3/24/2023 0 Comments Radbeacon dot size![]() ![]() The ground truth is produced by the actors themselves that record their position over a large number of pre-defined signposts on the ground, by using a mobile application that we developed for the EvAAL localization competition. The actors move in the area by following several different scripts that model different use cases, from simple movements to meetings, to both test localization algorithms and detect social relationships among users. Thus, the dataset can be used both in self positioning and in remote positioning scenarios. Each device (either fixed or wearable) acts both as transmitter and receiver of the signal from any other transmitting device. ![]() The installation includes a redundant number of fixed devices installed in each room or public space, and each actor carry two devices (in the hand and on the chest) to represent different kinds of use of wearable devices. The dataset was produced in a portion of our building that includes offices, corridors and public spaces by monitoring up to three human “actors” that act and move in these spaces. The dataset is freely accessible for research purposes without any limitation (the dataset can be freely downloaded from: ). The dataset was built with the purpose of providing an accurate ground truth and a large number of Received Signal Strength Indicator (RSSI) values obtained from both fixed and wearable BLE devices, so that is can be used to test solutions operating with different configurations (e.g., self/remote positioning and direct/indirect positioning) and a large number of use cases, including not only localization and tracking, but also room occupancy and social interactions among users. In particular, the dataset addresses BLE, since this is a low-cost technology available worldwide (all recent smartphones embed this technology), and RSS, since this is a simple and versatile way of measuring distances/proximity, which is also available to all wireless devices. In this work, we propose a novel dataset that addresses all these limits. Finally, almost all datasets (and competitions too) address localization and (to a minor extent) tracking problems or room occupancy, but they hardly provide specific data aimed at detecting social interactions among users. Furthermore, many datasets are produced for the specific aim of evaluating one solution, or they are produced in an exploratory way during events (for example, ) but they often miss an accurate and publicly available ground truth. In particular, datasets providing Received Signal Strength (RSS) data obtained with technologies such as WiFi and RFID are widely available, while only few datasets provide data obtained with more recent technologies such as Bluetooth Low Energy (BLE). However, in most cases, such datasets focus on a specific measurement method and established technologies. For this reason, researchers produce datasets that can be freely used for experimentation, and several competitions on indoor localization also offer the opportunity for researchers to test and compare their solutions, and to produce additional datasets. Since the very early period of research in the field, it became evident that the evaluation and comparison of different indoor localization solutions and technologies require experimentation in real environments with real equipment and users, with ensuing high costs for the setup and maintenance for the experimental testbeds. ![]() Currently, indoor localization is a live and active research area as witnessed by the large number of related publications and events. The need for wide-spread and low-cost solutions for indoor localization has pushed the research towards new technologies, solutions and algorithms. At the same time, raw localization data have also been used to infer information about social interactions among people, and this use of localization data is gaining more and more attention, because of its potential use in online social networks (such as Facebook, Google+ or Twitter), mobile social networking or mobile crowdsensing, where it can contribute to add a physical level of sociality in worlds that, up to now, had only been virtual. In this scenario, indoor localization plays a key role since it is a main enabler for location-based services that deliver services at the right time in the right place and that can thus better assist people in both their ordinary and extra-ordinary activities. The massive introduction of Internet of Things devices on the market-it is expected that a population of over 50 billions of devices will be connected to the Internet by 2020 -and the consequent widespread availability of wearable cyber-physical devices opens new ways of delivering services to people. ![]()
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |