Lean operated cornering lights?

[snowrx]

After a night of trying to make good time over a serpentine pass posted for 35 mph, I really feel the need for some lights to fill the blind spot inside the corner when you're leaned over. Such up and out pointing lights can't be on all the time, so there needs to be a control circuit.
I'm trolling for input from anyone working on such a DIY light project.

It would seem with the current availability of inexpensive flight control Inertial Measurement Unit's (multi-axis accelerometer/Gyroscope/Compass combos) for cell pones and drones, and cheap Arduino processors, this would just be a matter of some one writing the code to give you switched outputs for two (or sequentially more, ala KTM) cornering lights. If you can control four motors in stable flight with a $35 flight board, DIY lean sensitive lighting should be possible for under $100. I've not given it a lot of thought, but it would seem that hanging in the air under propellers would be the inverse of our gravity acting along the bike center line as far as lean sensing.

Ebay IMU's can be had for as little $8, Processors for $5, assorted relays and power switch, housing and wires maybe $20. Code/design; priceless at the moment.

I'd like to test a FPV camera gimble on a cornering bike. If the thing tracks OK, then it could be a matter of adapting the servos to control light weight reflectors or lights to a level attitude. Although this mechanical method would be much harder to adapt to different bikes than a set of fixed LED lights slapped on a fork or fairing. I think BMW does this with a reflector inside the head light unit.

[virgilmobile]

How bout a simple mercury tilt switch...One for each side and let it turn on the light.You can find them in the junk yard in older vehicle hood and trunk.

[snowrx]

Many have tried gravity switches, but the force vector of the cornering motorcycle at speed is mostly along the plane of the frame, so the switch action does not track with lean. Also the heavy blob of mercury (or more commonly these days, a metal ball) in a tilt switch will bounce around in it's tube, giving you a strobe light.

Someone suggested using the aerielon out puts from a R/C plane stabilizer unit, which has an IMU incorporated into it, to operate PWM switches.
I would think that by the time you added circuits to emulate the receiver input and the switches, you'd be into more complexity than a dedicated Arduino or similar.

[virgilmobile]

Good analysis. I'll look at the available sensor hardware available for the arduino.Some of us can do the generic programming needed.
I lost my programming board and computer in the flood but I still can research.

[WingAdmin]

I've thought a bit about this as well. The problem with using solid state accelerometers to detect lean angle is that the apparent force of gravity remains straight down through the frame during the turn. You could detect lateral acceleration (i.e. turning) by the increase of apparent "g" straight down. If the force of gravity straight down into the frame increases from 1.0 (stationary) to say a constant 1.4, you can safely assume that you are turning a corner, and the lateral acceleration diverging from a straight line of travel (i.e. centrifugal force) is causing the apparent gravity to increase.

The problem is...how do you detect which direction you're turning?

There's two ways: one is to monitor the steering angle. That takes mechanical hardware.

The other would to have a second accelerometer mounted to detect rotation of the bike along its longitudinal axis (i.e. lean angle). If it was mounted high up in the bike, it could simply detect lateral acceleration (i.e. displacement).

Perhaps it could also monitor speed pulses from the cruise control to work out the bike's speed.

I see the algorithm something like this:

- If speed is greater than a given amount (say 15 mph) then
- If downward "G" is 1.0 then assume the bike is upright and not turning, so reset the lateral displacement to "0" (upright)
- If lateral acceleration increases momentarily, record it as a change in bank angle, and identify the direction turned
- If downward "G" increases for a given amount of time, check the last lean angle recorded to identify the direction of the turn, and illuminate the appropriate cornering light
- If downward "G" decreases below a specific amount, extinguish the cornering light

There would need to be a fair amount of hysteresis management and damping written into the algorithm in order to manage false information recorded by bumps and other transients. The lateral (lean) accelerometer in particular would need to be continually reset to its baseline when no increase in downward "G" was detected for a given time period (indicating bike is upright and traveling in a straight line).

[alwrmcusn]

Just my two cents worth: If your light does not come on until you are already leaned over, you're a few seconds too late with it.

[WingAdmin]

Just my two cents worth: If your light does not come on until you are already leaned over, you're a few seconds too late with it.

That's a very good point. My wife's car has headlights that move left and right as you steer around corners. The problem is, by the time you're steering around the corner, you should have already looked to see what was there, so it's a little late!

So if the process is look-lean-steer, having the leaning lights come on only when the "steer" part is happening is kind of late. Perhaps what we need is a helmet mounted device that detects when your head has turned one way or the other for more than a second or two, and when it does, turns on the lean light in that direction. That way when you look - the bike lights up what you're looking at, BEFORE you lean into the turn.

There is one other aspect to this however. You can easily have lights that shine off to the side and illuminate what you are GOING to steer into. The problem is (unique to motorcycles) that when you lean into the steer, the light on that side is now shining into the ground, instead of illuminating where you are going. So perhaps the motorcycle-lean-actuated light IS the solution for this.

[offcenter]

The problem is (unique to motorcycles) that when you lean into the steer, the light on that side is now shining into the ground, instead of illuminating where you are going.

...and the reverse is also true. When you are NOT leaning, that side light will be shining up into
someone's face in the oncoming lane.

How about an infrared laser shining at the ground beneath the corning light, measuring the distance
from the light to the ground. As the bike leans, the distance to the ground decreases and the
servo angles the light upward to keep it pointing level out the side. When the bike is riding
straight and level, the servo would keep the angle of the light low and just light up the road
a bit to either side of the bike.
This technology already exists in cheap laser "tape measures" the carpenters
use to measure the length of a room.

[landisr]

I vote for low tech. Mount fixed lights (non-pencil beam) on each side focused at an appropriate angle to flood the lead angles, and connect momentary switches under (or on top of?) each switch housing.

My 2 cents.

Ron

[alwrmcusn]

I say give it up and ride a trike

[FM-USA]

How about an infrared laser shining at the ground beneath the corning light, measuring the distance
from the light to the ground. As the bike leans, the distance to the ground decreases and the
servo angles the light upward to keep it pointing level out the side. When the bike is riding
straight and level, the servo would keep the angle of the light low and just light up the road
a bit to either side of the bike.
This technology already exists in cheap laser "tape measures" the carpenters
use to measure the length of a room.

More electronics to go wrong.
What was TUCKER's center headlight solution?

MAYBE take that IR sensor and hook it up to your aux. lights. You tip, they both turn on.
The opposite side will shine skyward offering landing lights for Aliens.

[robb]

Always thought the cornering lights were a necessity living so far in country area where dark roads are all you get. Few years back there was a big discussion over converting cornering to running lights. Since I already had upper running lights I installed a toggle switch to break the wire to allow lights to run continually when in country but retain ability to still operate with turn signals while in high traffic conditions. Worked flawlessly without addition of future headache when things go wrong.

[waituntilthebeep]

Or even a lower tech solution... a momentary switch that acts like turn indicator selector switch that you just grab with your thumb to select which side the light turns on. That way you can preselect way before some semi-smart component decides you need it. Also makes it so if in the corner you encounter another vehicle, you can easily cancel the light for a second until the car clears the lights broadcast area. Let go of the switch or push button and they go dark. Depending on what model of bike you have and what switch setup is on it, you may even have a CB channel selector switch that you could repurpose just for that function.

[robb]

The angle or cornering lights is such that it does not interfere with oncoming traffic. Little forward light but they do cover the ditches.