Audio QR technology can be used by many at a time unlike image QR and does not require any extra hardware. The transmitter transmits continuously frequency modulated sound wave and receiver can pick up these waves anytime without asking for request from transmitter. This system will use dual tone FSK for modulation. Both the transmitter and receiver will be implemented on TMS320C5515 eZDSP board from Texas Instruments. At the transmitter the string to be transmitted is transmitted using 2-Tone FSK character by character. For each character 2 frequencies are assigned. These 2 frequencies are chosen from a pool of 16 different frequencies in the range 15KHz to 20KHz. The frequencies were so chosen as they correspond to the frequencies that are represented by bins in the 256 point DFT at output. We set the sampling frequency at the transmitter and receiver to 48KHz. The sinusoids were generated in MATLAB and saved in the form of 16x256 matrix in c code. The receiver is implementing FFT on the received samples. The output of FFT helps to detect the transmitted sinusoids. On receiver hardware, ping-pong buffer is used. The frame of ping is processed, while pong is being filled. First, FFT of the frame is found using hardware FFT. Then, scaling of FFT is done to bring down the values in between -128 to 128. The absolute value of the FFT is found by squaring and adding the real and imaginary parts. The location of highest and second highest peaks are found by traversing the FFT array. The locations are stored as highest location and second highest location. A variable which is equal to sum of squares of there variables is used to identify the character that was transmitted. The character matrix which was made at the transmitter is also made at the receiver along with the sum of squares of the locations of the highest and second highest peaks.
Shrihari Deshmukh
Comments