Virtual Booth

The rapid growth of wireless devices has increased congestion in multi-access-point (multi-AP) Wi-Fi networks, where conventional client association based on signal strength often leads to uneven load distribution and degraded Quality of Service (QoS). This project proposes an AI-driven wireless network optimization system using OpenWRT, consisting of two Xiaomi CR6608 routers and a centralized Ubuntu-based AI Controller. Real-time network metrics are collected using Netdata and iwinfo and analyzed to dynamically redistribute clients via SSH. Experimental results show significant performance improvement: download throughput increased from 183 Mbps to 240 Mbps (31.1%), latency decreased from 88.0 ms to 23.0 ms (73.9%), and packet loss remained at 0%.