| Title |
CNN-based Channel Estimation Techniques for Cellular V2X Systems |
| Authors |
황재동(Jaedong Hwang) ; 김명진(Myoung Jin Kim) |
| DOI |
https://doi.org/10.5573/ieie.2020.57.3.25 |
| Keywords |
V2X ; CNN ; Channel Estimation ; C-V2X ; Sidelink ; |
| Abstract |
V2X (Vehicle-to-Everything) technology is attracting attention as a key technology for autonomous driving. Especially a great deal of research has been carried out mainly on C-V2X (cellular V2X) based on LTE-A or 5G NR (New Radio) cellular network. To achieve autonomous driving beyond simple safe driving, ultra-low latency and high data throughput at high speeds must be ensured. When the relative speed of the vehicle is high, the channel tends to change its characteristics rapidly so that reliable communication in time varying frequency selective fading environments should be performed. Therefore, reliable channel estimation is required in high-speed environments. In this paper, we apply CNN (Convolutional Neural Network) based channel estimation technique to LTE-A based C-V2X sidelink communication and verify its validity. The proposed channel estimation scheme uses CNN weights learned in various Doppler frequency environments and operates adaptively to the Doppler frequency which varies according to the speed and moving direction of the UE. Simulation results show that the proposed channel estimation technique outperforms LS (Least Square) channel estimation technique in terms of MSE and BER. |