Due to the limitations that falls have on humans, early detection of these becomes essential to avoid further damage. In many applications, various technologies are used to acquire accurate information from individuals such as wearable sensors, environmental sensors or cameras, but all of these require high computational resources in many cases, delaying the response of the entire system. The complexity of the models used to process the input data and detect these activities makes them almost impossible to complete on devices with limited resources, which are the ones that could offer an immediate response avoiding unnecessary communications between sensors and centralized computing centers. In this work, we chose to reduce the models to detect falls using images as input data. We proceeded to use image sequences as video frames, using data from two open source datasets, and we applied the Sparse Low Rank Method to reduce certain layers of the Convolutional Neural Networks that were the backbone of the models. Additionally, we chose to replace a convolutional block with Long Short Term Memory to consider the latest updates of these data sequences. The results showed that performance was maintained decently while significantly reducing the parameter size of the resulting models.

Anticipating pedestrians’ activity is a necessary task for providing a safe and energy efficient environment in an urban area. By locating strategically sensors throughout the city useful information could be obtained. By knowing the average activity of those throughout different days of the week we could identify the typology of the buildings neighboring those sensors. For these type of purposes, clustering methods show great capability forming groups of items that have great similarity intra clusters and disimilarity inter cluster. Different approaches are made to classify the sensors depending on the typology of buildings sorrounding them and the mean pedestrians’ counts for different time intervals. By this way, sensors could be classified in different groups according to their activation patterns and the environment in which they are located through clustering processes and using graph convolutional networks. This study reveals that there is a close relationship between the activity pattern of the pedestrians’ and the type of environment sensors that collect pedestrians’ data are located. By this way, institutions could alleviate a great amount of effort needed to ensure safe and energy efficient urban areas, only knowing the typology of buildings of an urban zone.

Application of deep neural networks (DNN) in edge computing has emerged as a consequence of the need of real time and distributed response of different devices in a large number of scenarios. To this end, shredding these original structures is urgent due to the high number of parameters needed to represent them. As a consequence, the most representative components of different layers are kept in order to maintain the network’s accuracy as close as possible to the entire network’s ones. To do so, two different approaches have been developed in this work. First, the Sparse Low Rank Method (SLR) has been applied to two different Fully Connected (FC) layers to watch their effect on the final response, and the method has been applied to the latest of these layers as a duplicate. On the contrary, SLRProp has been proposed as a variant case, where the relevances of the previous FC layer’s components were weighed as the sum of the products of each of these neurons’ absolute values and the relevances of the neurons from the last FC layer that are connected with the neurons from the previous FC layer. Thus, the relationship of relevances across layer was considered. Experiments have been carried out in well-known architectures to conclude whether the relevances throughout layers have less effect on the final response of the network than the independent relevances intra-layer.

Application of Deep Neural Networks (DNN) in Edge Computing has emerged as a consequence of the need of real time and distributed response of different devices in a large number of scenarios. For this end, shredding these original structures is urgent due to the high number of parameters needed to represent them. As a consequence, the most representative components of different layers are kept in order to maintain the network's accuracy as close as possible from the entire network's ones. To do so, two different approaches have been developed in this work. First, Sparse Low Rank Method (SLR) has been applied to two different Fully Connected (FC) layers to watch their effect on the final response, and similarly the method has been applied to one of these layers. In the contrary, in this last case, the relevances of the previous FC layer's components were considered taking into account the connections to each of the components from the other FC layer, considering the relationship of relevances across layers. Experiments had been carried out in well-known architectures to conclude whether the relevances throughout layers have less effect on the final response of the network than the independent relevances intra-layer.