Task-Oriented Semantic Communication for Hazard Warning and Remote Operation in Connected Vehicle Platoons

S M Sabit Bananee; Shahriar Hasan; Muhammad Mahbub Alam; Nafiul Rashid

cs.NISep 5, 2026

Task-Oriented Semantic Communication for Hazard Warning and Remote Operation in Connected Vehicle Platoons

Authors: S M Sabit Bananee, Shahriar Hasan, Muhammad Mahbub Alam, Nafiul Rashid

Organizations: Department of Computer Science and Engineering, Islamic University of Technology, Gazipur 1704, Bangladesh · Department of Computer Science and Engineering, Mälardalen University, 721 23 Väster˚as, Sweden · Samsung Research America, Mountain View, CA 94043, USA

Abstract

Connected and automated vehicle platoons require reliable dissemination of task-relevant hazard information to enable appropriate downstream responses, with unresolved situations additionally requiring remote human intervention. However, the limited channel bandwidth of V2X communication makes high-volume sensor data transmission challenging, motivating the exchange of only task-relevant information rather than raw sensor data. This paper proposes an integrated semantic and task-oriented communication framework for cooperative hazard response and remote operation in automated vehicle platoons. The framework combines onboard instance segmentation and monocular depth estimation for task-oriented hazard decision-making with Stable Diffusion Variational Autoencoder (SD-VAE)-based semantic scene compression. Task-relevant hazard information and the compact semantic representation are jointly disseminated through a broadcast-based Emergency Semantic Message (ESM) over the existing V2X protocol stack. Evaluation demonstrates that SD-VAE consistently outperforms JPEG and JPEG 2000 on task-relevant reconstruction metrics, exhibiting graceful degradation rather than a sharp cliff effect under low-SINR conditions, demonstrating its feasibility for remote scene reconstruction when human intervention is required. Moreover, an extensive campaign of 7,290 simulation runs shows a low overall inter-vehicle collision rate of 0.99%, with collisions occurring mainly under extreme high-speed and short-gap conditions.

Figures & tables

Fig. 1: System architecture for semantic and task-oriented hazard warning and remote operation.

Fig. 2: End-to-end encoder-decoder architecture of the proposed semantic and task-oriented framework. The platoon LV extracts hazard features and sparse latent semantics, and the receiver reconstructs the scene and performs zone-based response.

Parameter	Values	Count
Platoon size	4, 6, 8 vehicles	3
non-platooning vehicles	50, 100, 200	3
Leader speed	80, 100, 120 km/h	3
CACC algorithm	Rajamani et al. [ 21 ] , Ploeg et al. [ 22 ]	2
Inter-platoon gap	50, 100, 150 m	3
Deceleration rate	$-5$ , $-6$ , $-8$ m/s 2	3

TABLE I: Simulation parameter space (

7{,}290

runs).

Fig. 3: SD-VAE BPP as a function of input resolution, illustrating continuous rate control via resolution-adaptive encoding.

Panel of Figure 5 — Fig. 4: Rate-distortion performance of SD-VAE compared against JPEG and JPEG 2000 across pixel-level ( SSIM, top left ), perceptual ( LPIPS, top right ), and task-relevant ( mIoU, bottom left ; [email protected], bottom right ) reconstruction quality as a function of BPP.

Panel of Figure 6 — Fig. 5: Decoded scene quality of SD-VAE, evaluated across all tested input resolutions (thin lines, 64–320 px, with 160 px highlighted in bold), compared to JPEG and JPEG 2000 for varying SINR (dB).

Class	Box Precision	Box Recall	Box F1 Score	Box mAP50	Mask mAP50
Pedestrian	0.756	0.419	0.539	0.500	0.464
Car	0.784	0.623	0.694	0.673	0.611
Truck	0.603	0.343	0.437	0.409	0.384
Bus	0.702	0.311	0.431	0.363	0.360
Motorcycle	0.638	0.255	0.364	0.325	0.324
Bicycle	0.397	0.062	0.107	0.094	0.079

TABLE II: YOLOv8n-seg validation results on BDD100K-10K.

Fig. 6: Decision path for collision occurrence across all 7,290 simulation runs, identified by a decision tree classifier. Every leaf reaches exactly 0% or 100% collision rate.

Task-Oriented Semantic Communication

Task-oriented semantic communication sends information needed for downstream tasks rather than faithfully reconstructing raw data.

See all 23 papers

Spatiotemporal Semantic V2X Framework for Cooperative Collision Prediction
Jan 23, 2026Murat Arda Onsu, Poonam Lohan, Burak Kantarci +3Intelligent Transportation Systems Semantic Communication
School of Electrical Engineering and Computer Science, University of Ottawa, Ottawa, ON, Canada. · Nokia Bell Labs, 600 March Road, Kanata, ON K2K 2E6, Canada.
SEG-JPEG: Simple Visual Semantic Communications for Remote Operation of Automated Vehicles over Unreliable Wireless Networks
Feb 16, 2026Sebastian Donnelly, Ruth Anderson, George Economides +4Semantic Communication Autonomous Driving
Autonomous Driving and Intelligent Transport Group, Oxford Brookes University, Oxford, UK · Oxfordshire County Council, Oxford, UK · Department for Transport, London, UK +2
SemRD-V2X: Closure-Guided Communication with Bounded Inference for Cooperative Perception
Sep 28, 2026Hu Xu, Chun Li, Siyuan Qiu +2Communication-Efficient Collaborative Perception Rate-Distortion Optimization
Shanghai Jiao Tong University, Shanghai, China

Jan 23, 2026J/K move · Enter open · S save

Spatiotemporal Semantic V2X Framework for Cooperative Collision Prediction

Murat Arda Onsu, Poonam Lohan, Burak Kantarci, Aisha Syed+2

School of Electrical Engineering and Computer Science, University of Ottawa, Ottawa, ON, Canada. · Nokia Bell Labs, 600 March Road, Kanata, ON K2K 2E6, Canada.

Intelligent Transportation Systems Semantic Communication V2X Communication

Open paper AskPDF arXiv

Intelligent Transportation Systems (ITS) demand real-time collision prediction to ensure road safety and reduce accident severity. Conventional approaches rely on transmitting raw video or high-dimensional sensory data from roadside units (RSUs) to vehicles, which is impractical under vehicular communication bandwidth and latency constraints. In this work, we propose a semantic V2X framework in which RSU-mounted cameras generate spatiotemporal semantic embeddings of future frames using the Video Joint Embedding Predictive Architecture (V-JEPA). To evaluate the system, we construct a digital twin of an urban traffic environment enabling the generation of d verse traffic scenarios with both safe and collision events. These embeddings of the future frame, extracted from V-JEPA, capture task-relevant traffic dynamics and are transmitted via V2X links to vehicles, where a lightweight attentive probe and classifier decode them to predict imminent collisions. By transmitting only semantic embeddings instead of raw frames, the proposed system significantly reduces communication overhead while maintaining predictive accuracy. Experimental results demonstrate that the framework with an appropriate processing method achieves a 10% F1-score improvement for collision prediction while reducing transmission requirements by four orders of magnitude compared to raw video. This validates the potential of semantic V2X communication to enable cooperative, real-time collision prediction in ITS.

Abstract

Task-Oriented Semantic Communication

Explore similar work

Spatiotemporal Semantic V2X Framework for Cooperative Collision Prediction

SEG-JPEG: Simple Visual Semantic Communications for Remote Operation of Automated Vehicles over Unreliable Wireless Networks

SemRD-V2X: Closure-Guided Communication with Bounded Inference for Cooperative Perception