With the advent of Industry 4.0, localization of materials and factory items will play important roles in factory automation. Since GPS signals are not available in indoor environments, a lot of indoor localization technologies have been proposed based on inertial sensors, audio signals, visible light, wireless signals, etc. In this research, we consider using magnetic fields, which usually exhibit high uniqueness at different locations especially in factories where a lot of stacks, machineries, materials, and metal partitions may coexist. These factors allow us to incorporate deep learning neural networks to learn locationrelated features. Existing works try to collect magnetic field data by human and leverage interpolation to augment dataset. However, our experiments show that the data generated by interpolation is usually different from the ground truth because magnetic fields may not be linear. Therefore, to collect a large enough dataset without human intervention, we dispatch a robot carrying a smartphone to collect dataset at a fine resolution. We use these collected data to train two localization models: deep neural network (DNN) and recurrent neural network (RNN). Besides, we augment our RNN training dataset by combining multiple single-point magnetic values to synthesize fake magnetic trajectories. We conduct field trials, which validate that our approach outperforms previous work.

Localization is the foundation of location-based services and has received considerable attention recently. Pedestrian dead-reckoning (PDR) systems are widely used to infer user positions. However, a main problem is that the errors may easily accumulate during multiple integrations. In this research, we exploit the frequency shift of a wave caused by relative movement, i.e., the Doppler effect, of feet to infer user movement for locating user's location. Specifically, when the two feet of a person move, the relative movements exhibit self-Doppler effect. This observation allows us to implement a PDR system by mounting two speakers and one microphone on shoes. The advantage is that the displacement can be inferred by integration once, as opposed to twice for accelerometers. To investigate the performance of our system, we evaluate its capability in measuring step event, step length, and step direction. Field trial results are presented, which show that our approach performs very well in step detection with 99.6% accuracy and step length estimation with error $\approx 0.0073$ m. However, its direction estimation is still not accurate enough and, thus, needs the help of other sensors.

Given virtual reality (VR)’s popularity, VR already impacts various applications such as education and tourism. In the above applications, users usually need to follow a specific target, such as a teacher or a tour guide, to explore and learn from the environment. However, because of the constraint with visual senses and unintuitive locomotion, following a target in VR is not as easy as doing it in real life. As a result, the user may not concentrate on the audio narration or the surroundings. Therefore, finding a following mechanism that can help the user concentrate and learn in VR is important. In this paper, we focus on the following techniques in VR. We propose four types of following mechanisms: limited (which restricts the user’s ability to move), semi-limited (which constrains the user’s range of activity), improved semi-limited (with visual assistance), and not limited (user’s movement is not constrained). This study consisted of 29 participants divided into four groups and aimed to evaluate how the following mechanisms affect user experience and make users more concentrated in the virtual world. The experiment shows that the semi-limited following mechanism with visual assistance is superior to the other three types in the performance of the touring experience and helping users pay attention to the tour content.

An ultra-wideband (UWB) positioning system consists of at least three anchors and a tag for the positioning procedure. Via the UWB transceivers mounted on all devices in the system, we can obtain the distance information between each pair of devices and further realize the tag localization. However, the uncertain measurement in the real world may introduce incorrect measurement information, e.g., time, distance, positioning, and so on. Therefore, we intend to incorporate the technique of ensemble learning with UWB positioning to improve its performance. In this paper, we present two methods. The experimental results show that our ideas can be applied to different scenarios and work well. Of note, compared with the existing research in the literature, our first algorithm was more accurate and stable. Further, our second algorithm possessed even better performance than the first. Moreover, we also provide a comprehensive discussion for an ill-advised point, which is often used to evaluate the positioning efficiency in the literature.