Engineering Stack Exchange is a question and answer site for professionals and students of engineering. It only takes a minute to sign up.
Sign up to join this community
I have 3-axis accelerometer and 3-axis gyroscope sensor data coming from my bike trip. Until now I have extracted only basic information like overspeeding, hard acceleration and sudden turns.
What additional motion characteristics can I infer from this data? I also have x, y coordinates.
This information you have, in theory, allows complete reconstruction of location and orientation of the bike along the entire path. This is basically what is referred to as inertial navigation.
However, the problem in real life is that there will be some error on all the signals. Since determining location and orientation require integrating the signals you have, these errors accumulate over time. Note that position is the second integral of acceleration. After some time, the resulting location and orientation data will be so far off as to be useless.
For small and cheap sensors like you can put on a bicycles, the integrals may only be useful for a few seconds, tens of seconds at most. I once worked on a device that tracked head motion during a golf swing, and it likewise had 6-axis acceleration and rotation rate data. The useful integration time window was about 2 seconds.
So to really answer the question, you can only use this kind of data for characterizing short term conditions or events. If used in conjunction with something that is long-term stable, like GPS, then you can use it to fill in the high frequency information between GPS readings. However, for an ordinary bike ride, the extra position accuracy is of little use.
You can use this data on its own for its high frequency information, like smoothness/roughness of the road, number of larger pothole incidents, how wobbly/steady the bike is, how much it sways side to side when you are pedaling, when and how hard you braked, etc.
