Real Mobility Data Research Articles

Analyzing and forecasting service demands using human mobility data: A two-stage predictive framework with decomposition and multivariate analysis

Accurate service demand forecasts at critical facilities are fundamental for efficiently managing resources and promptly providing essential services to people and community. However, it has received little attention in the literature, mainly due to the unavailability of granular data and the lack of sophisticated forecasting methods. To address this gap, we provide a new perspective on sensing service demands at critical facilities leveraging fine-grained human mobility data, and propose a novel data-driven framework to forecast mobility patterns at the neighborhood level. Specifically, we develop a two-stage forecasting scheme to manage large-scale and complex human movement information. The first stage is to decompose the large-scale mobility data into spatial and temporal patterns, whereas the second stage is to model complex temporal dynamics using multivariate time series analysis. The proposed framework is implemented using real human mobility data obtained from mobile phone users. The results show that our model demonstrates the best predictive performance for varying forecast horizons, when compared to multiple benchmark methods including traditionally-used statistical and deep learning models. We also performed model robustness checks, showing that the proposed model is robust in making short-term and long-term forecasts. The proposed predictive framework could help businesses and local governments accurately forecast service demands for critical facilities for better allocating their resources.

Regular routine aware routing in opportunistic mobile social networks

AbstractDue to intermittent connectivity, routing protocols in opportunistic mobile social networks (OMSN) forward messages through opportunistic contacts. Previous opportunistic routing protocols focus on the past contact history of nodes to make a forwarding decision. It was found that predicting temporally evolving contacts and communities according to different times of the day would improve the accuracy of the prediction. In this paper, a routing paradigm called RRoutine is proposed that allows message holders to exchange multiple copies of the message based on the predicted number of communities. This distribution continues until the remaining copies of the message become one. Each message holder applies either inter or intra‐community forwarding and selects a set of potential forwarders based on the prediction of future encounters in the community and an expected remaining duration between two nodes (ERIMD). The estimation of ERIMD takes into account the sequence of daily activities, the time scale of current activities, and the return time to regular meetings of the nodes. Also, unlike traditional routing, where a single forwarder is selected, a set of potential candidates is selected and prioritized according to their contact prediction, with the most likely candidate receiving the highest priority. Simulation results obtained from synthetic and real mobility data show that the proposed paper improves the performance of the system in terms of message delivery rate, throughput, and latency.

Polynomial-Time Carsharing Optimization: Linear Formulation and Large-Scale Simulations

Shared mobility services are useful for low-income people that are not satisfied with public transport and cannot comfortably afford their own vehicle. Among the low-cost shared mobility services, carsharing stands out because it avoids the cost of a driver. This work applies a polynomial-time linear programming formulation to optimize different carsharing business models for the Metropolitan Region of São Paulo. Real mobility data are used, focusing on inhabitants who are able to drive. Scenarios are evaluated varying distances that clients would be flexible to walk to get to an available vehicle or parking slot, comparing carsharing business models, and considering different rental prices. Results show that it is possible to offer a profitable low-cost carsharing service without performing vehicle relocations if clients are flexible enough to walk and if only a subset of trips is selected to be served. Results also demonstrate that trips selected to be served are similar among the different business models; are concentrated on São Paulo’s downtown region; are shorter than the average trip, but otherwise behave in a similar way as compared to the complete set of trips; and the lack of parking slots may be a risk to the carsharing company.

Assessing the effectiveness of perimeter lockdowns as a response to epidemics at the urban scale

From September 2020 to May 2021 Madrid region (Spain) followed a rather unique non-pharmaceutical intervention (NPI) by establishing a strategy of perimeter lockdowns (PLs) that banned travels to and from areas satisfying certain epidemiological risk criteria. PLs were pursued to avoid harsher restrictions, but some studies have found that the particular implementation by Madrid authorities was rather ineffective. Based on Madrid’s case, we devise a general, minimal framework to investigate the PLs effectiveness by using a data-driven metapopulation epidemiological model of a city, and explore under which circumstances the PLs could be a good NPI. The model is informed with real mobility data from Madrid to contextualize its results, but it can be generalized elsewhere. The lowest lockdown activation threshold Theta considered (14-day cumulative incidence rate of 20 cases per every 10^5 inhabitants) shows a prevalence reduction >20% with respect to the scenario Theta =10^3, more akin to the case of Madrid, and assuming no further mitigation. Only the combination of Theta =20 and mobility reduction >90% can avoid PLs for more than >20% of the system. The combination of low Theta and strong local transmissibility reduction is key to minimize the impact, but the latter is harder to achieve given that we assume a situation with highly mitigated transmission, resembling the one observed during the second wave of COVID-19 in Madrid. Thus, we conclude that a generalized lockdown is hard to avoid under any realistic setting if only this strategy is applied.

