Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Deep Reinforcement Learning-Based User Scheduling for Collaborative Perception

Yandi Liu, Guowei Liu, Le Liang, Hao Ye, Chongtao Guo, Shi Jin

TL;DR

This work proposes a deep reinforcement learning-based V2X user scheduling algorithm for collaborative perception that reformulates the conventional label-dependent objective into a label-free goal, based on characteristics of 3D object detection, and develops a double deep Q-Network-based user scheduling framework, named SchedCP.

Abstract

Stand-alone perception systems in autonomous driving suffer from limited sensing ranges and occlusions at extended distances, potentially resulting in catastrophic outcomes. To address this issue, collaborative perception is envisioned to improve perceptual accuracy by using vehicle-to-everything (V2X) communication to enable collaboration among connected and autonomous vehicles and roadside units. However, due to limited communication resources, it is impractical for all units to transmit sensing data such as point clouds or high-definition video. As a result, it is essential to optimize the scheduling of communication links to ensure efficient spectrum utilization for the exchange of perceptual data. In this work, we propose a deep reinforcement learning-based V2X user scheduling algorithm for collaborative perception. Given the challenges in acquiring perceptual labels, we reformulate the conventional label-dependent objective into a label-free goal, based on characteristics of 3D object detection. Incorporating both channel state information (CSI) and semantic information, we develop a double deep Q-Network (DDQN)-based user scheduling framework for collaborative perception, named SchedCP. Simulation results verify the effectiveness and robustness of SchedCP compared with traditional V2X scheduling methods. Finally, we present a case study to illustrate how our proposed algorithm adaptively modifies the scheduling decisions by taking both instantaneous CSI and perceptual semantics into account.

Deep Reinforcement Learning-Based User Scheduling for Collaborative Perception

TL;DR

This work proposes a deep reinforcement learning-based V2X user scheduling algorithm for collaborative perception that reformulates the conventional label-dependent objective into a label-free goal, based on characteristics of 3D object detection, and develops a double deep Q-Network-based user scheduling framework, named SchedCP.

Abstract

Stand-alone perception systems in autonomous driving suffer from limited sensing ranges and occlusions at extended distances, potentially resulting in catastrophic outcomes. To address this issue, collaborative perception is envisioned to improve perceptual accuracy by using vehicle-to-everything (V2X) communication to enable collaboration among connected and autonomous vehicles and roadside units. However, due to limited communication resources, it is impractical for all units to transmit sensing data such as point clouds or high-definition video. As a result, it is essential to optimize the scheduling of communication links to ensure efficient spectrum utilization for the exchange of perceptual data. In this work, we propose a deep reinforcement learning-based V2X user scheduling algorithm for collaborative perception. Given the challenges in acquiring perceptual labels, we reformulate the conventional label-dependent objective into a label-free goal, based on characteristics of 3D object detection. Incorporating both channel state information (CSI) and semantic information, we develop a double deep Q-Network (DDQN)-based user scheduling framework for collaborative perception, named SchedCP. Simulation results verify the effectiveness and robustness of SchedCP compared with traditional V2X scheduling methods. Finally, we present a case study to illustrate how our proposed algorithm adaptively modifies the scheduling decisions by taking both instantaneous CSI and perceptual semantics into account.

Paper Structure

This paper contains 14 sections, 27 equations, 13 figures, 5 tables, 1 algorithm.

Table of Contents

  1. Introduction
  2. System Model And Problem Formulation
  3. System Model
  4. Scheduling Problem Formulation
  5. SchedCP: DRL-Based V2X User Scheduling Framework for Collaborative Perception
  6. Feature Selection Based on Spatial Confidence Map
  7. Reinforcement Learning Formulation of User Scheduling
  8. Label-Free Design for User Scheduling
  9. Proposed DDQN-Based Scheduling Algorithm
  10. Simulation Results
  11. Convergence Behavior
  12. Perception Performance Comparison
  13. Case Study: An Occlusion Scenario
  14. Conclusion

Figures (13)

  • Figure 1: An illustrative example of collaborative perception.
  • Figure 2: An illustration of the collaborative perception framework for autonomous driving.
  • Figure 3: Relationship among the sensor sampling interval, scheduling slot and sub-time slot.
  • Figure 4: Relationship between spatial confidence map and detection results.
  • Figure 5: Illustration of the scheduler as an agent interacting with the environment.
  • ...and 8 more figures