2005 & 2006 TDM900A (ABS) still have the same speed sensor on the front sprocket along with the front & rear wheel speed sensors used by the ABS ECU.
Latter ABS models only have wheel speed sensors, no sprocket speed sensor.
Â
Â
Some time ago I worked on embeded GPS & accelerometers sensors.
The problem with the accelerometers is the noise & error building. Noise may be filtered (with a Kalman filter or just a low-pass filter) but the error increase remains a problem when trying to derive speed from acceleration.
The idea was to overcome the GPS speed data output - data is outputed once per second on most embeded GPS - to give a more frequent speed data using an accelerometer, and fixing increasing errors every second with the GPS data.
Â
One thing to know with a GPS is that the speed data is not a computed data using the distance between coordinates divided by time, which would be very inaccurate (up to 20% error with few satellites in sight). They actually use the doppler frequency shift on received signals to compute a much more accurate speed.
Maybe using only the GPS speed data could allow to compute acceleration data without the noise & error issue with an accelerometer, but that would be only once per second. No idea if this is enough for a brake warning signal...
Just a though on your very interesting project... :good:
Latter ABS models only have wheel speed sensors, no sprocket speed sensor.
Â
Â
Quote:Correct Steve ABS wheel speed sensor supercedes front sprocket sensor on ABS bikes. It should be simple to read the pulse output from it just the same. Although it may make some people nervous doing so !Â
I've wasted a perfectly good day off work tweaking the accelerometer version and its still not quite 100 %.....mind you i had to revisit digital filtering principles and thats not a 2 minute task !
Some time ago I worked on embeded GPS & accelerometers sensors.
The problem with the accelerometers is the noise & error building. Noise may be filtered (with a Kalman filter or just a low-pass filter) but the error increase remains a problem when trying to derive speed from acceleration.
The idea was to overcome the GPS speed data output - data is outputed once per second on most embeded GPS - to give a more frequent speed data using an accelerometer, and fixing increasing errors every second with the GPS data.
Â
One thing to know with a GPS is that the speed data is not a computed data using the distance between coordinates divided by time, which would be very inaccurate (up to 20% error with few satellites in sight). They actually use the doppler frequency shift on received signals to compute a much more accurate speed.
Maybe using only the GPS speed data could allow to compute acceleration data without the noise & error issue with an accelerometer, but that would be only once per second. No idea if this is enough for a brake warning signal...
Just a though on your very interesting project... :good:


![[Image: top_640.png]](http://jbx9.free.fr/tmp/top_640.png)