No abstract

Abstract-We believe that robots will take an active part in our daily lives in the near future. Once this time arrives, robots will be expected to have the social skills necessary for interacting with people in addition to the ability to carry out their own tasks. We call an interaction that increases familiarity and makes communication smoother a "social interaction." We have recently developed a human-size humanoid, called "Robovie-IV," which will have the ability to interact with people in their daily lives. This paper discusses the design requirements for humanrobot interaction by a communication robot, then introduces an overview of Robovie-IV's hardware and software architectures. There are also preliminary results of a daily experiment that we are currently conducting with Robovie-IV in our office.

Three-dimensional user interfaces (3D UIs) let users interact with virtual objects, environments, or information using direct 3D input in the physical and/or virtual space. In this article, the founders and organizers of the IEEE Symposium on 3D User Interfaces reflect on the state of the art in several key aspects of 3D UIs and speculate on future research.

An ideal augmented reality (AR) display for multi-user co-located collaboration should have following three features: 1) Any virtual object should be able to be shown at any arbitrary position, e.g. a user can see a virtual object in front of other users' faces. 2) Correct occlusion of virtual and real objects should be supported.3) The real world should be naturally and clearly visible, which is important for face-to-face conversation. We have been developing an optical see-through display, ELMO (Enhanced see-through display using an LCD panel for Mutual Occlusion), that satisfies these three requirements. While previous prototype systems were not practical due to their size and weight, we have come up with an improved optics design which has reduced size and is lightweight enough to wear. In this paper, the characteristics of typical multi-user three-dimensional displays are summarized and the design details of the latest optics are then described. Finally, a collaborative AR application employing the new display and its user experience are explained.

Abstract:In recent years, virtual and augmented reality have begun to take advantage of the high speed capabilities of data streaming technologies and wireless networks. However, limitations like bandwidth and latency still prevent us from achieving high fidelity telepresence and collaborative virtual and augmented reality applications. Fortunately, both researchers and engineers are aware of these problems and have set out to design 5G networks to help us to move to the next generation of virtual interfaces. This paper reviews state of the art virtual and augmented reality communications technology and outlines current efforts to design an effective, ubiquitous 5G network to help to adapt to virtual application demands. We discuss application needs in domains like telepresence, education, healthcare, streaming media, and haptics, and provide guidelines and future directions for growth based on this new network infrastructure.

Abstract. Augmented Reality (AR) is a technology that allows users to view and interact in real time with virtual images seamlessly superimposed over the real world. AR systems can be used to create unique collaborative experiences. For example, co-located users can see shared 3D virtual objects that they interact with, or a user can annotate the live video view of a remote worker, enabling them to collaborate at a distance. The overall goal is to augment the face-to-face collaborative experience, or to enable remote people to feel that they are virtually co-located. In this special issue on collaboration in augmented reality, we begin with the visions of science fiction authors of future technologies that might significantly improve collaboration, then introduce research articles which describe progress towards these visions, finally we outline a research agenda discussing the work still to be done.

We conducted two experiments comparing communication behaviors of co-located users in collaborative augmented reality (AR) interfaces. In the first experiment, we compared optical, stereo-and mono-video, and immersive head mounted displays (HMDs) using a target identification task. It was found that differences in the real world visibility severely affect communication behaviors. The optical see-through case produced the best results with the least extra communication needed. Generally, the more difficult it was to use non-verbal communication cues, the more people resorted to speech cues to compensate. In the second experiment, we compared three different combinations of task and communication spaces using a 2D icon designing task with optical see-through HMDs. It was found that the spatial relationship between the task and communication spaces also severely affected communication behaviors. Placing the task space between the subjects produced the most active behaviors in terms of initiatory body languages and utterances with least miscommunications.

In this article, a gait recognition algorithm is presented based on the information obtained from inertial sensors embedded in a smartphone, in particular, the accelerometers and gyroscopes typically embedded on them. The algorithm processes the signal by extracting gait cycles, which are then fed into a Recurrent Neural Network (RNN) to generate feature vectors. To optimize the accuracy of this algorithm, we apply a random grid hyperparameter selection process followed by a hand-tuning method to reach the final hyperparameter configuration. The different configurations are tested on a public database with 744 users and compared with other algorithms that were previously tested on the same database. After reaching the best-performing configuration for our algorithm, we obtain an equal error rate (EER) of 11.48% when training with only 20% of the users. Even better, when using 70% of the users for training, that value drops to 7.55%. The system manages to improve on state-of-the-art methods, but we believe the algorithm could reach a significantly better performance if it was trained with more visits per user. With a large enough database with several visits per user, the algorithm could improve substantially.