DEEP ACTOR-CRITIC FOR CONTINUOUS 3D MOTION CONTROL IN MOBILE RELAY BEAMFORMING NETWORKS

The paper studies the motion control for mobile relays implementing cooperative beamforming to aid the communication between a source-destination pair. We consider an urban communication scenario, where the channels exhibit spatiotemporal correlations and thus can be learned. The relays move in a time-slotted fashion within a three-dimensional cube. During every slot, the relays beamform optimally to maximize the Signal-to-Interference+Noise Ratio (SINR) at the destination and decide their positions for the next slot. Unlike prior works that assume knowledge of channel statistics, our proposed approach is model-free. Also, typically, prior approaches assume discrete motion on the two-dimensional plane. However, as discretization introduces the curse of dimensionality, those methods do not easily extend to three-dimensional motion. We propose a model-free, continuous control actor-critic approach that can be easily applied to 2D and 3D motion with the same complexity . To address the random nature of the channel, we propose to use Sinusoidal Representation Networks (SIRENs) for value function approximation. Our approach outperforms the direct application of the State-of-the-Art continuous control algorithms for both 2D and 3D cases.