In vehicular communications, reliable channel estimation is critical for the system performance due to the doubly-dispersive nature of vehicular channels. IEEE 802.11p standard allocates insufficient pilots for accurate channel tracking. Consequently, conventional IEEE 802.11p estimators suffer from a considerable performance degradation, especially in high mobility scenarios. Recently, deep learning (DL) techniques have been employed for IEEE 802.11p channel estimation. Nevertheless, these methods suffer either from performance degradation in very high mobility scenarios or from large computational complexity. In this paper, these limitations are solved using a long short term memory (LSTM)-based estimation. The proposed estimator employs an LSTM unit to estimate the channel, followed by temporal averaging (TA) processing as a noise alleviation technique. Moreover, the noise mitigation ratio is determined analytically, thus validating the TA processing ability in improving the overall performance. Simulation results reveal the performance superiority of the proposed schemes compared to recently proposed DL-based estimators, while recording a significant reduction in the computational complexity.

翻译：在车辆通信中,可靠的频道估计对于系统性能至关重要,因为车辆频道具有双重分散性。IEE 802.11p标准分配的试样不足以准确跟踪频道。因此,常规IEE 802.11p估计器的性能明显退化,特别是在高流动性情况下。最近,IEE 802.11p频道估计使用了深层次的学习(DL)技术。然而,这些方法在非常高流动性情况下或计算复杂程度很大,要么由于性能退化而受到影响。在本文件中,这些限制是通过长期短期内存(LSTM)的估计数解决的。拟议的估计器使用一个LSTM单位来估计频道,然后采用平均时间(TA)处理法作为缓解噪音的技术。此外,噪声缓解率是通过分析来确定的,从而验证了TA处理能力来改进总体性能。模拟结果显示,与最近提议的基于DL的估测器相比,拟议计划的业绩优于最近提出的高性,同时记录计算复杂性的显著下降。