This paper details the Wireless at Virginia Tech Center for Wireless Telecommunications' (CWT) design and implementation of its Smart Radio (SR) communication platform. The CWT SR can identify available spectrum within a pre-defined band, rendezvous with an intended receiver, and transmit voice and data using a selected quality of service (QoS). This system builds upon previous cognitive technologies developed by CWT for the public safety community, with the goal of providing a prototype mobile communications package for military and public safety First Responders. A master control (MC) enables spectrum awareness by characterizing the radio environment with a power spectrum sensor and an innovative signal detection and classification module. The MC also enables spectrum and signal memory by storing sensor results in a knowledge database. By utilizing a family radio service (FRS) waveform database, the CWT SR can create a new communication link on any designated FRS channel frequency using FM, BPSK, QPSK, or 8PSK modulations. With FM, it supports analog voice communications with legacy hand-held FRS radios. With digital modulations, it supports IP data services, including a CWT developed CVSD-based VoIP protocol. The CWT SR coordinates spectrum sharing between analog primary users and digital secondary users by applying a simple but effective channel-change protocol. It also demonstrates a novel rendezvous protocol to facilitate the detection and initialization of communications links with neighboring SR nodes through the transmission of frequency-hopped rendezvous beacons. By leveraging the GNU Radio toolkit, writing key modules entirely in Python, and utilizing the USRP hardware front-end, the CWT SR provides a dynamic spectrum test bed for future smart and cognitive radio research.© (2008) COPYRIGHT SPIE--The International Society for Optical Engineering. Downloading of the abstract is permitted for personal use only.