Showing papers by "Nokia published in 2015"

Science and technology roadmap for graphene, related two-dimensional crystals, and hybrid systems

Abstract: We present the science and technology roadmap for graphene, related two-dimensional crystals, and hybrid systems, targeting an evolution in technology, that might lead to impacts and benefits reaching into most areas of society. This roadmap was developed within the framework of the European Graphene Flagship and outlines the main targets and research areas as best understood at the start of this ambitious project. We provide an overview of the key aspects of graphene and related materials (GRMs), ranging from fundamental research challenges to a variety of applications in a large number of sectors, highlighting the steps necessary to take GRMs from a state of raw potential to a point where they might revolutionize multiple industries. We also define an extensive list of acronyms in an effort to standardize the nomenclature in this emerging field.

Semi-supervised learning with Ladder networks

Abstract: We combine supervised learning with unsupervised learning in deep neural networks. The proposed model is trained to simultaneously minimize the sum of supervised and unsupervised cost functions by backpropagation, avoiding the need for layer-wise pre-training. Our work builds on top of the Ladder network proposed by Valpola [1] which we extend by combining the model with supervision. We show that the resulting model reaches state-of-the-art performance in semi-supervised MNIST and CIFAR-10 classification in addition to permutation-invariant MNIST classification with all labels.

Universal linear optics

Abstract: Linear optics underpins fundamental tests of quantum mechanics and quantum technologies. We demonstrate a single reprogrammable optical circuit that is sufficient to implement all possible linear optical protocols up to the size of that circuit. Our six-mode universal system consists of a cascade of 15 Mach-Zehnder interferometers with 30 thermo-optic phase shifters integrated into a single photonic chip that is electrically and optically interfaced for arbitrary setting of all phase shifters, input of up to six photons, and their measurement with a 12-single-photon detector system. We programmed this system to implement heralded quantum logic and entangling gates, boson sampling with verification tests, and six-dimensional complex Hadamards. We implemented 100 Haar random unitaries with an average fidelity of 0.999 ± 0.001. Our system can be rapidly reprogrammed to implement these and any other linear optical protocol, pointing the way to applications across fundamental science and quantum technologies.

On the coverage of LPWANs: range evaluation and channel attenuation model for LoRa technology

Abstract: In addition to long battery life and low cost, coverage is one of the most critical performance metrics for the low power wide area networks (LPWAN). In this work we study the coverage of the recently developed LoRa LPWAN technology via real-life measurements. The experiments were conducted in the city of Oulu, Finland, using the commercially available equipment. The measurements were executed for cases when a node located on ground (attached on the roof rack of a car) or on water (attached to the radio mast of a boat) reporting their data to a base station. For a node operating in the 868 MHz ISM band using 14 dBm transmit power and the maximum spreading factor, we have observed the maximum communication range of over 15 km on ground and close to 30 km on water. Besides the actual measurements, in the paper we also present a channel attenuation model derived from the measurement data. The model can be used to estimate the path loss in 868 MHz ISM band in an area similar to Oulu, Finland.

Layered memristive and memcapacitive switches for printable electronics

Abstract: Memristors promise to emulate the appealing characteristics of biological neural systems. Solution-processed heterostructures are now shown to behave as memristive and memcapacitive switches compatible with printed electronics applications.

Opportunities and Challenges for Data Physicalization

Abstract: Physical representations of data have existed for thousands of years. Yet it is now that advances in digital fabrication, actuated tangible interfaces, and shape-changing displays are spurring an emerging area of research that we call Data Physicalization. It aims to help people explore, understand, and communicate data using computer-supported physical data representations. We call these representations physicalizations, analogously to visualizations -- their purely visual counterpart. In this article, we go beyond the focused research questions addressed so far by delineating the research area, synthesizing its open challenges and laying out a research agenda.

