Abstract: The present specification relates to a method for controlling vehicle interior devices. More specifically, the method comprises the steps of: acquiring image data of a vehicle passenger located within a predetermined distance from a vehicle through an object detection device provided in the exterior of the vehicle; extracting body structure information about the body structure of the vehicle passenger from the image data by using a skeletonization-related deep learning algorithm; and when opening of a specific door of the vehicle is detected, controlling passenger seat-related interior devices corresponding to the specific door on the basis of the extracted body structure information.

Abstract: Disclosed is a method and device for canceling noise by using a deep learning algorithm. The method includes collecting a noise signal, obtaining a first sound signal, which is obtained by extracting only a voice signal from the collected noise signal, and ‘P’ being a probability value indicating that a human voice signal is included in the collected noise signal, through a deep learning algorithm, and on a basis of a value of the ‘P’, outputting the first sound signal or a second sound signal obtained by converting an overall volume of the collected noise signal. At this time, the second sound signal may be a sound signal, of which a reduction ratio of a volume is converted to be great as the volume corresponds to a great portion, from among the collected noise signal.

Abstract: In order to predict power consumption, previous measurements, which indicate the actual amount of power consumed in the past, and errors between previous estimates and the previous measurements, are used as first input data, and power consumption estimates for each prediction technique are simultaneously calculated by using the first input data in at least two prediction techniques. Next, the power consumption estimates calculated by each prediction technique and errors between the power consumption estimates and an actual measurement are used as second input data, and the final power consumption is predicted by making an additional power consumption prediction based on the second input data.

Abstract: Provided is a method for operating a network node in a specific cluster of a tree-structured network based on distributed address assignment. In the method, a non-registered node is detected. A routing depth of the network node is determined. When the determined routing depth corresponds to a reference value, a sub cluster using a cluster address is generated, and the detected non-registered node is registered as a child node in the generated sub cluster using an internal address.

Abstract: Provided is a method for operating a network node of a tree-structured network based on distributed address assignment. Addresses are assigned. It is determined whether a routing depth of the network node corresponds to a reduced node depth. It is determined whether an authority for generating an inferior node exists using the assigned address when the routing depth of the network node corresponds to the reduced node depth. The inferior node is generated using a result of the determination.

Abstract: Provided is a method of preventing a data collision that occurs when two or more child nodes simultaneously transmit data to a parent node in a ZigBee network having a tree routing scheme and operating in a beacon mode. In the ZigBee system having the tree routing scheme and operating in the beacon mode, data communication to the parent node is performed in a beacon frame of the parent node. Here, when several child nodes exist under a single parent node, the child nodes transmit data to the parent node at the moment of receiving the beacon of the parent node, so that the data collision may occur when the child nodes simultaneously transmit the data to the parent node in the beacon frame of the parent node. In the method of preventing a data collision, each of the two or more child nodes is set to wait for its guard time (delay time) when transmitting data to the parent node to induce a normal slotted carrier sense multiple access with collision avoidance (CSMA-CA) operation.

Abstract: A node for self localization, a clustering method using the same, and a localization method are provided. The node, which is located in a specific space so as to constitute a sensor network, includes a location information messaging unit which receives one or more location information messages including information on spatial locations of one or more neighboring nodes in the sensor network from the neighboring nodes in the sensor network; a distance calculator which calculates a first distance to the neighboring node on the basis of the location information included in the received location information messages and calculates a second distance to one or more neighboring nodes on the basis of the received time or intensity of the message on the location information; and a clustering unit which forms clusters of the node and a plurality of neighboring nodes in which the difference between the first and second distances is less than a predetermined threshold.

Abstract: A node for self localization, a clustering method using the same, and a localization method are provided. The node, which is located in a specific space so as to constitute a sensor network, includes a location information messaging unit which receives one or more location information messages including information on spatial locations of one or more neighboring nodes in the sensor network from the neighboring nodes in the sensor network; a distance calculator which calculates a first distance to the neighboring node on the basis of the location information included in the received location information messages and calculates a second distance to one or more neighboring nodes on the basis of the received time or intensity of the message on the location information; and a clustering unit which forms clusters of the node and a plurality of neighboring nodes in which the difference between the first and second distances is less than a predetermined threshold.