TY - JOUR
T1 - Goal-oriented Semantic Communication for Robotic Arm Reconstruction in Digital Twin
T2 - Feature and Temporal Selections
AU - Chen, Shutong
AU - Spyrakos Papastavridis, Emmanouil
AU - Jin, Yichao
AU - Deng, Yansha
N1 - Publisher Copyright:
© 2024 IEEE.
PY - 2025/3/11
Y1 - 2025/3/11
N2 - As one of the most promising technologies in industry, the Digital Twin (DT) facilitates real-time monitoring and predictive analysis for real-world systems by precisely reconstructing virtual replicas of physical entities. However, this reconstruction faces unprecedented challenges due to the ever-increasing communication overhead, especially for digital robotic arm reconstruction. To this end, we propose a novel goal-oriented semantic communication (GSC) design to extract GSC information for the robotic arm reconstruction task in the DT, with the aim of minimising the communication load without sacrificing the reconstruction accuracy. Specifically, rather than transmitting a message that contains complete robotic arm states for reconstruction, we design a Wireless Feature Selection (WFS) algorithm that first segments the motion of the physical robot into several phases, and then extracts and transmits GSC information from this message according to the current phase. Our proposed design is validated through both Pybullet simulations and real-world experiments using the Franka Research 3 robotic arm, where their communication loads are reduced by 44.1% and 35%, respectively, while maintaining the reconstruction error in the same level. The accompanying demo is available online at: https://youtu.be/IAZjoTcbaFA.
AB - As one of the most promising technologies in industry, the Digital Twin (DT) facilitates real-time monitoring and predictive analysis for real-world systems by precisely reconstructing virtual replicas of physical entities. However, this reconstruction faces unprecedented challenges due to the ever-increasing communication overhead, especially for digital robotic arm reconstruction. To this end, we propose a novel goal-oriented semantic communication (GSC) design to extract GSC information for the robotic arm reconstruction task in the DT, with the aim of minimising the communication load without sacrificing the reconstruction accuracy. Specifically, rather than transmitting a message that contains complete robotic arm states for reconstruction, we design a Wireless Feature Selection (WFS) algorithm that first segments the motion of the physical robot into several phases, and then extracts and transmits GSC information from this message according to the current phase. Our proposed design is validated through both Pybullet simulations and real-world experiments using the Franka Research 3 robotic arm, where their communication loads are reduced by 44.1% and 35%, respectively, while maintaining the reconstruction error in the same level. The accompanying demo is available online at: https://youtu.be/IAZjoTcbaFA.
KW - Digital Twin
KW - Semantic Communication
KW - Goal-oriented
KW - Deep Reinforcement Learning (DRL)
KW - Feature Selection
KW - Temporal Selection
KW - Robot Arm
UR - http://www.scopus.com/inward/record.url?scp=105000831453&partnerID=8YFLogxK
U2 - 10.1109/GLOBECOM52923.2024.10901815
DO - 10.1109/GLOBECOM52923.2024.10901815
M3 - Article
SN - 0733-8716
SP - 4830
EP - 4835
JO - IEEE Journal on Selected Areas in Communications
JF - IEEE Journal on Selected Areas in Communications
ER -