基于空中计算CoMAC架构的不同计算场景叠加符号判决算法

Symbols decision metrics algorithms in different computing scenes based on the computation over multi-access channels architecture

本文研究不同场景下基于空中计算的多址信道计算(CoMAC)架构的覆盖符号决策算法。首先,从理论上分析了XOR、ADD、MOD三种场景中加性高斯白噪声(AWGN)多址接入信道下叠加符号的概率密度分布,提出了一种基于先验概率的最优门限判决策略。其次,推导了系统最优门限及对应误码率的理论表达式。最后,通过仿真验证了不同信噪比、传感器节点个数及先验概率对于该门限判决方案的鲁棒性和可靠性的影响。与通信计算相分离的传统方案相比,空中计算判决方案具有更好的检测性能,为多址接入信道下的信号识别提供了新的参考方案。

This study investigates overlay symbol decision algorithms via the computation over multi-access channels(CoMAC)architecture for different scenarios.Firstly,the probability density distributions of overlay symbols under additive white Gaussian noise(AWGN)multi-access channels in three scenarios,namely XOR,ADD,and MOD,are theoretically analyzed.A threshold decision strategy based on prior probabilities is proposed.Secondly,the theoretical expressions for the optimal threshold and corresponding error rate are derived.Finally,simulations are conducted to verify the robustness and reliability of the threshold decision scheme under different signal-to-noise ratios,numbers of sensor nodes,and prior probabilities.Compared to traditional approaches that separate communication and computation,the CoMAC decision scheme exhibits superior detection performance and provides a new reference for signal identification in multi-access channels.