A flaw in Inmotion safety alarms for V10/V10F and V11, demonstrated

[supercurio]

Following the in-depth discussion in this forum about @EUChristian's V11 cutout resulting in a crash at 53 km/h in this thread, here is a video I made which demonstrates an unfortunate design characteristic of the alarm system which contributed significantly to the poor outcome.

More details

As of December 2020, @Inmotion Global V10, V10F and V11 wheels exhibit a problematic behaviour affecting the way safety alarms are communicated to the user.
After being triggered once, an alert stops and can't be communicated to the user anymore during 5 seconds later in most cases, exposing a rider to significant risk of crash.
Even if the conditions triggering an alarm continue, the alarm cannot ring continuously.
Some conditions can reset the 5s delay timer but they remain unclear to me.

The video also demonstrates that despite the wheel is reaching its limits in power delivery due to my intense acceleration requests every second, tilt-back is not getting engaged in order to alert the user while the alarms are delayed.
Instead, the pedals are tilting forward as a acceleration assistance.

Here are two example scenarios highlighting how this alarm design is a safety hazard on a V10 or V10F.

Bump on the road

• The rider is cruising at 32 km/h, significantly below max speed
• A bump on the pavement triggers the alarm
• 2 second later, the rider decides to accelerate in order to keep a safe distance with a car behind
• During the 3 seconds following, the rider increases speed without being alerted
• The wheel was pushed above its top speed leading to a crash in front of the car following

Incline

• The rider is cruising at 25 km/h and decides to increase speed to 35 km/h
• As the rider is accelerating close to the max speed, the wheel rings an alert and briefly engage tilt-back
• The rider stops accelerating and remain at a steady speed, below max speed, which ends tilt-back deployment
• Within the next 3 seconds, the road transitions from flat to a steep incline.
• At alarm+3s, the wheel reaches its  power delivery limit to keep speed at 35km/h but can't alert the rider
• At alarm+4s, with battery sag increasing, the wheel doesn't have the power reserve to deploy tilt-back and still can't alert the user, leading to a crash near max speed.

I designed the demo protocol to avoid stressing continuously the wheel's power delivery in order to not crash for science, but these are real world potential results of this problematic alarm and tilt-back design.

 

And this is the original discussion which prompted this analysis.

 

Edited December 11, 2020 by supercurio

[Rawnei]

Great video, very well articulated! I hope Inmotion watches this and does internal analysis of their safety mechanisms. They also need to take a look at the tilt-back on the V11 at high speeds which happens very slowly.

[Freestyler]

Thanks for demonstrating this.

I have the V11 and I have also noticed that the tiltback in lower speeds is very aggressive, but in top speed (50km/h) is barely noticeable.

Sometimes I'm not sure if it was an actual tiltback or it was just in my head.

 

Having said that, that's why I choose a wheel with up-gradable firmware.

Both of these issues are easily fixable by software and I have faith that InMotion will fix them eventually.

[supercurio]

Good point @Rawnei and @Freestyler, since the alarms can be delayed significantly, tilt back becomes the only method the wheel can communicate its status with the rider.

Although the V10 and V11 have slightly different tiltback characteristics, they are both lacking in consistency and effectiveness. 

Including in this video, where tiltback gets disabled every time because the acceleration is not continuous and steady. The pedals tilt forward instead, while the alarm can't ring.

Delays and transitions between states need complete rework as various conditions end up disabling both safety alert systems. 

[Kekafuch]

What you explain is clearly demonstrated in your video.  Thanks for making that and sharing with the community.  

My experience with the V10F is having a "Please get off" warning where the wheel goes into a permanent tilt back that forces me off the wheel.  Then I have to shut the wheel off before it self balances again.  I am attributing it to colder weather but I am not sure what it means.  It seems to happen as I reach a hill and try to accelerate up it. 

[supercurio]

Thanks @Kekafuch, when making recording that video, with very little experience talking to the camera by myself I wasn't sure the message would come across. 

Same as you during this winter! I got about 4-5 "overload, please get off" with final tiltback when riding at usual, but near freezing temperature, mostly in uphill sections in conditions where battery sag accumulated rapidly. 

At my weight (65kg) it never happened before then. One time a car was right behind me which was awkward.

It's a good safety to have instead of an overlean, but in some cases I would prefer beeps to know to reduce my speed and acceleration in order to avoid triggering it since you have to handle a marked tiltback and get off. 

[Wraith Rider]

As I agree with this, from my experience INMOTION is still far safer than any other EUC company.  Inexperience not understanding the limits of the wheel is user error.  Of course INMOTION can push farther development to make the wheel safer but any wheel can be made to cut out.  Accelerating to keep up with traffic near max speed is not an excuse.  

[supercurio]

3 hours ago, Wraith Rider said:

As I agree with this, from my experience INMOTION is still far safer than any other EUC company.  Inexperience not understanding the limits of the wheel is user error.  Of course INMOTION can push farther development to make the wheel safer but any wheel can be made to cut out.  Accelerating to keep up with traffic near max speed is not an excuse.  

I agree the V10F appears to be pretty safe as baseline - ignoring this alarm issue.
With conservative speed limits at all battery voltages for instance, I found the pedals to be really solid and never dipping, which my 16X does before beeping in some scenarios.
You can see it in this video, I stomp on the pedals pretty hard, generating power peaks of 2500W+ and it doesn't bulge aside from the normal give of "commuter mode" (medium/soft mode for other manufacturers)

However when you say "Accelerating to keep up with traffic near max speed is not an excuse" for the Bump on the road scenario described earlier, that's where it's easy to miss what can happen then as side-effect.

Since the beep is disabled for 5 seconds after being triggered by the bump, the rider doesn't know he/she is riding near max speed.
There's no doubt that if the rider was ignoring a warning then, a cutout would be rider error.
But in this situation, there's no warning.
Especially when navigating traffic around other vehicles the perception of speed is altered, and it is difficult to guess if the current speed is close to or at the maximum - unless being told by the wheel.

[Rawnei]

13 hours ago, Wraith Rider said:

As I agree with this, from my experience INMOTION is still far safer than any other EUC company.  Inexperience not understanding the limits of the wheel is user error.  Of course INMOTION can push farther development to make the wheel safer but any wheel can be made to cut out.  Accelerating to keep up with traffic near max speed is not an excuse.  

That's a weird take, as presented in this thread there's evidence of the safety mechanisms not working as intended, it has nothing to do with experience or user error.

[wohal]

The new firmware is out, is this issue solved?