Public Safety Cognitive Radio (PSCR) is a Software Defined Radio(SDR) developed by the Center for Wireless Telecommunications (CWT) at Virginia Tech. PSCR can configure itself to interoperate with any public safety waveform it finds during the scan procedure. It also offers users the capability to scan/classify both analog and digital waveforms. The current PSCR architecture can only run on a general purpose processor and hence is not deployable to the public safety personnel. In the first part of this thesis an Android based control application for the PSCR on a Beagle Board(BB) and the GUI for the control application are developed. The Beagle Board is a low-cost, fanless single board computer that unleashes laptop-like performance and expandability. The Android based Nexus One connected to the Beagle Board via USB is used to control the Beagle Board and enable operations like scan, classify, talk, gateway etc. In addition to the features that exist in the current PSCR a new feature that enables interoperation with P25 (CPFSK modulation) protocol based radios is added. In this effort of porting the PSCR to Beagle Board my contributions are the following (i) communication protocol between the Beagle Board and the Nexus One (ii) PSCR control application on the Android based Nexus One (iii) detection/classification of P25 protocol based radios. In the second part of this thesis, a prototype testbed of a Dynamic Spectrum Access (DSA) broker that uses the Beagle Board PSCR based sensor/classifier is developed. DSA in simple terms is a concept that lets the user without license (secondary user) to a particular frequency access that frequency, when the licensed user (primary user) is not using it. In the proposed testbed we have two Beagle Board based sensor/classifiers that cooperatively scan the spectrum and report the results to the central DSA broker. The DSA broker then identifies the frequency spectrum without primary users and informs the secondary users about the free spectrum. The secondary users can then communicate among each other using the frequency band allocated by the DSA broker. When the primary user enters the spectrum occupied by the secondary user, the DSA broker instructs the secondary user to use a different spectrum. Based on the experiments conducted on the testbed setup in the CWT lab environment, the average time taken by the DSA broker to detect the presence of primary user is 0.636 secs and the average time taken for the secondary user to leave the frequency band that interferes with the primary user is 0.653 secs. / Master of Science