Double Q-learning based routing protocol for opportunistic networks

Opportunistic Delay Tolerant Networks also referred to as Opportunistic Networks (OppNets) are a subset of wireless networks having mobile nodes with discontinuous opportunistic connections. As such, developing a performant routing protocol in such an environment remains a challenge. Most research in the literature have shown that reinforcement learning-based routing algorithms can achieve a good routing performance, but these algorithms suffer from under-estimations and/or over-estimations. Toward addressing these shortcomings, in this paper, a Double Q-learning based routing protocol for Opportunistic Networks framework named Off-Policy Reinforcement-based Adaptive Learning (ORAL) is proposed, which selects the most suitable next-hop node to transmit the message toward its destination without any bias by using a weighted double Q-estimator. In the next-hop selection process, a probability-based reward mechanism is involved, which considers the node’s delivery probability and the frequency of encounters among the nodes to boost the protocol’s efficiency. Simulation results convey that the proposed ORAL protocol improves the message delivery ratio by maintaining a trade-off between underestimation and overestimation. Simulations are conducted using the HAGGLE INFOCOM 2006 real mobility data trace and synthetic model, showing that when time-to-live is varied, (1) the proposed ORAL scheme outperforms DQLR by 14.05%, 9.4%, 5.81% respectively in terms of delivery probability, overhead ratio and average delay; (2) it also outperforms RLPRoPHET by 16.17%, 9.2%, 6.85%, respectively in terms of delivery ratio, overhead ratio and average delay.

A study on the impact of infrastructure on transport emissions based on real driving data

The objective of the paper is to enable local authorities to improve air quality by considering road traffic polluting emissions in their decision to deal with road infrastructure. Innovative tools based on real mobility data from thousands of users were developed. These tools provide urban mobility departments with a clear vision of the impact of urban development and the associated regulations. They enable the classification of the impact of the different types of road development in place on the territory to identify best practices. The approach presented enables the visualization of the levels of polluting emissions, noise, emergency braking, and congestion on each section of the road. Then, recent infrastructure modifications, on which data could be collected before and after, were analyzed. Finally, infrastructure is classified in terms of polluting levels depending on congestion.

Using real mobility patterns to assess the impact of 100% electrified mobility in a German city

Until now, individual motorized mobility has been almost exclusively powered by fossil energy sources. The battle against climate change, however, requires a transformation of the mobility system with the ultimate objective of a full electrification of transport. Ultimately, this will increase considerably the load on the distribution grid both in overall size and through temporarily and locally distributed charging. This paper explores the effects of an assumed full electrification of individual motorized transport on the power grid in a major German city using real traffic data collected from a German traffic panel and employing a microscopic and dynamic travel simulation with the Simulation of Urban MObility tool. The main finding of the traffic simulation is that the local and temporary impact on the distribution grid is very sensitive to the distribution of charging stations and their geographical aggregation in transformers. However, behavior change in the form of charging at different points in time and locations or using less power can alleviate local peaks, up to 50% in the simulated scenario. Therefore, this paper addresses the previously uncovered need for a more profound and realistic computation that includes all important aspects of a proper traffic simulation. It further extends the barely covered field of dynamic simulations that operate on real mobility data. With that, limitations arise from the process of transposing driving data into geographical data as it is very sensitive to underlying assumptions. However, we are convinced that the spatialization of charging stations and the randomization of trip assignments represent the most transparent and meaningful way of overcoming these limiting factors.

Epidemic spreading in populations of mobile agents with adaptive behavioral response

Despite the advanced stage of epidemic modeling, there is a major demand for methods to incorporate behavioral responses to the spread of a disease, such as social distancing and adoption of prevention methods. Mobility plays an important role on epidemic dynamics and is also affected by behavioral changes, but there are many situations in which real mobility data is incomplete or inaccessible. We present a model for epidemic spreading in temporal networks of mobile agents that incorporates local behavioral responses. Susceptible agents are allowed to move towards the opposite direction of infected agents in their neighborhood. We show that this mechanism considerably decreases the stationary prevalence when the spatial density of agents is low. However, for higher densities, the mechanism causes an abrupt phase transition, where a new bistable phase appears. We develop a semi-analytic approach for the case when the mobility is fast compared to the disease dynamics, and use it to argue that the bistability is caused by the emergence of spatial clusters of susceptible agents. Finally, we characterize the temporal networks formed in the fast mobility regime, showing how the degree distributions and other metrics are affected by the behavioral mechanism. Our work incorporates results previously known from adaptive networks into population of mobile agents, which can be further developed to be used in mobility-driven models.

