Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Task-Oriented Co-Design of Communication, Computing, and Control for Edge-Enabled Industrial Cyber-Physical Systems

Yufeng Diao, Yichi Zhang, Daniele De Martini, Philip Guodong Zhao, Emma Liying Li

TL;DR

This work tackles the challenge of bandwidth-limited, latency-sensitive edge-enabled industrial CPS by proposing a task-oriented co-design that jointly optimizes communication and control. It introduces a variational information bottleneck-based JSCC to preserve task-relevant information while reducing data rate, and couples this with a Delay-Aware Trajectory-Guided Control Prediction (DTCP) framework that fuses trajectory planning and control prediction to mitigate end-to-end delay. The approach is trained via imitation learning and a joint VIB-based objective, and validated in the CARLA simulator where, under an end-to-end delay of $1$ second, it achieves a driving score of $48.12$, about $31.59$ points higher than BPG, while reducing bandwidth by $99.19 ext{%}$. The DTCP framework’ s selective symbol transmission and dynamic fusion of trajectory and control branches enable robust performance across varying channel conditions, highlighting practical benefits for real-time edge-enabled autonomous driving in industrial settings. Potential extensions include more realistic wireless environments (UMi/UMa), dynamically adaptive coding/modulation, and broader applications beyond autonomous driving.

Abstract

This paper proposes a task-oriented co-design framework that integrates communication, computing, and control to address the key challenges of bandwidth limitations, noise interference, and latency in mission-critical industrial Cyber-Physical Systems (CPS). To improve communication efficiency and robustness, we design a task-oriented Joint Source-Channel Coding (JSCC) using Information Bottleneck (IB) to enhance data transmission efficiency by prioritizing task-specific information. To mitigate the perceived End-to-End (E2E) delays, we develop a Delay-Aware Trajectory-Guided Control Prediction (DTCP) strategy that integrates trajectory planning with control prediction, predicting commands based on E2E delay. Moreover, the DTCP is co-designed with task-oriented JSCC, focusing on transmitting task-specific information for timely and reliable autonomous driving. Experimental results in the CARLA simulator demonstrate that, under an E2E delay of 1 second (20 time slots), the proposed framework achieves a driving score of 48.12, which is 31.59 points higher than using Better Portable Graphics (BPG) while reducing bandwidth usage by 99.19%.

Task-Oriented Co-Design of Communication, Computing, and Control for Edge-Enabled Industrial Cyber-Physical Systems

TL;DR

This work tackles the challenge of bandwidth-limited, latency-sensitive edge-enabled industrial CPS by proposing a task-oriented co-design that jointly optimizes communication and control. It introduces a variational information bottleneck-based JSCC to preserve task-relevant information while reducing data rate, and couples this with a Delay-Aware Trajectory-Guided Control Prediction (DTCP) framework that fuses trajectory planning and control prediction to mitigate end-to-end delay. The approach is trained via imitation learning and a joint VIB-based objective, and validated in the CARLA simulator where, under an end-to-end delay of second, it achieves a driving score of , about points higher than BPG, while reducing bandwidth by . The DTCP framework’ s selective symbol transmission and dynamic fusion of trajectory and control branches enable robust performance across varying channel conditions, highlighting practical benefits for real-time edge-enabled autonomous driving in industrial settings. Potential extensions include more realistic wireless environments (UMi/UMa), dynamically adaptive coding/modulation, and broader applications beyond autonomous driving.

Abstract

This paper proposes a task-oriented co-design framework that integrates communication, computing, and control to address the key challenges of bandwidth limitations, noise interference, and latency in mission-critical industrial Cyber-Physical Systems (CPS). To improve communication efficiency and robustness, we design a task-oriented Joint Source-Channel Coding (JSCC) using Information Bottleneck (IB) to enhance data transmission efficiency by prioritizing task-specific information. To mitigate the perceived End-to-End (E2E) delays, we develop a Delay-Aware Trajectory-Guided Control Prediction (DTCP) strategy that integrates trajectory planning with control prediction, predicting commands based on E2E delay. Moreover, the DTCP is co-designed with task-oriented JSCC, focusing on transmitting task-specific information for timely and reliable autonomous driving. Experimental results in the CARLA simulator demonstrate that, under an E2E delay of 1 second (20 time slots), the proposed framework achieves a driving score of 48.12, which is 31.59 points higher than using Better Portable Graphics (BPG) while reducing bandwidth usage by 99.19%.

Paper Structure

This paper contains 27 sections, 1 theorem, 37 equations, 10 figures, 1 table, 2 algorithms.

Table of Contents

  1. Introduction
  2. Related Works
  3. Task-oriented Joint Source-Channel Coding
  4. Edge Inference
  5. Predictions in URLLC Applications
  6. System Model and Problem Formulation
  7. Variational Information Bottleneck Approach
  8. Delay-Aware Trajectory-Guided Control Prediction
  9. Prediction for End-to-End Delay
  10. Trajectory Branch
  11. Control Branch
  12. Two Branch Combination
  13. Loss function
  14. Joint Training
  15. Performance Evaluation
  16. ...and 12 more sections

Key Result

Corollary 1

Assume the shown in fig_DPGM, let $q_{\psi}(\bm{a}|\hat{\bm{z}},\bm{m},\bm{h})$ be a variational approximation of $p(\bm{a}|\hat{\bm{z}},\bm{m},\bm{h})$, let $q(\hat{\bm{z}}) \sim \mathcal{N}(\bm{\mu}_{\hat{\bm{z}}}, \bm{\sigma}^2_{\hat{\bm{z}}}I)$ be a variational approximation of $p(\hat{\bm{z}})$

Figures (10)

  • Figure 1: General framework of edge-enabled autonomous driving.
  • Figure 2: The DPGM for edge-enabled autonomous driving.
  • Figure 3: The proposed task-oriented co-design framework based on JSCC and DTCP.
  • Figure 4: The illustration of a completed cycle of the communication, computing, and control process, along with the prediction structure.
  • Figure 5: The framework of the trajectory and control branch of DTCP.
  • ...and 5 more figures

Theorems & Definitions (1)

  • Corollary 1