I have an existing wrist-worn product that uses the MMA8451Q and I would like to add human fall detection as a feature. I've read some research papers and looked at existing products like the Philips Lifeline and all use a pendant or attach the sensor to the wait or torso. I do not want to waste my time trying to make fall detection work on a wrist device if the false detection rate is too high. Does anyone have any opinions or research on wrist-worn fall detectors? Also, I saw a NXP reference manual titled MMA7260QHFDRM Human Fall Detection Using 3-Axis Accelerometer but could not find any of the sample code it mentions. Is this manual still valid and can I get the sample code somewhere? Is there a sample algorithm for the MMA8451Q on fall detection? I do not believe the freefall detection interrupt is sufficient to cover human falls.
已解决! 转到解答。
Hello Jeff,
In this case, the false detections might be avoided using the FF_MT_THS and FF_MT_COUNT registers from the device.
Please find more detailed description at page 32 and 33 from the datasheet.
If you are interested, please let me know. I will be glad to help.
Have a great day.
David
Hello Jeff,
In this case, I would like to recommend using one of the embedded functions included in the MMA8451Q: Freefall detection.
The detection of “Freefall” involves the monitoring of the X, Y, and Z axes for the condition where the acceleration magnitude is below a user specified threshold for a user definable amount of time. Normally, the usable threshold ranges are between ±100 mg and ±500 mg.
Please find attached a sample project in which an interrupt is generated when a freefall event has been detected.
I used the FRDM-KL25Z and CodeWarrior IDE for this project.
I hope this information will be useful for you.
If I misunderstood your question, feel free to let me know. I will be glad to help.
Have a great day.
David Diaz.
Thanks David. My main concern is that a wrist-worn device will cause too many false detections. I was wondering if anyone has attempted to detect falling from a wrist sensor or know of any end product I can buy to test for myself before trying to implement this.
Hello Jeff,
In this case, the false detections might be avoided using the FF_MT_THS and FF_MT_COUNT registers from the device.
Please find more detailed description at page 32 and 33 from the datasheet.
If you are interested, please let me know. I will be glad to help.
Have a great day.
David