Modeling Adversarial Behavior Against Mobility Data Privacy

Privacy risk assessment is a crucial issue in any privacy-aware analysis process. Traditional frameworks for privacy risk assessment systematically generate the assumed knowledge for a potential adversary, evaluating the risk without realistically modelling the collection of the background knowledge used by the adversary when performing the attack. In this work, we propose Simulated Privacy Annealing (SPA), a new adversarial behavior model for privacy risk assessment in mobility data. We model the behavior of an adversary as a mobility trajectory and introduce an optimization approach to find the most effective adversary trajectory in terms of privacy risk produced for the individuals represented in a mobility data set. We use simulated annealing to optimize the movement of the adversary and simulate a possible attack on mobility data. We finally test the effectiveness of our approach on real human mobility data, showing that it can simulate the knowledge gathering process for an adversary in a more realistic way.

The importance of non-pharmaceutical interventions during the COVID-19 vaccine rollout.

The promise of efficacious vaccines against SARS-CoV-2 is fulfilled and vaccination campaigns have started worldwide. However, the fight against the pandemic is far from over. Here, we propose an age-structured compartmental model to study the interplay of disease transmission, vaccines rollout, and behavioural dynamics. We investigate, via in-silico simulations, individual and societal behavioural changes, possibly induced by the start of the vaccination campaigns, and manifested as a relaxation in the adoption of non-pharmaceutical interventions. We explore different vaccination rollout speeds, prioritization strategies, vaccine efficacy, as well as multiple behavioural responses. We apply our model to six countries worldwide (Egypt, Peru, Serbia, Ukraine, Canada, and Italy), selected to sample diverse socio-demographic and socio-economic contexts. To isolate the effects of age-structures and contacts patterns from the particular pandemic history of each location, we first study the model considering the same hypothetical initial epidemic scenario in all countries. We then calibrate the model using real epidemiological and mobility data for the different countries. Our findings suggest that early relaxation of safe behaviours can jeopardize the benefits brought by the vaccine in the short term: a fast vaccine distribution and policies aimed at keeping high compliance of individual safe behaviours are key to mitigate disease resurgence.

QoS-Aware Approximate Query Processing for Smart Cities Spatial Data Streams.

Large amounts of georeferenced data streams arrive daily to stream processing systems. This is attributable to the overabundance of affordable IoT devices. In addition, interested practitioners desire to exploit Internet of Things (IoT) data streams for strategic decision-making purposes. However, mobility data are highly skewed and their arrival rates fluctuate. This nature poses an extra challenge on data stream processing systems, which are required in order to achieve pre-specified latency and accuracy goals. In this paper, we propose ApproxSSPS, which is a system for approximate processing of geo-referenced mobility data, at scale with quality of service guarantees. We focus on stateful aggregations (e.g., means, counts) and top-N queries. ApproxSSPS features a controller that interactively learns the latency statistics and calculates proper sampling rates to meet latency or/and accuracy targets. An overarching trait of ApproxSSPS is its ability to strike a plausible balance between latency and accuracy targets. We evaluate ApproxSSPS on Apache Spark Structured Streaming with real mobility data. We also compared ApproxSSPS against a state-of-the-art online adaptive processing system. Our extensive experiments prove that ApproxSSPS can fulfill latency and accuracy targets with varying sets of parameter configurations and load intensities (i.e., transient peaks in data loads versus slow arriving streams). Moreover, our results show that ApproxSSPS outperforms the baseline counterpart by significant magnitudes. In short, ApproxSSPS is a novel spatial data stream processing system that can deliver real accurate results in a timely manner, by dynamically specifying the limits on data samples.

Heterogeneous interventions reduce the spread of COVID-19 in simulations on real mobility data

Major interventions have been introduced worldwide to slow down the spread of the SARS-CoV-2 virus. Large scale lockdown of human movements are effective in reducing the spread, but they come at a cost of significantly limited societal functions. We show that natural human movements are statistically diverse, and the spread of the disease is significantly influenced by a small group of active individuals and gathering venues. We find that interventions focused on these most mobile individuals and popular venues reduce both the peak infection rate and the total infected population while retaining high social activity levels. These trends are seen consistently in simulations with real human mobility data of different scales, resolutions, and modalities from multiple cities across the world. The observation implies that compared to broad sweeping interventions, more heterogeneous strategies that are targeted based on the network effects in human mobility provide a better balance between pandemic control and regular social activities.

Mobility Prediction Using a Weighted Markov Model Based on Mobile User Classification.

