Adaptive Transversal Filter Tap Scaling for Lower Quantization Noise
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional adaptive transversal filters suffer from reduced precision for small tap weights, leading to increased quantization noise, which affects the accuracy of signal processing.
Innovation Solution
The implementation of tap weights as products of corresponding tap coefficients and gains, with a filter control loop managing coefficients to minimize error signals and a tap control loop adjusting gains to satisfy specific conditions, such as maximizing coefficient magnitude within a constraint, reduces quantization noise.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional adaptive transversal filters use direct tap weight implementation, then the filter structure is simple, but precision for small tap weights deteriorates leading to increased quantization noise
Solution Approach 1:
The tap weight Wk is segmented into two separate components: a coefficient Ck and a gain Mk. This segmentation allows the coefficient to be optimized for precision while the gain handles the scaling, thereby reducing quantization noise for small tap weights without complicating the overall filter structure.
Solution Approach 2:
The invention changes the parameter representation from direct tap weights to a product of coefficients and gains. By transforming Wk = Ck × Mk, the system can independently optimize Ck for precision and Mk for scaling, effectively reducing quantization noise while maintaining filter functionality.
2Measurement precision
If tap weights are implemented directly without decomposition, then the filter control loop is simple, but precision for small tap weights deteriorates
Solution Approach 1:
The tap weight is segmented into coefficient and gain components, each managed by appropriate control mechanisms. This segmentation improves precision for small tap weights while distributing the control complexity across two manageable loops rather than one complex loop.
Solution Approach 2:
The invention introduces feedback mechanisms in both the filter control loop and tap control loop. The filter control loop minimizes error signals to optimize coefficients, while the tap control loop uses feedback to adjust gains based on signal conditions, achieving high precision through coordinated feedback control.
3Object-generated harmful factors
If tap weights are used directly in filtering operations, then computation is simpler, but quantization noise increases for small weights
Solution Approach 1:
By segmenting tap weights into coefficients and gains, the system reduces quantization noise for small weights. The gain component handles scaling efficiently, while the coefficient provides precise filtering, maintaining computation efficiency while eliminating quantization noise issues.
Solution Approach 2:
The parameter transformation from direct tap weights to coefficient-gain products enables more efficient computation. The gain can be optimized for computational efficiency while the coefficient provides precise filtering, and their product maintains the desired filtering characteristics with reduced quantization noise.
Data Source
AI summary
An adaptive transversal filter having tap weights Wj which are products of corresponding tap coefficients Cj and tap gains Mj is provided. A filter control loop controls all of the tap coefficients Cj such that an error signal derived from the filter output is minimized. One or more tap control loops controls a tap gain Mk such that the corresponding tap coefficient Ck satisfies a predetermined control condition. For example, |Ck| can be maximized subject to a constraint |Ck|≦Cmax, where Cmax is a predetermined maximum coefficient value. In this manner, the effect of quantization noise on the coefficients Cj can be reduced. Multiple tap control loops can be employed, one for each tap. Alternatively, a single tap control loop can be used to control multiple taps by time interleaving.