Decision Feedback Equalizer Buffer for ATSC DTV Convergence
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing equalizers face difficulties in converging under severe multipath conditions in ATSC DTV signal reception, leading to long convergence times and error propagation issues due to the need for long channel impulse responses and infrequent training sequences.
Innovation Solution
A method is introduced to improve equalizer convergence by continuously storing input data segments in a decision feedback equalizer buffer, supplying output data sections at a reduced rate with void times, and calculating tap weights based on both received and decoded symbols during these void times, allowing for more frequent updates of the channel impulse response and tap weights.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the equalizer uses a long channel impulse response to handle severe multipath conditions, then the equalization accuracy is improved, but the convergence time increases significantly
Solution Approach 1:
The patent segments the long channel impulse response into multiple shorter sections, each processed independently by separate equalizer units. This allows parallel processing of different signal portions, maintaining the effectiveness of handling long multipath channels while reducing the computational burden and convergence time of each individual equalizer unit.
Solution Approach 2:
The patent performs preliminary channel estimation and equalizer initialization using training sequences before actual data transmission. This preliminary action pre-configures the equalizer parameters, significantly reducing the convergence time required during actual operation while maintaining high equalization accuracy for the long channel impulse response.
2Speed
If training sequences are used frequently to improve equalizer convergence, then the convergence speed is improved, but the data transmission efficiency decreases
Solution Approach 1:
The patent applies partial training sequences rather than complete re-training, updating only the necessary portions of equalizer parameters that have drifted due to channel variations. This partial action maintains convergence speed while minimizing the overhead and preserving data transmission efficiency.
Solution Approach 2:
The patent implements periodic updates of equalizer parameters at optimized intervals rather than continuous re-training. This periodic action maintains adequate convergence speed by updating parameters only when channel variations exceed a threshold, thereby preserving data transmission efficiency during stable channel conditions.
3Measurement precision
If the equalizer adapts continuously to channel variations, then the tracking accuracy is improved, but the computational complexity increases
Solution Approach 1:
The patent implements dynamic adaptation where the equalizer continuously tracks channel variations by adjusting its parameters in real-time based on incoming signal statistics. This dynamic behavior maintains high tracking accuracy while the system automatically scales the adaptation rate to balance computational complexity, using faster adaptation when channel conditions change rapidly and slower adaptation when conditions are stable.
Data Source
AI summary
Input data segments of received symbols are continuously stored in a decision feedback equalizer buffer at a symbol rate S. Output data sections of received symbols are supplied from the decision feedback equalizer buffer at an output rate of nS such that void times separate the output data sections, and n>1. The received symbols supplied by the decision feedback equalizer buffer are equalized in a decision feedback equalizer to provide equalized symbols; and the equalized symbols are decoded by a decoder to provide decoded symbols. Adjustments for the decision feedback equalizer are calculated during the void times such that the adjustments are calculated based on both the received symbols supplied by the decision feedback equalizer buffer and the decoded symbols. The adjustments are applied to the decision feedback equalizer.