An important aspect of all spread spectrum receivers is synchronization. The harsh radio environment has necessitated the development of fast and robust techniques. Current techniques for direct sequence spread spectrum (DSSS) synchronization, although rapid, are not robust to multiple access interference. This is because they require the spreading codes of the multiple access network to have very good cross correlation properties. In this thesis, a novel technique for DSSS synchronization is investigated. This technique, termed as the eigenstructure technique, introduces a new approach to synchronization, called soft synchronization, by which spread spectrum despreading is performed with estimates of the spreading code obtained from the received signal, starting at an early stage. As more data is collected, the code estimate improves and so does the quality of the despread data. The key advantage of the soft synchronization technique is that excellent despread data can be obtained using imperfect, but early, code estimates. The eigenstructure technique exploits the cyclostationarity property of the DSSS signal. It can theoretically provide perfect code estimates for signals received in arbitrary levels of white background noise and with arbitrary timing offsets and is insensitive to frequency offsets on the received signal. It is applicable to DSSS signals with a constant modulus code and requires the knowledge of only the code repetition rate. The technique is robust to multiple access interference provided the users have different code repetition rates. / Master of Science