Key management in wireless networks is a very challenging problem because of the unreliable transmission medium, the lack of network infrastructure, and stringent energy and complexity limitations at terminals. Recently, there has been increased interest in exploiting physical layer dynamics to generate keys. However, existing schemes exploit only the channel dynamics in the point-to-point channel or basic small wireless systems, but not other rich dynamics inherent to multiuser wireless networks. Furthermore, current studies focus mainly on the pairwise key agreement problem, and did not address issues arising from user interactions in networks. This project develops practical key generation schemes for large scale networks, which 1) fully exploit rich opportunities inherent to network dynamics, and 2) properly address issues arising from interaction among multiple users in networks. Topics including unicast and/or multicast key generation, oblivious relay-assisted key generation, and secret sharing among multiple users are investigated. The goal of this project is to design transmission protocols and algorithms that have engineering impact and practical feasibility. Furthermore, these protocols and algorithms are implemented and tested on a testbed. This project also has strong education components. In addition to providing thesis topics for graduate students, this project provides hands-on lab opportunities for undergraduate education on communication security.