Spectrum Agile Radios project studies dynamic channel selection problem upon random encountering of uncontrolled interference in a general wireless network environment. This research targets problem formulation, development of channel selection algorithm, selection of platform and prototyping the proof-of-concept in lab level. Among several candidates, MadWiFi/Ath5K is selected for platform and is studied in-depth to learn its capabilities and limitations. Realtime statistics collection with respect to wireless channel and device buffer is an ongoing work. Algorithms will be designed on top of the statistics collected and economic behavior of users. Automatic Rate Fallback issue of WiFi is discussed to improve the rate adaptation on top of best channel selection. Finally an optimal AP selection algorithm is proposed for corporate type wireless networks.