Due to significant advances in sensor technology, studies towards activity recognition have gained interest and maturity in the last few years. Existing machine learning algorithms have demonstrated promising results ...
详细信息
Due to significant advances in sensor technology, studies towards activity recognition have gained interest and maturity in the last few years. Existing machine learning algorithms have demonstrated promising results by classifying activities whose instances have been already seen during training. Activity recognition methods based on real-life settings should cover a growing number of activities in various domains, whereby a significant part of instances will not be present in the training data set. However, to cover all possible activities in advance is a complex and expensive task. Concretely, we need a method that can extend the learning model to detect unseen activities without prior knowledge regarding sensor readings about those previously unseen activities. In this paper, we introduce an approach to leverage sensor data in discovering new unseen activities which were not present in the training set. We show that sensor readings can lead to promising results for zero-shot learning, whereby the necessary knowledge can be transferred from seen to unseen activities by using semantic similarity. The evaluation conducted on two data sets extracted from the well-known CASAS datasets show that the proposed zero-shot learning approach achieves a high performance in recognizing unseen (i.e., not present in the training dataset) new activities.
Navigation without Global Navigation Satellite Systems (GNSS) poses a significant challenge in aerospace engineering, particularly in the environments where satellite signals are obstructed or unavailable. This paper ...
详细信息
Navigation without Global Navigation Satellite Systems (GNSS) poses a significant challenge in aerospace engineering, particularly in the environments where satellite signals are obstructed or unavailable. This paper offers an in-depth review of various methods, sensors, and algorithms for Unmanned Aerial Vehicle (UAV) localization in outdoor environments where GNSS signals are unavailable or denied. A key contribution of this study is the establishment of a critical classification system that divides GNSS-denied navigation techniques into two primary categories: absolute and relative localization. This classification enhances the understanding of the strengths and weaknesses of different strategies in various operational contexts. Vision-based localization is identified as the most effective approach in GNSS-denied environments. nonetheless, it's clear that no single-sensor-based localization algorithm can fulfill all the needs of a comprehensive navigation system in outdoor environments. Therefore, it's vital to implement a hybrid strategy that merges various algorithms and sensors for effective outcomes. This detailed analysis emphasizes the challenges and possible solutions for achieving reliable and effective outdoor UAV localization in environments where GNSS is unreliable or unavailable. This multi-faceted analysis, highlights the complexities and potential pathways for achieving efficient and dependable outdoor UAV localization in GNSS-denied environments.
Smart environments refer to buildings or locations equipped with a multitude of sensors and processing mechanisms for improved security, efficiency or functionality. Often, these sensors serve distinct purposes and th...
详细信息
Smart environments refer to buildings or locations equipped with a multitude of sensors and processing mechanisms for improved security, efficiency or functionality. Often, these sensors serve distinct purposes and their data may be processed separately by entirely separate systems. We argue that integrated processing of data available from multiple types of sensors can benefit a variety of decision making processes. For example, smart building sensors such as occupancy or temperature sensors used for lighting or heating efficiency can benefit the security system, or vice versa. Recent industry standards in sensor networks such as ZigBee make it possible to collect and aggregate data from multiple, heterogeneous sensors efficiently. However, integrated information processing with a diverse set of sensor data is still a challenge. We provide an information processing scheme that offers data fusion for multiple sensors such as temperature sensors or motion detectors and visualsensors such as security cameras. The broader goal of multi-sensor data fusion in this context is to enhance security systems, improve energy efficiency by supporting the decision making process based on relevant and accurate information gathered from different sensors. In particular, we investigate a major data fusion technique, Bayesian network, and present a simulation tool for a "smart environment". In addition, we discuss the potential impact of data fusion on the processes of decision or detection, estimation, association, and uncertainty management.
暂无评论