De-embedding Motion Artifacts from Robotic Arm Assisted Propagation Measurements

This paper introduces a method to de-embed motion artifacts introduced by cable movement during millimeter wave propagation measurements that involve motion. In our work, the motion is performed by a Sawyer robotic arm emulating a UAV motion profile disturbed with gusts generated by the Dryden wind model. An RF cable connects the Rx from the end effector of the arm to a Keysight VNA taking time-swept continuous wave measurements. The cable moves with the arm, introducing phase changes to the measurement data. After determining the cable S-parameters in the continuous time domain, they are removed from the measurement data to get the channel parameters. Furthermore, the effects of de-emebedding is examined by comparing the Doppler spread of measurement data before and after de-embedding. The noise floor in the channel shifted from -62 dB before de-embedding to -58 dB after deembedding. The observed Doppler spread reduced by 31.67 Hz.