnecessarily, that is not correct. If the person in front of me keeps hitting their brakes, then yes I expect ACC to react each time that it happens, including with brake lights, no matter how often that is, since my car will be quickly slowing down as a result of what the person in front of me is doing. But if the person in front of me only slightly decreases their speed, and there is plenty of space between us, all ACC has to do is slightly reduce power, there is no reason for it to momentarily tap the brakes and cause the brake lights to come on. If you don't care if it does that is fine, I am not trying to convince anyone to care about this, as I said some people will be interested in this topic, others won't be.

I assume by the blue EV bar moving that you are referring to the white pointer that moves along the blue EV semi-circle, indicating the amount of power that is being delivered to the wheels. If so, then from your description it sounds like your Clarity works differently in ACC than mine does. Which is entirely possible. My car as I mentioned continues at the same power level for the first several seconds after the decline begins and only reacts after the speed begins to noticeably increase, and then it seems to be caught of guard, which gives me the impression that my car (not yours apparently) is not directly capable of detecting that the car is descending. Who knows, maybe Clarity does use accelerometers to detect vertical speed changes and mine is busted. That's why I am curious if anyone else has the same experience that I do.

 

I have been trying, maybe unsuccessfully so far, to make it clear that I am only referring to situations where a driver would not be touching the brakes, situations where all that is needed to maintain safe 3-second following distance is to slightly reduce power. Unlike situations where a driver needs to touch the brakes to maintain safe following distance, which are not the situations that I am talking about.

The same thing in downhill, I am only referring to a slight decline where a driver can easily maintain speed by slightly reducing power, no touching of the brakes is necessary to maintain speed. Unlike situations where a driver needs to touch the brakes to maintain speed, which are not the situations that I am talking about.

In the situations that I am referring to, where someone driving manually would not touch the brakes, ACC oftentimes is. The result is that in an otherwise smooth flowing group of cars that are all maintaining relatively steady speed, suddenly one car flashes its brake lights for no apparent reason.

The cause of this, as I have observed repeatedly now that I am aware of it and watching for it (or in my case listening for it), is that in many of these situations Clarity ACC seems to momentarily misjudge the situation for the first fraction of a second, resulting in the brake lights being activated unnecessarily (thus the title of this thread).

Again I understand that many don't see what's the big deal. But my concern is that brake lights should mean something. There should not be a "cry wolf" situation where brake lights come on for no apparent reason, as this diminishes their effectiveness. Brake lights are designed to get other drivers attention, to let them know that a car ahead of them is slowing down so that they can begin immediately reacting or at least preparing to slow down themselves if necessary. Flashing a brake light when there is no need to creates an unnecessary distraction to other drivers and can cause a ripple effect as other drivers react to it.

I am not trying to overdramatize this or say that it is the single biggest safety concern of the decade, I am simply pointing out that I have noticed something that is not easy to discern since we generally don't know when our brake lights are coming on. I brought it up in the hopes that maybe some others will begin to look at it as well and see if they experience something similar. I am thinking that if those using the OBD-2 devices turn on the new brake sensor, that possibly they may be just as surprised as I was to see how often ACC is tapping the brakes in situations where it is not warranted.

 

I meant to mention that I completely agree with this. In fact one of the primary reasons that I use ACC is because it helps me keep a safe distance from the car in front of me, in many cases it detects before I do that the car in front of me is slowing down. In fact sometimes it is so smooth that I will glance down at the speedometer and be surprised to realize that I am now ten mph below the speed limit because the car in front of me had gradually slowed down and I didn't realize it. So like I said earlier I think it does a good job as far as that goes. But that fraction of a second hiccup that can cause the brake lights to flash unnecessarily means that by adding this extra level of safety to my driving by using ACC I can become somewhat of a nuisance to other drivers. Of course the balance tilts towards safety so I continue to use ACC, but I look forward to when these type of systems become more refined.

 

The tough thing about all of this is that it is subjective. Which means that unless we are sitting alongside someone in their car, we really don't know what they mean when they say "too abruptly", too quickly", "too slowly", "overreacting", "erratic", etc. We think we know, because those words paint a certain picture in our mind, and we assume that is what the other person is experiencing. But it may actually be something completely different.

That being said, what you are describing, and the picture that it paints in my mind, does not sound like how my ACC works. Based primarily on the fact that you said that other drivers would notice and be annoyed by the way that your car is behaving. Even on its worst day I don't think my ACC is ever erratic enough that it would bother another driver. I notice it sometimes, but that's only because our bodies are very sensitive to motion and we feel the slightest accelerations in any direction. Whereas an outside observer will likely be completely unaware of what my car is doing unless there is a major sudden change in speed.

I will try and quantify what my car typically does. You probably can relate to what four-chevron regen feels like (three pulls on the paddle). You feel it, but it's far from what anyone would call heavy braking. Even lighter of course is one-chevron regen which is what you get when you simply take your foot off of the accelerator pedal. One-chevron regen attempts to duplicate how a regular car works when high-gear engine braking is occurring.

So when a car in front of me slows down say 2-3 mph, all that is needed is to back off the accelerator a little to match their speed, and in some cases maybe briefly use about one-chevron of regen (i.e. coasting) for a couple of seconds to more quickly reduce the speed. None of this would seem out of the ordinary to another driver who is observing. However what my ACC often does in this situation is to abruptly initiate maybe about three to four-chevron of regen, sometimes maybe five-chevron (if that existed). However it only does this for a fraction of a second and then it backs off to around neutral (power pointer at 270 degree position, no regen) and then settles in with the proper power setting to maintain the new speed. This all happens very fast, less than a second. I feel it doing this if I am paying attention, but it's easy to tune out as it is not in any way forceful. And it happens so fast that there is no way that a driver behind me would know that within that half-second interval my car had done a quick overreaction followed by a correction.

