The definition of Intelligent Environments has always been focused around their users, aiming at helping them in a smart and transparent way, and avoiding bothering them or acting against their will. The complexity of IEs, whose technologies range from sensors to machine learning, from distributed architectures to tangible interfaces, from communication protocols to data analysis, challenges researchers from various fields to contribute innovative and effective solutions. In this quest for technical solutions to the myriad requirements of an intelligent environments, user expectations are often left behind, and while researchers tend to focus on niche technical aspects, they risk of losing the big picture of an IE “helping users in their daily life”.
In this paper, we propose a consistency adjustment algorithm for a new localization method based on a client-server model of mobile devices such as indoor mobile robots and smart phones to realize intelligent environments. The distance of movement is calculated using feature points of consecutive images, and subsequently the range of the estimated position is reduced based on the distance. We already have developed an indoor location estimation infrastructure called Universal Map (UMap) using the pre-map. UMap generates a two-dimensional image as a database in advance. The system performs an indoor location estimation by matching the database image and the sensor image. While the maximum error in the previous scheme using only UMap was 76.11 m, the maximum error in this proposed method was reduced to 1.42 m.
Sensor devices are inevitable for realizing intelligent environments by obtaining the physical-world status. Their functions are normally self-contained and merely provide input values to the processing unit. However, sensors are becoming easy to be connected to the Internet and thus our group proposed the notion of SWNW, abstracted sensors utilizing the knowledge at remote servers. In this paper, we describe an enhanced implementation of SWNW using HTTP and an instance of SWNW as HTTP-based Location Sensor using Images (HLSI). HLSI consists of a server maintaining an image database and clients, normally smartphones, capturing images and exchanging information with the server. This paper describes the design and implementation of HLSI.
Voice-activated devices are becoming common place: people can use their voice to control smartphones, smart vacuum robots, and interact with their smart homes through virtual assistant devices like Amazon Echo or Google Home. The spread of such voice-controlled devices is possible thanks to the increasing capabilities of natural language processing, and generally have a positive impact on the device accessibility, e.g., for people with disabilities. However, a consequence of these devices embracing voice control is that people with dysarthria or other speech impairments may be unable to control their intelligent environments, at least with proficiency. This paper investigates to which extent people with dysarthria can use and be understood by the three most common virtual assistants, namely Siri, Google Assistant, and Amazon Alexa. Starting from the sentences in the TORGO database of dysarthric articulation, the differences between such assistants are investigated and discussed. Preliminary results show that the three virtual assistants have comparable performance, with an accuracy of the recognition in the range of 50–60%.
Recently, scientists have shown a special interest in the integration of agent-oriented technologies with the Internet of Things (IoT) in order to manage smarter IoT-objects and their resources. This paper proposes an agent architecture of Multiagent Systems based on Resource-Oriented Architecture (MAS-ROA), where the behavior of each agent is driven by a specific control workflow. This workflow enables agents to be able to perform sensing and control actions over IoT-objects by means of collaborative processes. In this way, IoT-objects managed by MAS-ROA agents can behave proactively, collaboratively, adaptively and smartly. In order to validate the proposal, an IoT ecosystem composed of many “things” (IoT-objects) is modeled and managed by agents based on MAS-ROA. The Multiagent System based on MAS-ROA was then contrasted with an implementation based on Service-Oriented Architecture (MAS-SOA) addressed by Devices Profile for Web Services (DPWS) to get insight about the differences in capabilities and performance.