Deep Reinforcement Learning for Power Controlled Channel Allocation in Wireless Avionics Intra-Communications

Abstract

Wireless avionics intra-communications (WAIC) can play an important role in alleviate the issue of fuel consumption, stability and maintenance costs over traditional wired avionics systems. However, in order for WAIC system to coexist with other systems in aircraft, the transmitted power level of WAIC is strictly limited to 50 mW. Meanwhile, WAIC require extremely low outage probability for the safety-critical avionics applications. Hence, it is urgently needed to effectively allocate channels and utilize the limited power while ensuring the reliability and real-time performance of the WAIC system. In this paper, a deep reinforcement learning (DRL)-based power controlled channel allocation (DRL-PCCA) scheme for WAIC network is proposed, with physical layer of frequency hopping orthogonal frequency division multiplexing (FH-OFDM). First, we formulate a sub-bands allocation and power control optimization problem whose aim is to minimize the overall transmit power provided that all end nodes achieve their requested data rates and desired bit error rate (BER). However, the problem formulated is non-convex and NP-hard. To tackle this problem, we propose a DRL-based scheme, which can effectively solve the optimization problem of sequence decision making in complex environment by using neural networks to extract spatial correlation features of WAIC. In particular, a reliability framework for WAIC is presented to analyse the performance of the proposed scheme against the system requirements of the flight certification. Simulation results demonstrate that the performance of proposed DRL-based scheme is superior to the traditional power control and channel allocation scheme.

Keywords

• Channel allocation
• deep reinforcement learning
• UWB
• WAIC
• wireless communication