insfilterasync for smartphone pose estimation without GPS

Hi, there! I am using inertial sensors on smartphones to estimate their pose and trajectories. Although there is a tutorial about orientation estimation (e.g., insfiltersync), the pose estimation is not available with smartphones without GPS. Does anyone have such tutorials? Thanks!

Answer 1

[Brian Fanous]

I'd like to be sure I understand what you are trying to do. Do you want to estimate pose (orientation and position) based only on IMU data? Is that correct? Generally that's not feasible with low cost MEMS IMUs you'd find in a phone because of sensor biases.

[qniuaa]

Thanks for your kind reply, Brian. Yes, I'd like to estimate the pose of a smartphone in the indoor environment where GPS is not always available. I agree that it is not feasible to estimate the pose accurately with low-cost IMU in smartphones with long walking distances. But I am wondering if it is possible to estimate the pose with certain accuracy in a short distance, say, 3 or 5 meters.

[Brian Fanous]

I can’t offer you a suggestion as to how to proceed. If you plan to double integrate the accelerometer signal to determine position, your accelerometer bias will also double integrate and cause the error in your position estimate to increase quadratically over time.