Heuristic Approaches to the Controller Placement Problem in Large Scale SDN Networks

Abstract: Software Defined Networking (SDN) marks a paradigm shift towards an externalized and logically centralized network control plane. A particularly important task in SDN architectures is that of controller placement, i.e., the positioning of a limited number of resources within a network to meet various requirements. These requirements range from latency constraints to failure tolerance and load balancing. In most scenarios, at least some of these objectives are competing, thus no single best placement is available and decision makers need to find a balanced trade-off. This work presents POCO, a framework for Pareto-based Optimal COntroller placement that provides operators with Pareto optimal placements with respect to different performance metrics. In its default configuration, POCO performs an exhaustive evaluation of all possible placements. While this is practically feasible for small and medium sized networks, realistic time and resource constraints call for an alternative in the context of large scale networks or dynamic networks whose properties change over time. For these scenarios, the POCO toolset is extended by a heuristic approach that is less accurate, but yields faster computation times. An evaluation of this heuristic is performed on a collection of real world network topologies from the Internet Topology Zoo. Utilizing a measure for quantifying the error introduced by the heuristic approach allows an analysis of the resulting trade-off between time and accuracy. Additionally, the proposed methods can be extended to solve similar virtual functions placement problems which appear in the context of Network Functions Virtualization (NFV).

Why we use and abandon smart devices

Abstract: Smart devices are becoming increasingly commercially available. However, uptake of these devices has been slow and abandonment swift, which indicates that smart devices may not currently meet the needs of users. To advance an understanding of the ways users benefit from, are challenged by, and abandon smart devices, we asked a group of users to purchase smart sensing devices to advance themselves towards a personal, self-defined goal. We found that participants abandoned devices because they did not fit with the their conceptions of themselves, the data collected by devices were perceived to not be useful, and device maintenance became unmanageable. Participants used devices because they had developed routines and because devices were useful, satisfied curiosity, and held hope for potential benefit to them. We propose ways to reduce barriers, motivate use, and argue for envisioning an additional function of these devices for short-term interventions, in addition to standard long-term use.

Semi-Supervised Learning with Ladder Networks

Abstract: We combine supervised learning with unsupervised learning in deep neural networks. The proposed model is trained to simultaneously minimize the sum of supervised and unsupervised cost functions by backpropagation, avoiding the need for layer-wise pre-training. Our work builds on the Ladder network proposed by Valpola (2015), which we extend by combining the model with supervision. We show that the resulting model reaches state-of-the-art performance in semi-supervised MNIST and CIFAR-10 classification, in addition to permutation-invariant MNIST classification with all labels.

CID2013: A Database for Evaluating No-Reference Image Quality Assessment Algorithms

Abstract: This paper presents a new database, CID2013, to address the issue of using no-reference (NR) image quality assessment algorithms on images with multiple distortions. Current NR algorithms struggle to handle images with many concurrent distortion types, such as real photographic images captured by different digital cameras. The database consists of six image sets; on average, 30 subjects have evaluated 12–14 devices depicting eight different scenes for a total of 79 different cameras, 480 images, and 188 subjects (67% female). The subjective evaluation method was a hybrid absolute category rating-pair comparison developed for the study and presented in this paper. This method utilizes a slideshow of all images within a scene to allow the test images to work as references to each other. In addition to mean opinion score value, the images are also rated using sharpness, graininess, lightness, and color saturation scales. The CID2013 database contains images used in the experiments with the full subjective data plus extensive background information from the subjects. The database is made freely available for the research community.

Recent advancements in M2M communications in 4G networks and evolution towards 5G

Abstract: Machine-to-Machine (M2M) communications is considered to be one of the key enablers for the provisioning of advanced applications and services such as smart cities and hospitals, as well as automated vehicular and industrial automation operation. Currently, in LTE-Advanced systems, the main focus has been on supporting massive deployment of low cost devices, with enhanced radio access network coverage. In this work, we present the recent MTC enhancements in LTE system. Detailed performance analysis based on LTE system settings is also presented. The LTE capacity evaluations performed based on devices per physical resource block indicate that significantly large number of devices can already be supported in an LTE system, based on the assumptions used, with minimal system overhead. We also present an overview of some of the key scenarios, requirements and use cases currently being considered for M2M communication in fifth generation (5G) systems. The performance requirements currently being considered for massive and ultra-reliable M2M communication are also discussed.

