Next-generation home networks are expected to evolve into multi-technology, multipath mesh networks to support future services and devices with extended coverage and ultra broadband data rates throughout the home area. However this extended architecture is likely to significantly increase management and control complexity, which may become a serious problem for both end-users and operators. End-users will be affected since they are not necessarily network experts, whereas operators will have to manage an ever-expanding network incorporating millions of devices. It is therefore important to redesign such networks in order to embed self-management capabilities. This thesis applies the principles of ìautonomicî systems to provide a complete framework for the distributed, knowledge-based and locally triggered management of home-network equipment. A concrete architectural model for an autonomic home network is defined. Contributions to QoS management, self-healing, supervision and root cause analysis are also presented.