The vast amounts of mobile communication data collected by mobile operators can provide important insights regarding epidemic transmission or traffic patterns. By analyzing historical data and extracting user location information, various methods can be used to predict the mobility of mobile users. However, existing prediction algorithms are mainly based on the historical data of all users at an aggregated level and ignore the heterogeneity of individual behavior patterns. To improve prediction accuracy, this paper proposes a weighted Markov prediction model based on mobile user classification. The trajectory information of a user is extracted first by analyzing real mobile communication data, where the complexity of a user’s trajectory is measured using the mobile trajectory entropy. Second, classification criteria are proposed based on different user behavior patterns, and all users are classified with machine learning algorithms. Finally, according to the characteristics of each user classification, the step threshold and the weighting coefficients of the weighted Markov prediction model are optimized, and mobility prediction is performed for each user classification. Our results show that the optimized weighting coefficients can improve the performance of the weighted Markov prediction model.

EmuCD: An Emulator for Content Dissemination Protocols in Vehicular Networks

The development of protocols for mobile networks, especially for vehicular ad-hoc networks (VANETs), presents great challenges in terms of testing in real conditions. Using a production network for testing communication protocols may not be feasible, and the use of small networks does not meet the requirements for mobility and scale found in real networks. The alternative is to use simulators and emulators, but vehicular network simulators do not meet all the requirements for effective testing. Aspects closely linked to the behaviour of the network nodes (mobility, radio communication capabilities, etc.) are particularly important in mobile networks, where a delay tolerance capability is desired. This paper proposes a distributed emulator, EmuCD, where each network node is built in a container that consumes a data trace that defines the node’s mobility and connectivity in a real network (but also allowing the use of data from simulated networks). The emulated nodes interact directly with the container’s operating system, updating the network conditions at each step of the emulation. In this way, our emulator allows the development and testing of protocols, without any relation to the emulator, whose code is directly portable to any hardware without requiring changes or customizations. Using the facilities of our emulator, we tested InterPlanetary File System (IPFS), Sprinkler and BitTorrent content dissemination protocols with real mobility and connectivity data from a real vehicular network. The tests with a real VANET and with the emulator have shown that, under similar conditions, EmuCD performs closely to the real VANET, only lacking in the finer details that are extremely hard to emulate, such as varying loads in the hardware.

Vehicle-Assisted Data Delivery in Smart City: A Deep Learning Approach

Collecting the massive internet of things data produced in a large smart city is quite challenging, and recent advances in vehicle-to-everything communication makes urban vehicles to be a good candidate to conduct this task. Hence, this paper proposes a novel deep learning algorithm called DeepVDD to facilitate vehicle-assisted data delivery. First, a theoretical analysis is presented to quantitatively reveal the correlation between vehicle mobility and the success ratio of vehicle-assisted data delivery. Based on the findings in analysis, DeepVDD adopts a novel multi-headed neural network to determine the strategies for vehicles to deliver data. Comprehensive evaluations have been executed based on the real taxi mobility data in Tokyo, Japan. The results have validated that, compared with other state-of-art algorithms, DeepVDD not only improves the success ratio of data delivery, but also significantly reduces the communication overhead of vehicular networks.

The rhythm of the crowd: Properties of evolutionary community detection algorithms for mobile edge selection

The Multi-access Edge Computing (MEC) paradigm increases the computational capabilities of distributed sensing architectures, such as Mobile CrowdSensing platforms, which are designed to collect heterogeneous data from the crowd by exploiting mobile devices. In this context, our work focusses on the impact of three community detection algorithms to our edge selection strategy. In particular, we study TILES, Infomap, and iLCD which are specifically designed to identify evolving communities of users in dynamic networks. Our analysis is based on the ParticipAct data set that offers real human mobility data. We first measure the quality of the data set during an observation period of 1 year, during which the data set provides the 75% of the expected traces collected by approximately 170 users. We then compare some structural properties of the communities detected, namely Similarity, Forward Stability, Cohesion and Coverage. We conclude our study with a performance analysis of the selected Mobile MECs by varying the community detection algorithms adopted. In particular, we measure the latency and the number of satisfied requests and we show that the average latency obtained with Infomap is slightly lower than that of the other algorithms, while the average number of satisfied requests is higher when we adopt the TILES algorithm.

