Inertial Navigation System with Rift
Hey everyone,
I'm working on developing an application to allow for the Rift to act as a sort of inertial navigation system. The only really important information that is necessary for the program to function correctly is to have accurate and consistent readings on the velocity the rift is moving at. The rift will be moving in linear translations--not leaning forward or backwards but walking.
I've been able to get acceleration readings and have sanitized these to get rid of the gravity component (battling with the Equivalence Principle). The issue comes in that the acceleration readings are coming out to be very strange--lots of positive accelerations while the object is slowing down in the positive direction or negative acceleration while the object is speeding up in the positive direction. I use these readings to determine velocity using kinematics and the concepts of infinitesimals. I believe that the velocity (and the acceleration) are being thrown off by noise readings and plan on putting in a Kalman filter to try and fix the problem, but the wrong signing of the accelerations and the jumpiness of the signage is worrying me. Does anyone know if the readings from the Rift accelerometer can be accurately used to determine velocity?
I'm working on developing an application to allow for the Rift to act as a sort of inertial navigation system. The only really important information that is necessary for the program to function correctly is to have accurate and consistent readings on the velocity the rift is moving at. The rift will be moving in linear translations--not leaning forward or backwards but walking.
I've been able to get acceleration readings and have sanitized these to get rid of the gravity component (battling with the Equivalence Principle). The issue comes in that the acceleration readings are coming out to be very strange--lots of positive accelerations while the object is slowing down in the positive direction or negative acceleration while the object is speeding up in the positive direction. I use these readings to determine velocity using kinematics and the concepts of infinitesimals. I believe that the velocity (and the acceleration) are being thrown off by noise readings and plan on putting in a Kalman filter to try and fix the problem, but the wrong signing of the accelerations and the jumpiness of the signage is worrying me. Does anyone know if the readings from the Rift accelerometer can be accurately used to determine velocity?