Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Joint Uplink and Downlink Rate Splitting for Fog Computing-Enabled Internet of Medical Things

Jiasi Zhou, Yan Chen, Cong Zhou, Yanjing Sun

TL;DR

This work tackles end-to-end latency in fog computing-enabled IoMT by jointly optimizing offloading decisions, edge computing, and result feedback through a joint uplink and downlink RS framework. It introduces a surrogate optimization approach based on the quadratic transform to linearize rate expressions, complemented by first-order Taylor expansions to convexify energy and processing constraints, and solves the resulting problem via a two-tier alternating optimization scheme. A closed-form per-user computation resource allocation is derived, and the BS-side problem is recast into a convex set, enabling scalable optimization. Numerical results show that the RS-based scheme significantly outperforms NOMA/SDMA and SDR baselines, especially as data sizes grow or fog processing is leveraged, highlighting practical gains for real-time IoMT.

Abstract

The Internet of Medical Things (IoMT) facilitates in-home electronic healthcare, transforming traditional hospital-based medical examination approaches. This paper proposes a novel transmit scheme for fog computing-enabled IoMT that leverages uplink and downlink rate splitting (RS). Fog computing allows offloading partial computation tasks to the edge server and processing the remainder of the tasks locally. The uplink RS and downlink RS utilize their flexible interference management capabilities to suppress offloading and feedback delay. Our overarching goal is to minimize the total time cost for task offloading, data processing, and result feedback. The resulting problem requires the joint design of task offloading, computing resource allocation, uplink beamforming, downlink beamforming, and common rate allocation. To solve the formulated non-convex problem, we introduce several auxiliary variables and then construct accurate surrogates to smooth the achievable rate. Moreover, we derive the optimal computation resource allocation per user with closed-form expressions. On this basis, we recast the computing resource allocation and energy consumption at the base station to a convex constraint set. We finally develop an alternating optimization algorithm to update the auxiliary variable and inherent variable alternately. Simulation results show that our transmit scheme and algorithm exhibit considerable performance enhancements over several benchmarks.

Joint Uplink and Downlink Rate Splitting for Fog Computing-Enabled Internet of Medical Things

TL;DR

This work tackles end-to-end latency in fog computing-enabled IoMT by jointly optimizing offloading decisions, edge computing, and result feedback through a joint uplink and downlink RS framework. It introduces a surrogate optimization approach based on the quadratic transform to linearize rate expressions, complemented by first-order Taylor expansions to convexify energy and processing constraints, and solves the resulting problem via a two-tier alternating optimization scheme. A closed-form per-user computation resource allocation is derived, and the BS-side problem is recast into a convex set, enabling scalable optimization. Numerical results show that the RS-based scheme significantly outperforms NOMA/SDMA and SDR baselines, especially as data sizes grow or fog processing is leveraged, highlighting practical gains for real-time IoMT.

Abstract

The Internet of Medical Things (IoMT) facilitates in-home electronic healthcare, transforming traditional hospital-based medical examination approaches. This paper proposes a novel transmit scheme for fog computing-enabled IoMT that leverages uplink and downlink rate splitting (RS). Fog computing allows offloading partial computation tasks to the edge server and processing the remainder of the tasks locally. The uplink RS and downlink RS utilize their flexible interference management capabilities to suppress offloading and feedback delay. Our overarching goal is to minimize the total time cost for task offloading, data processing, and result feedback. The resulting problem requires the joint design of task offloading, computing resource allocation, uplink beamforming, downlink beamforming, and common rate allocation. To solve the formulated non-convex problem, we introduce several auxiliary variables and then construct accurate surrogates to smooth the achievable rate. Moreover, we derive the optimal computation resource allocation per user with closed-form expressions. On this basis, we recast the computing resource allocation and energy consumption at the base station to a convex constraint set. We finally develop an alternating optimization algorithm to update the auxiliary variable and inherent variable alternately. Simulation results show that our transmit scheme and algorithm exhibit considerable performance enhancements over several benchmarks.
Paper Structure (12 sections, 31 equations, 6 figures, 1 table, 1 algorithm)

This paper contains 12 sections, 31 equations, 6 figures, 1 table, 1 algorithm.

Table of Contents

  1. Introduction
  2. Rate splitting (RS) approach
  3. Problem statement and contributions
  4. Related works
  5. System model and problem formulation
  6. Rate splitting-based uplink transmission model
  7. Data processing model
  8. Rate splitting-based downlink transmission model
  9. Problem formulation
  10. Proposed algorithm and properties analysis
  11. Simulation results
  12. Conclusion

Figures (6)

  • Figure 1: An uplink and downlink RS-based transmit model for fog computing-enabled IoMT.
  • Figure 2: (a) The total time cost versus the number of users. (b) The number of infeasible initial points versus the number of users.
  • Figure 3: Total time cost versus maximum data size
  • Figure 4: Total time cost versus computation capacity per user.
  • Figure 5: Total time cost versus transmit power.
  • ...and 1 more figures