How to proceed with a IMU based odometry ?
I have finished making a robot and its URDF model . The robot consists of 2x250 watt ebike motors , sabertooth motor controller , logitech USB webcam and a MPU6050 .
How do i proceed with a pure IMU based implementation for receiving the odometry feedback ?
I am reluctant to use wheel encoders because , its too difficult to add to the current design (attaching the encoder to the wheel or the motor ).
You don't really have an easy way to get odometry directly from the IMU, since you will need to integrate the acceleration data from the x axis twice, which is not really recommended, you can do it, but it would be very unreliable and noisy.
One option would be to run some kind on visual odometry like viso2
A simple VO , will it run on a RPI3 ?
VO uses SIFT/SURF descriptors to update the pose/position of the robot .
My current robot uses a RPI3 , do you think i will be able to handle localization within say a room .
I also want to get odometry (pose and twist) using imu (mpu 6050)? Were you able to do this?