Abstract: A security and automation unit for building, offices or homes includes multiple sensors and multiple resolvers. The resolvers include one or more of a presence resolver, an alarm resolver, a climate resolver, and an event resolver, and at least one resolver includes a neural network. The event resolver includes a self-learning module and at least one of a cognitive module and a proactive module. The self-learning module finds clusters of activation of pieces of equipment at recorded events. The cognitive module analyzes the clusters and modifies an associated scenario for at least one piece of equipment. The proactive module analyzes the clusters and creates at least one new scenario for at least one piece of equipment. The sensors include some of a movement sensor, an acoustic sensor, an image sensor, a light sensor, a CO2 sensor, a thermometer, and a humidity sensor.

Abstract: A voice resolver unit processes audio data. The unit includes a noise detection and cancellation module operative on the audio data, a voice recognition engine to recognize voice in the output of the noise detection and cancellation module, a noise model scorer to match received audio data against at least two noise models to determine the type or types of noise in the audio data and a recognition analyzer to analyze the quality of the output of the recognition engine as a function of the operation of the noise detection and cancellation module according to the type or types of noise detected. The voice resolver may be used in an automation unit for a premises and an alarm system.

Abstract: Various automation units for a premises include a sensor, at least two resolvers, and a plurality of models of the data of the sensor. Each resolver has its associated models which define how the resolver uses the data of the sensor. The various resolvers are an alarm resolver, a presence resolver, a climate resolver and a voice resolver. Other automation units are cognitive or proactive, and include a data base to store at least recorded events per controlled equipment, an event analyzer to analyze recorded events to find patterns of activation of the equipment, and either a cognitive module to modify the associated scenario of a piece of equipment if the patterns do not match the scenario or a proactive module to create a new scenario for a piece of equipment based on the analysis.

Abstract: A smart premises controller unit for processing data received from a plurality of sensors and for controlling and monitoring one or more pieces of equipment in at least one room in a premises responsive to the processing includes an alarm resolver to activate an alarm, a climate resolver to control a climate in the room and a presence resolver to determine at least the presence of a human in the room. Each resolver receives input from a subset of the plurality of sensors and each the resolver comprises a sensor processor and scorer for at least one of its associated subset of sensors. Each resolver comprises a set of models for each of its associated sensors according to its type of resolver. Each sensor processor and scorer matches its sensor data against its models to produce a score for each of its associated models.