Public (original) (raw)
Exploiting context information for estimating the performance of vehicular communications
2013 IEEE Vehicular Networking Conference, 2013
The performance of wireless vehicular communications can depend on multiple context factors, such as the propagation conditions, the traffic density, or the location of communication infrastructure units. This paper proposes and evaluates two techniques that are able to identify and quantify such dependencies, and uses them to estimate the vehicular communications performance exploiting context information. The techniques proposed have been evaluated using real-world traces from a Vehicle to Infrastructure (V2I) IEEE 802.11p measurement campaign. The obtained results show that the number of context factors considered in the estimation process can influence its accuracy, and that not all context factors have the same importance in the estimation process.
Situation Aware Multi-Task Learning for Traffic Prediction
—Due to the recent vast availability of transportation traffic data, major research efforts have been devoted to traffic prediction, which is useful in many applications such as urban planning, traffic management and navigations systems. Current prediction methods that independently train a model per traffic sensor cannot accurately predict traffic in every situation (e.g., rush hours, constructions and accidents) because there may not exist sufficient training samples per sensor for all situations. To address this shortcoming, our core idea is to explore the com-monalities of prediction tasks across multiple sensors who behave similarly in a specific traffic situation. Instead of building a model independently per sensor, we propose a Multi-Task Learning (MTL) framework that aims to first automatically identify the traffic situations and then simultaneously build one forecasting model for similar-behaving sensors per traffic situation. The key innovation here is that instead of the straightforward application of MTL where each " task " corresponds to a sensor, we relate each MTL's " task " to a traffic situation. Specifically, we first identify these traffic situations by running clustering algorithms on all sensors' data. Subsequently, to enforce the commonalities under each identified situation, we use the group Lasso regularization in MTL to select a common set of features for the prediction tasks, and we adapt efficient FISTA algorithm with guaranteed convergence rate. We evaluated our methods with a large volume of real-world traffic sensor data; our results show that by incorporating traffic situations, our proposed MTL framework performs consistently better than naively applying MTL per sensor. Moreover, our holistic approach, under different traffic situations, outperforms all the best traffic prediction approaches for a given situation by up to 18% and 30% in short and long term predictions, respectively.
IEEE ICDM 2010 Contest TomTom Traffic Prediction for Intelligent GPS Navigation
In this foreword, we summarize the IEEE ICDM 2010 Contest: “TomTom Traffic Prediction for Intelligent GPS Navigation”. The challenge was held between Jun 22, 2010 and Sep 7, 2010 as an interactive on-line competition, using the TunedIT platform (http://tunedit.org). We present the scope of the ICDM contest series in general, the scope of this year’s contest, description of its tasks, statistics about participation, details about the TunedIT platform and the Traffic Simulation Framework. A detailed description of winning solutions is part of this proceeding series.
Context-Aware Location Prediction
Over the past few years, predicting the future location of mobile ob-jects has become an important and challenging problem. With the widespread use of mobile devices, applications of location prediction include location-based ser-vices, resource allocation, smooth handoffs in cellular networks, animal migra-tion research, and weather forecasting. Most current techniques try to predict the next location of moving objects such as vehicles, people or animals, based on their movement history alone. However, ignoring the dynamic nature of mobile behavior and trying to repeatedly exploit the same common patterns, may yield wrong results, at least part of the time. Analyzing movement in its context and choosing the best movement pattern by the current situation, can reduce some of the errors and improve prediction accuracy. In this paper, we present a context-aware location prediction algorithm that utilizes various types of context infor-mation to predict future location of vehicles. We use five contextual features re-lated to either the object environment or its current movement data: current loca-tion; object velocity; day of the week; weather conditions; and traffic congestion in the area. Our algorithm incorporates these context features into the trajectory-clustering phase as well as in the location prediction phase. We evaluate our al-gorithm using two real-world GPS trajectory datasets. The experimental results demonstrate that the context-aware approach can significantly improve the accu-racy of location predictions.
European Transport Research Review
Background European cities are placing a larger emphasis on urban data consolidation and analysis for optimizing public transport in response to changing urban mobility dynamics. Despite the existing efforts, traffic data analysis often disregards vital situational context, including large-scale events, weather factors, traffic generation poles, social distancing norms, or traffic interdictions. Some of these sources of context data are still private, dispersed, or unavailable for the purpose of planning or managing urban mobility. Addressing the above observation, the Lisbon city Council has already established efforts for gathering historic and prospective sources of situational context in standardized semi-structured repositories, triggering new opportunities for context-aware traffic data analysis. Research questions The work presented in this paper aims at tackling the following main research question: How to incorporate historical and prospective sources of situational context...
Journal of Communications and Information Networks, 2018
Spectrum occupancy information is necessary in a cognitive radio network (CRN) as it helps in modeling and predicting the spectrum availability for efficient dynamic spectrum access (DSA). However, in a CRN, it is difficult to ascertain a priori the pattern of the spectrum usage of the primary user due to its stochastic behavior. In this context, the spectrum occupancy prediction proves to be very useful in enhancing the quality of experience of the secondary user. This paper investigates the practical prowess of various time-series modeling approaches and the machine learning (ML) techniques for predicting spectrum occupancy, based on a spectrum measurement campaign conducted in Jaipur, Rajasthan, India. Moreover, the comparison analysis conducted between the above two approaches highlights the trade-off in terms of the respective performance depending upon the nature of the spectrum occupancy data. Nevertheless, prediction through ML-based recurrent neural network proves to perform reasonably well, thereby providing an accurate future spectrum occupancy information for DSA.
