Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

A Deep Q-Network based power control mechanism to Minimize RLF driven Handover Failure in 5G Network

Kotha Kartheek, Shankar K. Ghosh, Megha Iyengar, Vinod Sharma, Souvik Deb

TL;DR

This paper tackles RLF-driven handover failure in 5G NR by augmenting conditional handover with a Deep Q-Network based power-control agent. The method defines a 10‑dimension state, two actions (power up or stay), and a reward structure that penalizes RLF and HF while rewarding successful handovers, using a 3×64‑neuron DQN with Double DQN and Huber loss. Training across diverse handover and RLF parameter settings enables robust generalization, and evaluation shows significant reductions in both RLF and HF compared with baselines, including an RL-based dynamic BS approach. The findings demonstrate that adaptive transmit-power adjustments can substantially improve handover robustness in NR, with practical implications for reducing latency and improving reliability in high-mobility or dense deployments.

Abstract

The impact of Radio link failure (RLF) has been largely ignored in designing handover algorithms, although RLF is a major contributor towards causing handover failure (HF). RLF can cause HF if it is detected during an ongoing handover. The objective of this work is to propose an efficient power control mechanism based on Deep Q-Network (DQN), considering handover parameters (i.e., time-to-preparation, time-to-execute, preparation offset, execution offset) and radio link monitoring parameters (T310 and N310) as input. The proposed DRL based power control algorithm decides on a possible increase of transmitting power to avoid RLF driven HF. Simulation results show that the traditional conditional handover, when equipped with the proposed DRL based power control algorithm can significantly reduce both RLFs and subsequent HFs, as compared to the existing state of the art approaches.

A Deep Q-Network based power control mechanism to Minimize RLF driven Handover Failure in 5G Network

TL;DR

This paper tackles RLF-driven handover failure in 5G NR by augmenting conditional handover with a Deep Q-Network based power-control agent. The method defines a 10‑dimension state, two actions (power up or stay), and a reward structure that penalizes RLF and HF while rewarding successful handovers, using a 3×64‑neuron DQN with Double DQN and Huber loss. Training across diverse handover and RLF parameter settings enables robust generalization, and evaluation shows significant reductions in both RLF and HF compared with baselines, including an RL-based dynamic BS approach. The findings demonstrate that adaptive transmit-power adjustments can substantially improve handover robustness in NR, with practical implications for reducing latency and improving reliability in high-mobility or dense deployments.

Abstract

The impact of Radio link failure (RLF) has been largely ignored in designing handover algorithms, although RLF is a major contributor towards causing handover failure (HF). RLF can cause HF if it is detected during an ongoing handover. The objective of this work is to propose an efficient power control mechanism based on Deep Q-Network (DQN), considering handover parameters (i.e., time-to-preparation, time-to-execute, preparation offset, execution offset) and radio link monitoring parameters (T310 and N310) as input. The proposed DRL based power control algorithm decides on a possible increase of transmitting power to avoid RLF driven HF. Simulation results show that the traditional conditional handover, when equipped with the proposed DRL based power control algorithm can significantly reduce both RLFs and subsequent HFs, as compared to the existing state of the art approaches.

Paper Structure

This paper contains 16 sections, 8 equations, 5 figures, 6 tables, 1 algorithm.

Table of Contents

  1. Introduction
  2. Proposed DQN-based Power Control for Handover Failure Minimization
  3. DQN Framework for Dynamic Power Control
  4. State Space Representation
  5. Action space definition
  6. Reward function design
  7. Deep Q-Network Agent Architecture
  8. Neural Network Model
  9. Training enhancements
  10. System aspects
  11. Results and Discussions
  12. Simulation setup
  13. Training the DQN Algorithm
  14. Testing the DQN algorithm
  15. Simulation results
  16. ...and 1 more sections

Figures (5)

  • Figure 1: 2-gNB model demonstrating RLF induced HF.
  • Figure 2: Effect of RSRP averaging window size $N$ on HF Count.
  • Figure 3: Heatmap of RSRP across the 2D gNB deployment. The UE trajectory (from $x=0$ to $3000$ m at $y\approx500$ m) is overlaid.
  • Figure 4: RLF counts for CHO and CHO+DRL across different parameter sweeps
  • Figure 5: Total HF counts under CHO and CHO+DRL across different parameter sweeps.