Comprehensive Study of Least Squares Estimation in Fast Fading MIMO-OFDM Systems

MIMO-OFDM technology is now the foundation for modern wireless communication systems, allowing dramatic improvements in spectral efficiency, power efficiency, and transmission rate. On the other hand, the least accurate channel estimation is still a critical task, especially when they are in fast-fading environments. This study gives a comprehensive survey of LSE-based approaches and their applications on different fast-fading channel models for MIMO-OFDM systems. Techniques like Pilot Assisted Channel Estimation (PACE), Decision Directed Channel Estimation (DDCE), and Data Aided Channel Estimation (DACE) are examined, their advantages and disadvantages, and to what extend these techniques can be suited for application in dynamic environments. Deep learning-based approaches, including Fully Connected Deep Neural Networks (FDNN) and Convolutional Neural Networks Auto Encoder (CNNAE), offer a transformative approach in raising estimation accuracy. Additionally, emerging trends like compressive sensing for sparse millimeter-wave channels and 6G technologies are explored. This study aims to bridge existing research gaps, offering a roadmap for future innovations in channel estimation techniques and their applications in next-generation wireless networks.