Search

Sound pollution is an ever growing problem in modern society, and especially in urban environments. In this paper, we investigate if and how artificial sounds can improve the experience of visitors of an urban park with a lot of traffic noise. By using a mobile app, park visitors can control the sound playback by selecting the natural sounds they like, such as birds or a waterfall, and setting the volume. This process of adding artificial sounds to the existing sound environment results in an augmented soundscape. Comparison of the environment with and without this sound playback showed that most visitors experience this as an improvement of the park environment and enjoy controlling the sounds. An experiment with 165 users identified various correlations between the visitors’ subjective evaluations of the sound environment and objective measures of their usage behavior with the app, such as the number of interactions and the spent time.

Recommender systems have been investigated for many years, with the aim of generating the most accurate recommendations possible. However, available data about new users is often insufficient, leading to inaccurate recommendations; an issue that is known as the cold-start problem. A solution can be active learning. Active learning strategies proactively select items and ask users to rate these. This way, detailed user preferences can be acquired and as a result, more accurate recommendations can be offered to the user. In this study, we compare five active learning algorithms, combined with three different predictor algorithms, which are used to estimate to what extent the user would like the item that is asked to rate. In addition, two modes are tested for selecting the items: batch mode (all items at once), and sequential mode (the items one by one). Evaluation of the recommender in terms of rating prediction, decision support, and the ranking of items, showed that sequential mode produces the most accurate recommendations for dense data sets. Differences between the active learning algorithms are small. For most active learners, the best predictor turned out to be FunkSVD in combination with sequential mode., Comment: 11 pages, 12 figures, ORSUM@ACM RecSys 2021, 4th Workshop on Online Recommender Systems and User Modeling, in conjunction with the 15th ACM Conference on Recommender Systems

Road surface state monitoring is of main concern for road infrastructure owners. Hence dedicated measurement campaigns using laser scanning and image analysis are performed on a regular basis. Yet, this type of monitoring comes at a high labor cost and thus it is often limited in coverage and update frequency. This paper proposes opportunistic sensing as an alternative approach. Using sound and vibration sensing in cars that are on the road for other purposes and exploiting the advent of cheap communication, big data, and machine learning, timely information on road state is obtained. Results are compared to laser scanning for spatial frequencies between 0.1 and 100 cycles/m showing the applicability of the method. Results are also used for classification and labeling of road surfaces regarding their effect on rolling noise. Mapping illustrates the coverage of highways and local roads obtained in a few months with as few as seven cars.

User authentication and feedback gathering are crucial aspects for recommender systems. The most common implementations, a username / password login and star rating systems, require user interaction and a cognitive effort from the user. As a result, users opt to save their password in the interface and optional feedback with a star rating system is often skipped, especially for applica- tions such as video watching in a home environment. In this article, we propose an alternative method for user authentication based on facial recognition and an automatic feedback gathering method by detecting various face characteristics. Using facial recognition with a camera in a tablet, smartphone, or smart TV, the persons in front of the screen can be identified in order to link video watch- ing sessions to their user profile. During video watching, implicit feedback is automatically gathered through emotion recognition, attention measurements, and behavior analysis. An emotion finger- print, which is defined as a unique spectrum of expected emotions for a video scene, is compared to the recognized emotions in order to estimate the experience of a user while watching. An evaluation with a test panel showed that happiness can be most accurately detected and the recognized emotions are correlated with the user’s star rating.

Proceedings of the ICA 2019 and EAA Euroregio : 23rd International Congress on Acoustics, integrating 4th EAA Euroregio 2019 : 9-13 September 2019, Aachen, Germany / proceedings editor: Martin Ochmann, Michael Vorländer, Janina Fels 23rd International Congress on Acoustics, integrating 4th EAA Euroregio 2019, ICA 2019, Aachen, Germany, 9 Sep 2019 - 13 Sep 2019; Aachen (2019)., Published by Aachen

Proceedings of the ICA 2019 and EAA Euroregio : 23rd International Congress on Acoustics, integrating 4th EAA Euroregio 2019 : 9-13 September 2019, Aachen, Germany / proceedings editor: Martin Ochmann, Michael Vorländer, Janina Fels 23rd International Congress on Acoustics, integrating 4th EAA Euroregio 2019, ICA 2019, Aachen, Germany, 9 Sep 2019 - 13 Sep 2019; Aachen (2019)., Published by Aachen

Proceedings of the ICA 2019 and EAA Euroregio : 23rd International Congress on Acoustics, integrating 4th EAA Euroregio 2019 : 9-13 September 2019, Aachen, Germany / proceedings editor: Martin Ochmann, Michael Vorländer, Janina Fels 23rd International Congress on Acoustics, integrating 4th EAA Euroregio 2019, ICA 2019, Aachen, Germany, 9 Sep 2019 - 13 Sep 2019; Aachen (2019)., Published by Aachen

As car and truck engines are becoming quieter due to noise emission regulations and new propulsion systems, rolling noise is becoming the dominant contribution of traffic noise. The interaction of tires and pavement causes rolling noise ; thus mitigation is possible in both domains. In Europe, quiet tires are promoted at the EU level, amongst others by careful labelling. Pavement choice and maintenance remains the responsibility of local authorities. Typically, the information available on the acoustic quality of these pavements is scarce. Hence, we designed an opportunistic sound and vibration monitoring approach that allows to monitor pavements continuously. Several cars that drive regularly on the roads are equipped with a low- cost sensor box that collects noise, acceleration, and GPS data. Data analytics of the large datasets thus collected allows to classify and label pavements in a way that is relevant for rolling noise production. The classification method combines a set of carefully chosen sound and vibration features using blind clustering algorithms. Spatial connectivity is added to the clustering to represent the higher probability for similar pavements to be found on adjacent road segments. Action plans based on rolling noise labelling of pavements could become an important traffic noise mitigation approach.