Spread spectrum acoustic communication techniques

An apparatus comprising:

• an encoder configured to:
• store a set of data eigenvectors that are based on a Prometheus Orthonormal Set (PONS) code construction and orthogonal to each other, wherein each of the data eigenvectors is mapped to a unique multi-bit word;
• generate a pilot sequence representing a pilot eigenvector that is based on the PONS code construction and orthogonal to each of the data eigenvectors;
• group input data into multi-bit words and selecting ones of the data eigenvectors mapped to the multi-bit words;
• and
• generate a spread data sequence including the selected ones of the data eigenvectors and that is synchronized to the pilot sequence;
• a modulator configured to generate an acoustic signal including the synchronized pilot sequence and the spread data sequence;
• and
• a loudspeaker configured to transmit the acoustic signal.

A method comprising:

• storing (i) a set of data eigenvectors that are based on a Prometheus Orthonormal Set (PONS) code construction and orthogonal to each other, wherein each of the data eigenvectors is mapped to a unique multi-bit word, and (ii) a replica of a pilot eigenvector that is also based on the PONS code construction and is orthogonal to each of the data eigenvectors;
• receiving an acoustic signal including a pilot sequence representing the pilot eigenvector and at least one of the data eigenvectors synchronized to the pilot sequence;
• detecting the pilot sequence and timing of the pilot sequence using the replica, the detecting including correlating the acoustic signal with the replica of the pilot eigenvector;
• based on the timing of the pilot sequence, identifying a data frame in the acoustic signal that is occupied by the at least one of the data eigenvectors;
• determining which data eigenvector in the set of data eigenvectors is a best match to the at least one of the data eigenvectors in the data frame;
• and
• outputting the multi-bit word that is mapped to the data eigenvector determined to be the best match to the at least one of the data eigenvectors.