thalmiclabs / myo-bluetooth Goto Github PK

I noticed that the quaternion definition in this header and in the Myo SDK is different:

/// Orientation data, represented as a unit quaternion. Values are multiplied by MYOHW_ORIENTATION_SCALE.
    struct MYOHW_PACKED {
        int16_t w, x, y, z;
    } orientation;

While the SDK defines it as

/// Index into orientation data, which is provided as a quaternion.
/// Orientation data is returned as a unit quaternion of floats, represented as `w + x * i + y * j + z * k`.
typedef enum {
    libmyo_orientation_x = 0, ///< First component of the quaternion's vector part
    libmyo_orientation_y = 1, ///< Second component of the quaternion's vector part
    libmyo_orientation_z = 2, ///< Third component of the quaternion's vector part
    libmyo_orientation_w = 3, ///< Scalar component of the quaternion.
} libmyo_orientation_index;

This is either inconsistent or a even a bug, because using the SDK definition with the Bluetooth data instead gives reasonable results while using the BLE header definition seems to result in wrong data for me.

I hope opening an issue is the right way to report this. When using this header I noticed that myos actual UUID of MotionEventCharacteristic is 0x0502, not0x0a02. This could be a firmware issue (using V1.5).

I've been trying to gather data from the Myo for a project, however, it keeps on going to sleep when in use. This is particularly problematic as I need to restart my PC in order to have it recognise any gestures again, which is unacceptable when gathering large quantities of data from a large number of people. What would the resolution to this be?