Secure Symbol-Level Miso Precoding

While constructive interference offers indirect advantages in physical layer security by reducing the transmit power required to achieve a desired performance level, additional gains are possible by choosing the symbols to degrade the eavesdropper's ability to decode the desired data. An algorithm was recently proposed for this purpose, but it assumes an eavesdropper that employs simple nearest-neighbor decoding, and that only exploits a portion of the space available for destructive interference (DI). In this paper, we modify the technique to exploit the full DI region, but we show that even with this improvement, the general approach is vulnerable to an intelligent eavesdropper who can perform maximum likelihood detection. Based on this observation, we propose an alternative approach that, while requiring increased transmit power, can yield the best possible security.