Channel polarization-oriented multi-modal data semantic coding and decoding and reliable transmission method

By extracting anchor points and target modal features from multimodal data, and adjusting the path metric by combining channel polarization effects and cross-modal semantic distortion, the problem of pruning errors in traditional decoders under poor channel conditions is solved, thus achieving reliable transmission and semantic coherence of multimodal data.

CN122339629APending Publication Date: 2026-07-03NANJING LUKOU INT AIRPORT AIRPORT TECH CO LTD
0 Cites 0 Cited by

Patent Information

Application Number
CN202610416628.2
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-04-01
Publication Date
2026-07-03

AI Technical Summary

Technical Problem

In existing communication systems, under harsh channel polarization conditions or deep fading environments, traditional SCL decoders are susceptible to noise interference, leading to erroneous pruning and making it impossible to achieve reliable transmission of multimodal data. Furthermore, the CRC check mechanism results in a high communication interruption rate, failing to meet the requirements for semantic coherence and reliable recovery.

Method used

By extracting anchor points and target modal semantic features from multimodal data, calculating the reliability of polarized subchannels based on channel polarization effects, prioritizing the decoding of anchor point features, and adjusting path metrics in conjunction with cross-modal semantic distortion, polarization coding and reliable transmission are achieved.

Benefits of technology

Under deep fading limits, ensure the preservation of core semantics, reduce false pruning rate, achieve reliable communication in complex scenarios, and meet the needs of continuous semantic transmission.

✦ Generated by Eureka AI based on patent content.
Patent Text Reader

Abstract

The invention discloses a channel polarization-oriented multi-modal data semantic coding and decoding and reliable transmission method. The method comprises the following steps of: acquiring anchor points and target modal semantic features of multi-modal data; calculating the reliability of polarization sub-channels, asymmetrically mapping the two types of features to different reliable intervals according to priorities, and encoding and sending the features; a receiving end preferentially decodes the anchor point features to reconstruct a cross-modal semantic prior vector; carrying out decoding splitting on the target features by adopting an SCL algorithm, reconstructing local semantic features of a current splitting path, and calculating a cross-modal semantic distortion degree with a prior vector; and combined path measurement is dynamically generated in combination with polarization characteristic parameters to perform list pruning, so that a cyclic redundancy check red line is broken, retransmission is blocked, and a path with minimum semantic distortion is directly output. According to the scheme, splitting of a physical layer and a semantic layer is broken, false pruning and retransmission caused by local noise under a deep fading channel are avoided, and semantic coherence and reliable transmission under a complex environment are achieved.
Need to check novelty before this filing date? Find Prior Art