Place AR elements on top of a video feed using sensor data from different device

Question

I need to overlay AR elements on top of a video feed from a general-purpose camera and adjust them using an external sensor unit. For example, Let's say I mount a webcam on top of a mobile phone and move them as a single unit. The webcam feed serves as the video input, and I use the mobile phone's sensors to track the camera's movement. I have tried a few things with the help of several YouTube videos and tutorials, and I am currently able to track the device's rotation using Unity Remote 5 app and move the virtual camera with the rotation using unity 3d. However, I am having trouble detecting the device’s movement when I move the unit left/right, forward/backward, etc. I think although gyro and accel can be used to track the rotation they cannot handle the moment well (as I think)

At the moment, I have tried several things but still couldn't find any path to proceed. If you have any suggestions or ideas I could explore, it would be greatly appreciated. I am open to any suggestions. (even any other methods to achieve this)