Leveraging diverse propagation and context for multi-modal vehicular applications
2013 IEEE 5th International Symposium on Wireless Vehicular Communications (WiVeC), 2013
Vehicular wireless channels have a high degree of variability, presenting a challenge for vehicles and infrastructure to remain connected. The emergence of the white space bands for data usage enables increased flexibility for vehicular networks with distinct propagation characteristics across frequency bands from 450 MHz to 6 GHz. Since wireless propagation largely depends on the environment in operation, a historical understanding of the frequency bands' performance in a given environment could expedite band selection as vehicles transition across diverse scenarios. In this paper, we leverage knowledge of in-situ operation across frequency bands with real-time measurements of the activity level to select the the band with the highest throughput. To do so, we perform a number of experiments in typical vehicular topologies. With two models based on machine learning algorithms and an in-situ training set, we predict the throughput based on: (i.) prior performance for similar context information (e.g., SNR, GPS, relative speed, and link distance), and (ii.) real-time activity level and relative channel quality per band. In the field, we show that training on a repeatable route with these machine learning techniques can yield vast performance improvements from prior schemes.
Road Traffic Prediction Using Bayesian Networks
2012
Having prior road condition knowledge for planned or unplanned journeys will be beneficial in terms of not only time but potentially cost. Being able to obtain real-time information will further enhance these benefits. Current systems rely on huge infrastructure investments by governments to install cameras, road sensors and billboards to keep motorists informed. These efforts can only be, at best, available at pre-identified hotspots. Radio broadcast is an alternative, where they rely on reports by other motorists. However, such reports are often delayed and not tailored to individual motorist. Seeing the limitations of existing approaches to obtain real-time road conditions, this research work leverages on mobile devices that provide context sensitive information to propose a predictive analytics framework based on a Bayesian Network for road condition prediction. This paper aims to contribute to (i) defining a set of evidences (variables) that could potentially be utilized for road condition prediction and (ii) construction of a Bayesian Network model to predict road conditions. In conclusion, we presented a novel approach to provide potentially unlimited coverage of road traffic conditions with substantially reduced infrastructure investments.
Trajectory Prediction of Traffic Agents: Incorporating context into machine learning approaches
2020 IEEE 91st Vehicular Technology Conference (VTC2020-Spring), 2020
For a vehicle to navigate autonomously, it needs to perceive its surroundings and estimate the future state of the relevant traffic-agents with which it might interact as it navigates across public road networks. Predicting the future state of the perceived entities is a challenge, as these might appear to move in a stochastic manner. However, their motion is constrained to an extent by context, in particular the road network structure. Conventional machine learning methods are mainly trained using data from the perceived entities without considering roads, as a result trajectory prediction is difficult. In this paper, the notion of maps representing the road structure are included into the machine learning process. For this purpose, 3D LiDAR points and maps in the form of binary masks are used. These are used on a recurrent artificial neural network, the LSTM encoder-decoder based architecture to predict the motion of the interacting traffic agents. A comparison between the proposed solution with one that is only sensor driven (LiDAR) is included. For this purpose, NuScenes dataset is utilised, that includes annotated 3D point clouds. The results have demonstrated the importance of context to enhance our prediction performance as well as the capability of our machine learning framework to incorporate map information.
Transfer Learning for Tilt-Dependent Radio Map Prediction
IEEE Transactions on Cognitive Communications and Networking, 2020
Machine learning will play a major role in handling 1 the complexity of future mobile wireless networks by improving 2 network management and orchestration capabilities. Due to the 3 large number of parameters that can be monitored and config-4 ured in the network, collecting and processing high volumes of 5 data is often unfeasible or too expensive at network runtime, 6 which calls for taking resource management and service orches-7 tration decisions with only a partial view of the network status. 8 Motivated by this fact, this paper proposes a transfer learning 9 framework for reconstructing the radio map corresponding to a 10 target antenna tilt configuration by transferring the knowledge 11 acquired from another tilt configuration of the same antenna, 12 when no or very limited measurements are available from the 13 target. The performance of the framework is validated against 14 standard machine learning techniques on a data set collected 15 from a 4G commercial base stations. In most of the tested scenar-16 ios, the proposed framework achieves notable prediction accuracy 17 with respect to classical machine learning approaches, with a 18 mean absolute percentage error below 8%. 19 Index Terms-Radio map prediction, antenna tilt, transfer 20 learning. 21 I. INTRODUCTION 22 F IFTH generation wireless networks (5G) are expected to 23 improve the performance of cellular systems, achieving 24 higher data rates, reduced latency, higher reliability and sup-25 port for greater numbers of users. To achieve this, 5G resorts 26 to dense and heterogeneous deployments, coupled with higher 27 flexibility in the network access and core domains, which can 28 be dynamically managed in either a centralized or distributed 29 manner. To cope with such a complex scenario, it is foreseen 30 that machine learning tools will play a major role in enabling 31