An open ecosystem for mobile-cloud convergence

Abstract: We show how a disruptive force in mobile computing can be created by extending today’s unmodified cloud to a second level consisting of self-managed data centers with no hard state called cloudlets. These are located at the edge of the Internet, just one wireless hop away from associated mobile devices. By leveraging lowlatency offload, cloudlets enable a new class of real-time cognitive assistive applications on wearable devices. By processing high data rate sensor inputs such as video close to the point of capture, cloudlets can reduce ingress bandwidth demand into the cloud. By serving as proxies for distant cloud services that are unavailable due to failures or cyberattacks, cloudlets can improve robustness and availability. We caution that proprietary software ecosytems surrounding cloudlets will lead to a fragmented marketplace that fails to realize the full business potential of mobile-cloud convergence. Instead, we urge that the software ecosystem surrounding cloudlets be based on the same principles of openness and end-to-end design that have made the Internet so successful.

Method, apparatus and computer program product for input detection

Abstract: One or more operating conditions of a device such as a mobile phone are determined (802). A selection of a mode of operation of the device from at least a first mode and a second mode is facilitated (806, 808) based on the one or more operating conditions of the device such as the time delay between times of arrival of the audio signals (804). In the first mode; the device is configured to detect an operation input received from an audio source based on two or more audio sensors of the device. In the second mode, the device is configured to detect the operation input based on at least one of the two or more audio sensors and at least one non-audio sensor of the device.

A Diary Study on Combining Multiple Information Devices in Everyday Activities and Tasks

Abstract: As people possess increasing numbers of information devices, situations where several devices are combined and used together have become more common. We present a user study on people's current practices in combining multiple information devices in their everyday lives, ranging from pragmatic tasks to leisure activities. Based on diaries and interviews of 14 participants, we characterize the usage practices of the most common devices, including smartphones, computers, tablets, and home media centers. We analyze 123 real-life multi-device use cases and identify the main usage patterns, including Sequential Use, Resource Lending, Related Parallel Use, and Unrelated Parallel Use. We discuss the practical challenges of using several information devices together. Finally, we identify three levels of decisions that determine which devices are used in a particular situation, including acquiring, making available, and selecting the devices for use.

Awareness and determinants of electronic cigarette use among Finnish adolescents in 2013: a population-based study

Abstract: Background A wide range of electronic cigarettes (e-cigarettes) are now on the market. We studied e-cigarette awareness and use, determinants and sources of e-cigarettes, the e-liquids used in them and exposure to e-cigarette advertisements among adolescents in Finland. Among smokers, we studied the association of e-cigarette use and interest in smoking cessation. Method Data were obtained from a national survey of 12–18-year-old Finnish adolescents in 2013 (N=3535, response rate 38%). Descriptive statistics and logistic regression analysis were used. Results Of the respondents, 85.3% knew what ecigarettes were; 17.4% had tried them. E-liquids with nicotine were used most often (65.7%); also those who had never tried conventional cigarettes had used them. Of e-cigarette ever users, 8.3% had never tried smoking. Parents’ high level of education, being in employment, and intact family protected against children’s e-cigarette use. In the final model, daily smoking (OR 41.35; 95% CI 25.2 to 67.8), snus use (2.96; 2.4–4.0), waterpipe use (2.21; 1.6–3.0), children’s vocational education (2.06; 1.4–3.1) and poor school performance (1.92; 1.4–3.0) were associated with e-cigarette experimentation. Those smokers with most experience of e-cigarettes were least likely to consider smoking cessation. Conclusions Awareness and experimentation with e-cigarettes are high among adolescents, especially in older age groups and boys. Nicotine e-liquids are easy to acquire for youth. Having similar risk factors, e-cigarette use seems to follow the model of conventional smoking initiation. Among adolescent smokers, use of e-cigarettes does not clearly relate to interest in smoking cessation. Preventive policies are needed to protect the youth.

