NS3 simulation of a real-time IoT-Edge vision system
● Examined the WiFi network latency associated with an IoT-Edge computing infrastructure for distributed real-time video analytics.
● The novel Edge computing infrastructure enables real-time Deep Learning video analytics by performing object detection/classification, feature extraction, local and global re-identification across many cameras.
● Simulated the IoT-Edge vision system including IoT-camera nodes and edge servers using NS3 discrete event network simulator.
● Constructed a configurable network topology including WiFi STA and CSMA nodes and access points that use 802.11ac standard and supports multiple spatial streams.
● Analysed the throughput and node to server delay of the packet transmission for three different task mapping and resource allocation configurations of the Edge vision system.
This project was done as a part of the paper titled: "A Novel Application/Infrastructure Co-design Approach for Real-time Edge Video Analytics"
Link to the paper: https://ieeexplore.ieee.org/abstract/document/9020639