Abstract: In one embodiment, a technique for dynamic transmission power in wireless mesh networks using supervised and semi-supervised learning is provided. A first device of a mesh communication network may receive a set of transmission power metrics indicative of network conditions between a second device of the mesh communication network and a plurality of child nodes associated with the second device. The first device may provide the set of transmission power metrics as input to a supervised machine learning process that probabilistically determines one or more minimum transmission power thresholds for nodes of the mesh communication network. The first device may obtain an output from the supervised machine learning process comprising an indication of a particular minimum transmission power threshold for the second device. The first device may control the second device to exchange packets with the plurality of child nodes using the particular minimum transmission power threshold.

Abstract: In one embodiment, a technique for dynamic transmission power in wireless mesh networks using supervised and semi-supervised learning is provided. A first device of a mesh communication network may receive a set of transmission power metrics indicative of network conditions between a second device of the mesh communication network and a plurality of child nodes associated with the second device. The first device may provide the set of transmission power metrics as input to a supervised machine learning process that probabilistically determines one or more minimum transmission power thresholds for nodes of the mesh communication network. The first device may obtain an output from the supervised machine learning process comprising an indication of a particular minimum transmission power threshold for the second device. The first device may control the second device to exchange packets with the plurality of child nodes using the particular minimum transmission power threshold.