pocl: A Performance-Portable OpenCL Implementation

Abstract: OpenCL is a standard for parallel programming of heterogeneous systems The benefits of a common programming standard are clear; multiple vendors can provide support for application descriptions written according to the standard, thus reducing the program porting effort While the standard brings the obvious benefits of platform portability, the performance portability aspects are largely left to the programmer The situation is made worse due to multiple proprietary vendor implementations with different characteristics, and, thus, required optimization strategies In this paper, we propose an OpenCL implementation that is both portable and performance portable At its core is a kernel compiler that can be used to exploit the data parallelism of OpenCL programs on multiple platforms with different parallel hardware styles The kernel compiler is modularized to perform target-independent parallel region formation separately from the target-specific parallel mapping of the regions to enable support for various styles of fine-grained parallel resources such as subword SIMD extensions, SIMD datapaths and static multi-issue Unlike previous similar techniques that work on the source level, the parallel region formation retains the information of the data parallelism using the LLVM IR and its metadata infrastructure This data can be exploited by the later generic compiler passes for efficient parallelization The proposed open source implementation of OpenCL is also platform portable, enabling OpenCL on a wide range of architectures, both already commercialized and on those that are still under research The paper describes how the portability of the implementation is achieved We test the two aspects to portability by utilizing the kernel compiler and the OpenCL implementation to run OpenCL applications in various platforms with different style of parallel resources The results show that most of the benchmarked applications when compiled using pocl were faster or close to as fast as the best proprietary OpenCL implementation for the platform at hand

Characterization of Quasi-Stationarity Regions for Vehicle-to-Vehicle Radio Channels

Abstract: We analyze the nonwide-sense-stationarity (non-WSS) of vehicle-to-vehicle (V2V) radio channels using three metrics: 1) the correlation matrix distance (CMD); 2) the wideband spectral divergence (SD); and 3) the shadow fading correlation. The analysis is based on measurements carried out at 5.3 GHz using a ${\mathbf 30} \times {\mathbf 4}$ MIMO system in suburban, urban, and underground parking areas. Several factors such as the existence of a line-of-sight (LOS), the speed of cars, and the antenna array size and configuration are considered in the analysis of non-WSS. It is found that quasi-stationarity region ranges from 3 to 80 m in different V2V scenarios, and is strongly affected by the above factors. Based on the comparison of the equivalent quasi-stationarity region size estimated by the three metrics, it is suggested to use SD and shadowing correlation metrics for systems with small electrical array apertures and to use CMD metric only for arrays with large electrical apertures.

Bidirectional recurrent neural networks as generative models

Abstract: Bidirectional recurrent neural networks (RNN) are trained to predict both in the positive and negative time directions simultaneously. They have not been used commonly in unsupervised tasks, because a probabilistic interpretation of the model has been difficult. Recently, two different frameworks, GSN and NADE, provide a connection between reconstruction and probabilistic modeling, which makes the interpretation possible. As far as we know, neither GSN or NADE have been studied in the context of time series before. As an example of an un-supervised task, we study the problem of filling in gaps in high-dimensional time series with complex dynamics. Although unidirectional RNNs have recently been trained successfully to model such time series, inference in the negative time direction is non-trivial. We propose two probabilistic interpretations of bidirectional RNNs that can be used to reconstruct missing gaps efficiently. Our experiments on text data show that both proposed methods are much more accurate than unidirectional reconstructions, although a bit less accurate than a computationally complex bidirectional Bayesian inference on the unidirectional RNN. We also provide results on music data for which the Bayesian inference is computationally infeasible, demonstrating the scalability of the proposed methods.

Overview of the EVS codec architecture

Abstract: The recently standardized 3GPP codec for Enhanced Voice Services (EVS) offers new features and improvements for low-delay real-time communication systems. Based on a novel, switched low-delay speech/audio codec, the EVS codec contains various tools for better compression efficiency and higher quality for clean/noisy speech, mixed content and music, including support for wideband, super-wideband and full-band content. The EVS codec operates in a broad range of bitrates, is highly robust against packet loss and provides an AMR-WB interoperable mode for compatibility with existing systems. This paper gives an overview of the underlying architecture as well as the novel technologies in the EVS codec and presents listening test results showing the performance of the new codec in terms of compression and speech/audio quality.

