研究目的
Designing an extreme learning machine (ELM) based receiver to jointly handle the LED nonlinearity and cross-LED interference in LED MIMO communications.
研究成果
The ELM-based receivers effectively handle LED nonlinearity and cross-LED interference, offering significant performance gains over conventional methods. The low-complexity ELM receiver with circulant input weight matrix performs comparably to the conventional ELM but with reduced computational complexity.
研究不足
The study focuses on LOS propagation and assumes a specific LED nonlinearity model. The complexity reduction technique may have limitations in scenarios not suited for FFT implementation.
1：Experimental Design and Method Selection:
The study employs an ELM with a single-hidden layer feedforward neural network to mitigate LED nonlinearity and cross-LED interference. A circulant input weight matrix is proposed for low complexity implementation using FFT.
2：Sample Selection and Data Sources:
The study uses a practical LOS LED communication system setup with parameters based on a commercial LED (Kingbright AA3022EC-4.5SF).
3：5SF).
List of Experimental Equipment and Materials:
3. List of Experimental Equipment and Materials: Includes a 5.0 m × 5.0 m × 3.0 m room setup, LEDs, PDs, and a commercial LED for nonlinearity modeling.
4：0 m × 0 m × 0 m room setup, LEDs, PDs, and a commercial LED for nonlinearity modeling.
Experimental Procedures and Operational Workflow:
4. Experimental Procedures and Operational Workflow: The ELM-based receiver is trained with a sequence, and its performance is compared with conventional techniques using SER metrics.
5：Data Analysis Methods:
Performance is evaluated using symbol error rate (SER) comparisons between the proposed ELM-based receivers and conventional methods.
独家科研数据包，助您复现前沿成果，加速创新突破
获取完整内容
