Urban Air Mobility (UAM) expands roads from the ground to the near-ground space, envisioned as a revolution for congestion alleviation and fast commutes. However, UAM is still in its infancy, encountering the safety flying challenge: high environmental perception requirements for Autonomous Air Vehicles (AAVs) conflict with limited onboard computing and poor communication quality. The primary purpose of this paper is attempted to facilitate the perception of AAVs by exploring collaborative learning through learning model propagation and integration. Specifically, we leverage well-trained perception learning models of local AAVs to boost the onboard training of new entrants by Knowledge Distillation (KD) technology. Therefore, a knowledge distillation-based model propagation protocol is proposed to obtain the well-trained learning models of nearby AAVs. This protocol also takes transmission errors between AAVs into account. Moreover, we develop an error repair scheme to reduce the upper bound delay of model propagation. Simulation results verify the efficiency of the proposed scheme to improve the AAV onboard learning performance whenever it encounters new environments.


    Access

    Check access

    Check availability in my library

    Order at Subito €


    Export, share and cite



    Title :

    Knowledge Distillation-Based Learning Model Propagation for Urban Air Mobility


    Contributors:
    Xiong, Kai (author) / Xie, Juefei (author) / Wang, Zhihong (author) / Leng, Supeng (author)


    Publication date :

    2024-06-24


    Size :

    927879 byte





    Type of media :

    Conference paper


    Type of material :

    Electronic Resource


    Language :

    English




    Generative Dataset Distillation Based on Self-knowledge Distillation

    Li, Longzhen / Li, Guang / Togo, Ren et al. | ArXiv | 2025

    Free access

    Urban traffic flow prediction method and system based on graph knowledge distillation

    DU HANGYUAN / YU RONG / WANG WENJIAN et al. | European Patent Office | 2024

    Free access

    KNOWLEDGE DISTILLATION TECHNIQUES

    MALACH ERAN / KAPLUN GAL / SHALEV-SHWARTZ SHAI | European Patent Office | 2023

    Free access

    Decoupled Knowledge Distillation

    Zhao, Borui / Cui, Quan / Song, Renjie et al. | ArXiv | 2022

    Free access