Method, apparatus and computer program product for joining the displays of multiple devices

Abstract: A method is provided for enhancing a user interface with a mobile device by joining the displays of multiple devices together to function together with one another. In particular, example methods may provide for detecting a first touch or drag event, receiving a message indicating a second touch or drag event from a nearby device occurring during the first touch or drag event, creating a join event, and providing for transmission of the join event to the nearby device. The method may also include updating a registry of nearby devices in response to receiving a message regarding a touch or drag event from a nearby device. The method may further include providing for presentation of content on a display of a device and direction of content to be presented on a display of the nearby device, wherein the content to be presented on the display of the nearby device is related to the content on the display of the device.

Path Loss, Shadow Fading, and Line-of-Sight Probability Models for 5G Urban Macro-Cellular Scenarios

Abstract: This paper presents key parameters including the line-of-sight (LOS) probability, large-scale path loss, and shadow fading models for the design of future fifth generation (5G) wireless communication systems in urban macro- cellular (UMa) scenarios, using the data obtained from propagation measurements at 38 GHz in Austin, US, and at 2, 10, 18, and 28 GHz in Aalborg, Denmark. A comparison of different LOS probability models is performed for the Aalborg environment. Alpha- betagamma and close-in reference distance path loss models are studied in depth to show their value in channel modeling. Additionally, both single-slope and dual-slope omnidirectional path loss models are investigated to analyze and contrast their root-mean-square (RMS) errors on measured path loss values. While the results show that the dual-slope large-scale path loss model can slightly reduce RMS errors compared to its singleslope counterpart in non-line-of-sight (NLOS) conditions, the improvement is not significant enough to warrant adopting the dual- slope path loss model. Furthermore, the shadow fading magnitude versus distance is explored, showing a slight increasing trend in LOS and a decreasing trend in NLOS based on the Aalborg data, but more measurements are necessary to gain a better knowledge of the UMa channels at centimeter- and millimeter-wave frequency bands.

Full-duplex self-backhauling for small-cell 5G networks

Abstract: We consider in-band self-backhauling for small cell 5G systems. In-band self-backhauling enables efficient usage of frequency resources. When coupled with a flexible frame format, it also enables efficient time-division duplexing of uplink, downlink, and backhaul transmissions. Self-backhauling is particularly efficient when coupled with FD relaying. Antenna design, as well as cancellation in radio frequency and digital domains at an FD relay enables reuse of the same resources for backhaul and access hops. The use of radio resources in the self-backhauling and access hops can be coordinated to maximize end-to-end performance. We evaluate FD in-band self-backhauling in indoor 5G scenarios, targeting mobile broadband and ultrareliable communication use cases. Self-backhauling shows considerable promise for reaching 5G targets in these scenarios.

Objection recognition in a 3d scene

Abstract: A method comprising: obtaining a three-dimensional (3D) point cloud about at least one object of interest; detecting ground and/or building objects from 3D point cloud data using an unsupervised segmentation method; removing the ground and/or building objects from the 3D point cloud data; and detecting one or more vertical objects from the remaining 3D point cloud data using a supervised segmentation method.

Divide and conquer: Partitioning OSPF networks with SDN

Abstract: Software Defined Networking (SDN) is an emerging network control paradigm focused on logical centralization and programmability. At the same time, distributed routing protocols, most notably OSPF and IS-IS, are still prevalent in IP networks, as they provide shortest path routing, fast topological convergence after network failures, and, perhaps most importantly, the confidence based on decades of reliable operation. Therefore, a hybrid SDN/OSPF operation remains a desirable proposition. In this paper, we propose a new method of hybrid SDN/OSPF operation. Our method is different from other hybrid approaches, as it uses SDN nodes to partition an OSPF domain into sub-domains thereby achieving the traffic engineering capabilities comparable to full SDN operation. We place SDN-enabled routers as subdomain border nodes, while the operation of the OSPF protocol continues unaffected. In this way, the SDN controller can tune routing protocol updates for traffic engineering purposes before they are flooded into sub-domains. While local routing inside sub-domains remains stable at all times, inter-sub-domain routes can be optimized by determining the routes in each traversed sub-domain. As the majority of traffic in non-trivial topologies has to traverse multiple subdomains, our simulation results confirm that a few SDN nodes allow traffic engineering up to a degree that renders full SDN deployment unnecessary.

