Real-Time Inference Approach Based on Gateway-Centric Edge Computing for Intelligent Services

For providing intelligent services in the network edge, in this paper, we propose a real-time inference approach base on gateway-centric edge computing. The edge computing is deployed on the edge gateway that operates intelligent services in the network entry. The intelligent services are provided based on the inference approaches that are included in the edge gateway. The key component is an intelligent control algorithm based on deep learning. The training of the model is offloaded to the high-performance computing machine, and the resulting inference model is deployed on the edge gateway, which closes the control loop with the IoT device. The inference model is derived by user data and learning model through training the user data on the learning model. Based on deploying the inference model to the edge gateway, the intelligent approach is enabled close to the environment where the data is generated. Therefore, the edge computing architecture reduces the latency to get the control factor for updating the environment based on the real-time sensing data.