S-HUB: Scalable Deep Neural Network Fusion for Smart Home Hubs

Yuxing Yao, Dong Zhao, Ningcai Xu, Zhengyuan Zhang

Abstract: Smart home hubs have significantly improved everyday home life by serving as central control units that connect and manage various devices, such as lights, door locks, curtains, and cameras. However, the limited CPU and memory resources of these hubs hinder the execution of multiple intelligent tasks, such as human activity recognition, image classification, and speech recognition. To fully utilize these resources, we propose S-HUB, a scalable deep neural network (DNN) fusion framework for smart home hubs capable of handling multiple tasks. In the offline phase, we apply DNN pruning, weight virtualization, and refusion to create a unified model that dynamically scales across tasks. In the online phase, we design a DNN scheduling optimizer to achieve optimal multitask inference while adhering to resource constraints. Finally, comparative experiments and evaluations in smart home scenarios are conducted to assess the performance of S-HUB across various tasks, including speech recognition, object detection, gesture recognition, food recognition, and fall detection. Experimental results show that compared to two state-of-the-art baselines, S-HUB achieves an average task processing time of 3 s (at least 14% faster) under accuracy constraints and an average accuracy loss rate (ALR) of 4.21% (at least 68% lower) under latency constraints. In unconstrained scenarios, it consistently delivers the best overall performance (2.99 s and 2.94% loss), demonstrating the scalability and effectiveness of our fusion method in handling performance–resource tradeoffs across tasks.

External IDs:dblp:journals/iotj/YaoZXZ26