But what they would see is my brake lights flash during the moment of indecision that my ACC went through. Even if it doesn't go into heavy regen the brake light can still come on, so I suspect that in some cases it's the abrupt initiation of slowing that triggers the brake light. But either way, to another driver, seeing my brake lights come on momentarily would seem inconsistent with the situation that they are observing, as they would see no reason why I would be tapping my brakes.

If your car at times behaves erratically enough in ACC to frustrate other drivers, then that leads to another complication with all of this which is that I suspect that not all ACC systems work the same, perhaps due to poor calibration or something (I have no idea how radar would be calibrated). Or maybe some other type of hardware or software issue. I just know that if I try various ACC setting and switch to different modes, it doesn't have a major impact on how my ACC works. There may be some subtle differences, but there is no combination of settings that make my ACC behave as erratically as you are describing.

 

That is yet a whole other topic, stop-and-go driving. All of my positive experiences with ACC that I was referring to are in what I consider normal driving situations where traffic is flowing. Sadly I realize that for some people stop-and-go traffic is their normal. I never use ACC in stop-and-go traffic because it does not work well. It is somewhat marginal in slow-and-go traffic depending on the traffic flow so I use it sometimes in those situations.

I use ACC in any traffic situation where traffic is flowing relatively smoothly, even heavy traffic. Many people are wary of doing this which I can understand since many people are still not completely comfortable with radar cruise control and so they only use it in lower volume highway type of situations. But I have found that even in heavy traffic as long as things are moving ACC works pretty well and helps me maintain a safe distance with the car in front of me. No it doesn't react very well if someone ahead of me jams on their brakes, or if someone cuts in front of me, both of which are more likely in heavy traffic, so like you I take over at those moments.

I don't use ACC for stoplights either (which of course only works when someone is in front of you). ACC is not very smooth when stopping. If the person ahead of me comes to a very smooth stop it's not so bad, but most people slow down faster at signal lights than I prefer, and ACC amplifies this. I doubt if it affects the person behind me, but it annoys me enough for the both of us so when I see a light turn red or see traffic ahead of me slowing for a light I cancel ACC and manually control the coasting. And starting off from a signal light is no good either, I'm no leadfoot but it is way to sluggish so I manage the acceleration myself. Most of my surface street driving has an average of one to two miles between signal lights, and since engaging/disengaging cruise control at stoplights is easy to do, that's no big deal. However if I am in a denser urban area where there is only a few hundred yards between signal lights then I don't use ACC.

 

Thanks for the information.

 

Update on this, I was on some slight rolling hills this morning and I did see something similar to what you described. There were several points on my drive where I was going slightly uphill, then going over the "crest" for lack of a better word, then going slightly downhill. Of course while driving it's not always easy to watch the power meter but there were three different crests that I was able to safely watch the power meter as I crested the hill. Like your situation the power needle seemed to react correctly, at the crest when things started to flatten out the power needle moved down. Then when the downhill section started the needle moved further down to a tiny amount of regen, similar to coasting. Through all of this the speedometer was steady at 55 mph, and holding as the decline continued. So it all worked fine - EXCEPT that about five seconds into the decline, with speed steady at 55 mph and no one in front of me, all of a sudden and for no reason the regen increased momentarily, which I could feel, but like previous situations it only lasted for a moment. But during that fraction of a second I heard the telltale click which tells me that the brake lights flashed. This happened on two different crests, in both cases I was staring right at the needle when it happened. A third time that I was able to watch, it did not do the overactive regen thing on the decline and the brake lights did not come on as I can tell.

Really seems like a software issue. I wish I had a video camera running, if so you could see that there was no one in front me, steady speed, no reason to suddenly go into extra regen for a fraction of a second. Again I have a 2018, I wonder if they improved it on later models. Then again at least for the downhill issue it's solved on 2020 and 2021, since those models apparently no longer do regen at all on downhills while in ACC, if I understand correctly. I wonder how those models know that the car is going downhill, accelerometers?

 

I have never seen the paddles activate the brake lights. However @Ray B says that he has so apparently there are some scenarios where it will.

In theory the paddles and ACC are doing the same thing, using regen to slow down. The paddles however are limited in the amount of regen that they will do. ACC will do much heavier regen than the maximum four chevrons of the paddles. Also most people seem to believe that the paddles don't use the friction brakes, but I'm not sure if anyone knows if ACC ever mixes in friction brakes with regen like the brake pedals do. I think it might because my current theory is that the click that I am hearing is a solenoid that activates the electric brake booster or ABS or some other piece of hardware in anticipation of friction braking. I am hoping that the OBD-2 sensor can help confirm if ACC sometimes uses the friction brakes.

I am currently operating under the assumption that the brake lights will come on in three scenarios:

1. The brake pedal is pressed by the driver, no matter how lightly.

2. ACC activates the friction brakes (if it uses friction brakes). I say this only because I suspect that it might be viewed by lawyer types that the original brake light law applies in any situation where the friction brakes are applied, whether by the driver or automatically.

3. ACC deceleration exceeds whatever threshold Honda has set. We know only that by law regen should not activate the brake lights if there is less than 1.5 mph per second of deceleration, and regen must activate the brake lights if there is more than 3 mph per second of deceleration.

Other than having friends following you, do you have any other way to know if ACC never activates the brake lights on your car?