ABSTRACT Industry 5.0 demands ultra‐low latency and intelligent collaboration in sports Industrial Internet of Things (IIoT) systems. However, existing cloud‐based or single‐modality approaches suffer from high latency, while many edge‐based methods underutilize multimodal data and adaptive scheduling. This paper proposes an edge‐intelligent sports IIoT framework that integrates multimodal sensor fusion with dynamic edge scheduling to enable real‐time perception and decision‐making. The system jointly optimizes latency, prediction accuracy, and resource utilization through a multi‐objective formulation. Experimental results on a custom sports IIoT dataset and a public health monitoring dataset show that the proposed method reduces end‐to‐end latency by up to 31.7% compared with state‐of‐the‐art methods, while maintaining competitive accuracy (MSE = 0.034) and achieving high edge resource utilization (90.2%). These results demonstrate the effectiveness of collaborative multimodal fusion and edge intelligence for ultra‐low latency sports IIoT applications under Industry 5.0.
Wang et al. (Sun,) studied this question.