A Dynamic Wideband Directional Channel Model for Vehicle-to-Vehicle Communications

Abstract: Vehicle-to-vehicle (V2V) communications have received a lot of attention due to their numerous applications in traffic safety. The design, testing, and improvement of the V2V system hinge critically on the understanding of the propagation channels. An important feature of the V2V channel is the time variance. To statistically model the time-variant V2V channels, a dynamic wideband directional channel model is proposed in this paper, based on measurements conducted at 5.3 GHz in suburban, urban, and underground parking environments. The model incorporates both angular and delay domain properties and the dynamic evolution of multipath components (MPCs). The correlation matrix distance is used to determine the size of local wide-sense stationary (WSS) region. Within each WSS time window, MPCs are extracted using the Bartlett beamformer. A multipath distance-based tracking algorithm is used to identify the “birth” and “death” of such paths over different stationarity regions, and the lifetime of MPC is modeled with a truncated Gaussian distribution. Distributions for the number of multipaths and their positions are statistically modeled. Within each path lifetime, the initial power is found to depend on the excess delay, and a linear polynomial function is used to model the variations within the lifetime. In addition, a Nakagami distribution is suggested to describe the fading behavior. Finally, the model implementation is validated by comparison of second-order statistics between measurements and simulations.

Fusion of color images and LiDAR data for lane classification

Abstract: Lane classification is a fundamental problem for autonomous driving and map-aided localization. Many existing algorithms rely on special designed 1D or 2D filters to extract features of lane markings from either color images or LiDAR data. However, these handcrafted features could not be robust under various driving and lighting conditions. In this paper, we propose a novel algorithm to fuse color images and LiDAR data together. Our algorithm consists of two stages. In the first stage, we segment road surfaces and register LiDAR data with the corresponding color images. In the second stage, we train convolutional neural networks (CNNs) to classify image patches into lane markings and non-markings. Comparing with the algorithms based on handcrafted features, our algorithm learns a set of kernels to extract and integrate features from two different modalities. The pixel-level classification rate in our experiments shows that our algorithm is robust to different conditions such as shadows and occlusions.

Citywide traffic congestion estimation with social media

Abstract: Conventional traffic congestion estimation approaches require the deployment of traffic sensors or large-scale probe vehicles. The high cost of deploying and maintaining these equipments largely limits their spatial-temporal coverage. This paper proposes an alternative solution with lower cost and wider spatial coverage by exploring traffic related information from Twitter. By regarding each Twitter user as a traffic monitoring sensor, various real-time traffic information can be collected freely from each corner of the city. However, there are two major challenges for this problem. Firstly, the congestion related information extracted directly from real-time tweets are very sparse due both to the low resolution of geographic location mentioned in the tweets and the inherent sparsity nature of Twitter data. Secondly, the traffic event information coming from Twitter can be multi-typed including congestion, accident, road construction, etc. It is non-trivial to model the potential impacts of diverse traffic events on traffic congestion. We propose to enrich the sparse real-time tweets from two directions: 1) mining the spatial and temporal correlations of the road segments in congestion from historical data, and 2) applying auxiliary information including social events and road features for help. We finally propose a coupled matrix and tensor factorization model to effectively integrate rich information for Citywide Traffic Congestion Eestimation (CTCE). Extensive evaluations on Twitter data and 500 million public passenger buses GPS data on nearly 700 mile roads of Chicago demonstrate the efficiency and effectiveness of the proposed approach.

Towards 5G Security

Abstract: This paper discusses potential security requirements and mechanisms for 5G mobile networks. It does not intend to do so exhaustively, but rather aims at initiating and spurring the work towards a 5G security architecture.

Device-to-Device Link Admission Policy Based on Social Interaction Information

Abstract: We propose an admission policy for device-to-device (D2D) communication link based on a social interaction model. We utilize prior information on user interactive statistics such as contact frequency and contact duration among potential D2D pairs to determine the success rate of message delivery. Targeting a particular delivery success rate, we further propose methods to identify admissible D2D links according to their statistical channel state information. The social-interaction-based D2D admission policy provides a real-life guideline for practical D2D underlay operation in cellular networks.