Dependency-Aware Reliability Allocation for Open-Vocabulary Scene-Graph Semantic Packets over Latency- and Energy-Constrained Wireless Visual Uplinks

2026-08-08

AI Summary

This work addresses the challenge of preserving semantic interpretability in wireless visual uplink transmissions under stringent delay and energy constraints, where open-vocabulary scene graph semantic packets suffer from degraded reliability due to vocabulary expansion and triplet dependencies. To tackle this, the authors propose constructing a semantic packet dependency graph and jointly optimizing modulation and coding schemes, transmit power, and the maximum number of HARQ retransmissions to minimize dependency-aware semantic distortion while satisfying frame-level latency and energy budgets. The approach introduces a shared downstream enablement value mechanism and a semantic dependency Lagrangian block method, enabling precise modeling of dependency structures and yielding near-global optimal solutions within a limited candidate set. Experiments demonstrate that the method achieves global optimality in 28 out of 30 test instances, with an average relative error of merely 0.052%, significantly outperforming baseline approaches by reducing semantic failure rates and enhancing critical query success rates.

Abstract

Wireless visual perception uplinks increasingly transmit structured semantic packets rather than raw images or opaque feature tensors. In an open-vocabulary scene graph, an indexed subject-relation-object triplet is interpretable only if its own packet and all prerequisite vocabulary-increment packets are recovered, creating a dependency structure not captured by conventional packet-loss, layer-priority, or semantic-feature HARQ rules. We study dependency-aware reliability allocation for such packets under expected frame-delay and frame-energy constraints. The model separates vocabulary-increment and indexed-triplet packets, builds their prerequisite graph, and derives Chase-combining HARQ residual failure, feedback-aware delay and energy, and dependency-aware semantic distortion. The resulting finite discrete problem jointly selects each packet's modulation and coding scheme, transmit power, and maximum HARQ depth. Unlike fixed-priority UEP, a vocabulary packet receives a shared downstream enabling value that depends on the current reliabilities of other prerequisite and triplet packets. Exploiting the resulting multi-affine distortion structure, the proposed semantic-dependency Lagrangian block method performs exact finite-candidate minimization for each packet update. In a small-scale exhaustive audit over 30 reduced-menu instances, the method reaches the global optimum in 28 cases, with mean and worst observed relative gaps of 0.052% and 1.10%. Analytical link-budget sweeps, public scene-graph traces, multi-session and multi-seed validation, and table-driven BLER tests show lower vocabulary-induced semantic failure and higher critical-query success than dependency-agnostic HARQ, semantic-priority HARQ, priority UEP, UEP-only, and uniform HARQ.

Problem

• dependency-aware reliability allocation
• open-vocabulary scene graph
• semantic packets
• wireless visual uplinks
• HARQ

Innovation

• dependency-aware reliability allocation
• open-vocabulary scene graph
• semantic HARQ
• multi-affine distortion
• wireless visual uplink