Movable antennas keep gradient learning accurate during user motion
Keep Pace to Learn: Movable Antennas Restore the Backpropagated Gradient Under User Mobility
Information Theory
Summary
When people move while using devices that learn from wireless signals, their motion can twist important feedback signals, making learning less accurate. The researchers show that antennas that move along with the user can undo this twisting effect without needing detailed signal information. They provide mathematical rules for how fast the antenna should move and explain how small measurement errors affect learning. Their simulations confirm that this method nearly matches ideal conditions and that careful calibration is more important than reducing random noise.
What this means in practice
- •For wireless communication engineers: Build antenna systems that move with users in wireless training to maintain accurate signal-based learning without complex channel estimation.
- •For robotics hardware designers: Design movable antenna actuators to compensate user motion effects on wireless feedback when implementing over-the-air machine learning in mobile robots.
Authors
Xianzhong Li, Yuzhi Yang, Tianhao Guo
Abstract
When a mobile terminal takes part in over-the-air training, its own motion rotates the gradient that travels back through it; a movable antenna that keeps pace with the terminal rotates it back---without any channel estimate. We give a speed law: under a scalar conditional-mean model of the physical return, landing the residual displacement inside the first positive lobe of the spatial correlation keeps the conditional mean gradient a descent direction, which, given sufficient travel, fixes a critical actuator-to-terminal speed ratio. We further prove a bias--jitter law: under the same scalar conditional-mean model and with zero effective bias, independent zero-mean Gaussian jitter of any finite amplitude scales the conditional mean gradient without reversing its direction, whereas a sub-wavelength systematic bias can reverse it. This result motivates prioritizing bias calibration to preserve gradient direction. Simulations on 3GPP CDL-A channels show that an antenna keeping pace with the terminal achieves near-CSI accuracy without digital channel correction, reaching 98.58 % compared with 98.59% for the ideal-CSI reference. The simulations also show that partial compensation can support learning when the residual displacement falls within a later positive correlation lobe, so a faster actuator does not necessarily yield higher accuracy. Separate positioning simulations show a sharp accuracy loss under systematic bias but little degradation under zero-mean Gaussian jitter of comparable spatial scale.