Channel Estimation Optimization Model in Internet of Things based on MIMO/OFDM with Deep Extended Kalman Filter

Document Type : Original Article


Department of Computer Science, Saigon University, Ho Chi Minh City 700000, Vietnam


One of the most important parameters in the performance of wireless communication systems is the channel estimation. However, apart from the usual OFDM modes, there are also orthogonal conditions for modulation-based multi-channel systems which making channel estimation on networks such as the Internet of Things (IoT) more complex. In order to estimate the IoT channel, its type is considered as narrow or wide band. The narrowband IoT based on OFDM is considered in the present work. There are various classical methods for channel estimation such as Least Squares (LS) and Linear Minimum Mean Square Error (LMMSE). However, due to high computational complexity as well as inaccurate channel estimation and remaining weaknesses such as latency and other quality of service criteria, especially Bit Error Rate (BER), Signal to Noise Ratio (SNR) and Maximum to Average Power Ratio (PAPR), this research Improves these two methods based on the Deep Extended Kalman filter.