Code needs to be written, as stated in the momentum page, that can act as a speedometer based on values from the accelerometers in the phone. This will be a difficult task to accomplish, and it will involve a lot of integral math. AI could potentially be used to determine whether data from the accelerometers is "noise" or actual movement. The accelerometer generally seems to have a ±0.2 accuracy, which shouldn't be too hard to deal with considering that significant movements and even friction is much greater. The difficult part will most certainly be gravity compensation.
Generally, accelerometers are not accurate enough to perform this task. However, the GPS module could be used instead, with help from the accelerometers on when significant motion begins.
Code needs to be written, as stated in the momentum page, that can act as a speedometer based on values from the accelerometers in the phone. This will be a difficult task to accomplish, and it will involve a lot of integral math. AI could potentially be used to determine whether data from the accelerometers is "noise" or actual movement. The accelerometer generally seems to have a ±0.2 accuracy, which shouldn't be too hard to deal with considering that significant movements and even friction is much greater. The difficult part will most certainly be gravity compensation.