Adaptive Filter Core Structure for Low-Complexity RLS Processing

The patent divides the filter into multiple filter cores, each responsible for specific calculation operations on input signals. The auxiliary core generates cross-terms between block correlation matrices and provides them to the filter cores, segmenting the complex matrix inversion task into manageable parts that can be processed in parallel, thereby reducing overall device complexity while maintaining performance.

The auxiliary core acts as an intermediary that generates cross-terms between block correlation matrices corresponding to different kernels. This intermediary component handles the complex matrix operations separately, allowing the main filter cores to focus on effective channel estimation without directly performing complex matrix inversions, thus reducing their complexity.

The patent designs filter cores and auxiliary cores that can be universally applied across multiple signal routes. Each filter core can process different kernels corresponding to different signal routes, and the auxiliary core can generate cross-terms for any combination of kernels. This multi-functional design allows the same hardware structure to handle multiple signal routes without requiring separate dedicated circuits for each route.

The patent employs a dynamic structure where filter cores and auxiliary cores can be selectively activated based on the number and characteristics of signal routes. The system can adaptively configure which kernels are used and which filter cores are active, allowing it to scale from simple single-route scenarios to complex multi-route scenarios without maintaining a fixed maximum-complexity structure.