kriswiner
commented
6 years ago The user selects which edge of the board is to be defined as North, then everything else floows from that. This is why different users might have different definitions of NED.
On Tue, May 1, 2018 at 8:52 AM, ifrew notifications@github.com wrote:
Hi Kris,
First off thanks for all your work in this area!
I'm having a hard time trying to.get my head around why in your examples the mappings for NED change for the the same 9250 chip. Currently using your mpu9250-5611 code for the breakout board Im using. If the x axis points forward away from you that is North, correct? Based on your comments in the code and in the datasheet, the x-y accel and gyro axis are aligned with the y-x of the magnetometer and the + Z axis of the mag points down while the accel points up.
Given that, why wouldn't the mappings then stay the same for each example.
For example, in the 9250-5637 example you have
// Sensors x (y)-axis of the accelerometer/gyro is aligned with the y (x)-axis of the magnetometer; // the magnetometer z-axis (+ down) is misaligned with z-axis (+ up) of accelerometer and gyro! // We have to make some allowance for this orientation mismatch in feeding the output to the quaternion filter. // For the MPU9250+MS5637 Mini breakout the +x accel/gyro is North, then -y accel/gyro is East. So if we want te quaternions properly aligned // we need to feed into the Madgwick function Ax, -Ay, -Az, Gx, -Gy, -Gz, My, -Mx, and Mz. But because gravity is by convention // positive down, we need to invert the accel data, so we pass -Ax, Ay, Az, Gx, -Gy, -Gz, My, -Mx, and Mz into the Madgwick // function to get North along the accel +x-axis, East along the accel -y-axis, and Down along the accel -z-axis. // This orientation choice can be modified to allow any convenient (non-NED) orientation convention. // Pass gyro rate as rad/s MadgwickQuaternionUpdate(-ax, ay, az, gxPI/180.0f, -gyPI/180.0f, -gz*PI/180.0f, my, -mx, mz);
And this is different from the 9250-5611 example. I hope you can elaborate as im getting wrong results in my code. Also, why are the gyro values negative for y and z. Rotational direction around the axis wouldn't change with an accelerometer axis sign change?
Of all the coding I've done with these devices this ned mapping has always confused me.
Cheers Iain
— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/kriswiner/MPU9250/issues/272, or mute the thread https://github.com/notifications/unsubscribe-auth/AGY1qm9uzNcDUTIZEFfH5fmeo9g7ADrcks5tuISrgaJpZM4TuHNF .
ifrew
eminyavuzaslan
Hi Kris,
First off thanks for all your work in this area!
I'm having a hard time trying to.get my head around why in your examples the mappings for NED change for the the same 9250 chip. Currently using your mpu9250-5611 code for the breakout board Im using. If the x axis points forward away from you that is North, correct? Based on your comments in the code and in the datasheet, the x-y accel and gyro axis are aligned with the y-x of the magnetometer and the + Z axis of the mag points down while the accel points up.
Given that, why wouldn't the mappings then stay the same for each example.
For example, in the 9250-5637 example you have
// Sensors x (y)-axis of the accelerometer/gyro is aligned with the y (x)-axis of the magnetometer; // the magnetometer z-axis (+ down) is misaligned with z-axis (+ up) of accelerometer and gyro! // We have to make some allowance for this orientation mismatch in feeding the output to the quaternion filter. // For the MPU9250+MS5637 Mini breakout the +x accel/gyro is North, then -y accel/gyro is East. So if we want te quaternions properly aligned // we need to feed into the Madgwick function Ax, -Ay, -Az, Gx, -Gy, -Gz, My, -Mx, and Mz. But because gravity is by convention // positive down, we need to invert the accel data, so we pass -Ax, Ay, Az, Gx, -Gy, -Gz, My, -Mx, and Mz into the Madgwick // function to get North along the accel +x-axis, East along the accel -y-axis, and Down along the accel -z-axis. // This orientation choice can be modified to allow any convenient (non-NED) orientation convention. // Pass gyro rate as rad/s MadgwickQuaternionUpdate(-ax, ay, az, gxPI/180.0f, -gyPI/180.0f, -gz*PI/180.0f, my, -mx, mz);
And this is different from the 9250-5611 example. I hope you can elaborate as im getting wrong results in my code. Also, why are the gyro values negative for y and z. Rotational direction around the axis wouldn't change with an accelerometer axis sign change?
Of all the coding I've done with these devices this ned mapping has always confused me.
Cheers Iain