Mobile phone GPS data in urban ride-sharing: An assessment method for emission reduction potential

Spreading green and low-consumption transportation methods is becoming an urgent priority. Ride-sharing, which refers to the sharing ofcarjourneys so that more than one person travel in a car, and prevents the need for others to drive to a location themselves, is a critical solution to this issue. Before being introduced into one place, it needs a potential analysis. However, current studies did this kind of analysis based on home and work locations or social ties between people, which is not precise and straight enough. Few pieces of research departed from real mobility data, but uses time-consuming methodology. In this paper, we proposed an analysis framework to bridge this gap. We chose the case study of Tokyo area with over 1 million GPS travel records and trained a deep learning model to find out this potential. From the computation result, on average, nearly 26.97% of travel distance could be saved by ride-sharing, which told us that there is a significant similarity in the travel pattern of people in Tokyo and there is considerable potential of ride-sharing. Moreover, if half of the original public transit riders in our study case adopt ride-sharing, the quantity of CO2 is estimated to be reduced by 84.52%; if all of the original public transit riders in our study case adopt ride-sharing, 83.56% of CO2 emission reduction can be expected with a rebound effect because of increase of participants from public transit. Ride-sharing can not only improve the air quality of these center business districts but also alleviate some city problems like traffic congestion. We believe the analysis of the potential of ride-sharing can provide insight into the decision making of ride-sharing service providers and decision-makers.

Impact of Frequency of Location Reports on the Privacy Level of Geo-indistinguishability

Abstract Location privacy has became an emerging topic due to the pervasiveness of Location-Based Services (LBSs). When sharing location, a certain degree of privacy can be achieved through the use of Location Privacy-Preserving Mechanisms (LPPMs), in where an obfuscated version of the exact user location is reported instead. However, even obfuscated location reports disclose information which poses a risk to privacy. Based on the formal notion of differential privacy, Geo-indistinguishability has been proposed to design LPPMs that limit the amount of information that is disclosed to a potential adversary observing the reports. While promising, this notion considers reports to be independent from each other, thus discarding the potential threat that arises from exploring the correlation between reports. This assumption might hold for the sporadic release of data, however, there is still no formal nor quantitative boundary between sporadic and continuous reports and thus we argue that the consideration of independence is valid depending on the frequency of reports made by the user. This work intends to fill this research gap through a quantitative evaluation of the impact on the privacy level of Geo-indistinguishability under different frequency of reports. Towards this end, state-of-the-art localization attacks and a tracking attack are implemented against a Geo-indistinguishable LPPM under several values of privacy budget and the privacy level is measured along different frequencies of updates using real mobility data.

Using Reinforcement Learning to Minimize the Probability of Delay Occurrence in Transportation

Reducing traffic delay is of crucial importance for the development of sustainable transportation systems, which is a challenging task in the studies of stochastic shortest path (SSP) problem. Existing methods based on the probability tail model to solve the SSP problem, seek for the path that minimizes the probability of delay occurrence, which is equal to maximizing the probability of reaching the destination before a deadline (i.e., arriving on time). However, they suffer from low accuracy or high computational cost. Therefore, we design a novel and practical Q-learning approach where the converged Q-values have the practical meaning as the actual probabilities of arriving on time so as to improve the accuracy of finding the real optimal path. By further adopting dynamic neural networks to learn the value function, our approach can scale well to large road networks with arbitrary deadlines. Moreover, our approach is flexible to implement in a time dependent manner, which further improves the performance of returned path. Experimental results on some road networks with real mobility data, such as Beijing, Munich and Singapore, demonstrate the significant advantages of the proposed approach over other methods.

Message trust‐based secure multipath routing protocol for opportunistic networks

SummaryOpportunistic networks (OppNets) are composed of wireless nodes opportunistically communicating with each other. These networks are designed to operate in a challenging environment characterized by high delay, intermittent connectivity, and no guarantee of fixed path between the sender and the destination nodes. One of the most vital issues in designing and maintaining practical networks over a time period is the security of the messages flowing in OppNets. This paper proposes a new method called message trust‐based secure multipath routing protocol (MT‐SMRP) for opportunistic networks. Various routing protocols such as ProPHet, Epidemic, and HiBOp, to name a few, have been proposed for OppNets, but none of these have applied a secure multipath routing technique. The proposed MT‐SMRP scheme relays the message to the destination through the disjoint paths, each applying a soft‐encryption technique to prevent message fabrication attacks. Simulations are conducted using the Haggle Infocom'06 real mobility data traces, showing that when time‐to‐live is varied, (1) the proposed MT‐SMRP scheme outperforms D‐MUST by 18.10%, 7.55%, 3.275%, respectively, in terms of delivery probability, messages dropped, and average latency; (2) it also outperforms SHBPR by 21.30%, 7.44%, and 4.85%, respectively, in terms of delivery probability, messages dropped, and average latency. Under similar performance metrics, the performance of MT‐SMRP is also shown to be better than that of D‐MUST and SHBPR when the buffer size (respondents. the message generation interval) is varied.