User Trajectory Research Articles

Physics-Informed Neural Network Model for Predictive Risk Assessment and Safety Analysis

This paper presents a physics-informed neural network (PINN) designed to predict the future locations of both vehicles and pedestrians, providing critical insights into road safety risks. By forecasting potential trajectories of road users, the proposed model informs preemptive strategies to avoid accidents. The physics model incorporates the intelligent driver model for vehicles and the social force model for pedestrians. The stochastic nature of risk evaluation is addressed by probabilistically predicting future locations based on the expected distribution in a two-dimensional open space. The framework accurately assesses the risk by predicting the future locations of vehicles and pedestrians within a 2- to 4-s time frame with approximately 2% error rates. The risk evaluation performance of the proposed framework was tested by calculating the time to collision (TTC) between vehicles and pedestrians and analyzing traffic conflicts. Leveraging the probabilistic predictions, the TTC was evaluated stochastically using Monte Carlo simulations and the Kolmogorov–Smirnov test, enabling a more granular and effective traffic conflict analysis. The developed method demonstrated over 95% accuracy when evaluating potentially dangerous events occurring within 3 s or less, providing actionable insights for improving road safety. The framework was deployed in a real-world setting, demonstrating reliable and robust test results. This comprehensive approach is expected to pave the way for more effective risk evaluation and mitigation at intersections and on roads.

Exploring the Roles of Informational and Emotional Language in Online Government Interactions to Promote Citizens’ Continuous Participation

Effective government responses are crucial for fostering citizens’ online engagement, yet our understanding of the impacts of the language used in government responses remains limited. Based on Language Expectancy Theory, this study investigates how emotional and informational language on electronic government petition platforms (EGPPs) influence citizens’ willingness to engage continuously. The study employs a dual-method approach, combining user trajectory analysis and experimental data, to test the developed model and hypotheses. Findings reveal that both emotional and informational language significantly enhance citizen engagement, mediated by the alignment of government responses with citizen expectations. This research provides novel insights into digital governance, extends existing literature on e-participation, and offers practical recommendations for optimizing governmental communication strategies to foster greater public involvement.

Interactions among cyclists riding the wrong way on the bicycle path

Cycling in the opposite direction can lead to many critical interaction situations and sometimes to severe crashes among cyclists. Unfortunately, no official statistics are kept of such situations in Germany. Since the number of cyclists increases in many locations in Germany faster than the cycling infrastructure improves, we can expect more of such dangerous situations in the near future. To reduce their number and severity and to develop realistic simulation models, it is essentially important to understand, how cyclists interact with each other in this particular scenario of wrong-way cycling and what consequences result for safety and cycling behaviour. This paper presents methodology and descriptive results of a traffic observation study at a signalled urban intersection in Braunschweig, Germany, with separated bicycle and footpaths. At this instrumented intersection, road user trajectories were recorded and analysed with regard to identify interactions between normal and wrong-way cyclists, and to find behavioural patterns. It appeared that several different behavioural patterns, for instance switching from bicycle path to footpath, occurred, speeds of wrong-way cyclists were slower. The distances before switching appeared to be different in some of the patterns while in others they appeared to be similar.

UeLP: accurate user linkage across social platforms against location errors

Abstract User linkage across social platforms can connect the accounts of the same user across different social networks, which is crucial for the identification of users’ multiple social identities and cross-platform association analysis. Cross-platform user linkage based on location is a typical method in current research. These methods typically rely on check-in data to calculate user similarity. However, different from check-in location, the location data obtained from instant messaging social platforms may contain random errors, leading to low accuracy of user linkage of such methods. To solve this problem, this paper proposes an accurate user linkage method across social platforms against location errors. First, unlike existing methods that employ fixed-size grids, this paper uses a multi-grained spatio-temporal grid to organize data, in order to accurately extract user features from error locations. Then, by extracting coarse-grained movement pattern features from user trajectories, candidate users are filtered out, and a small subset of candidate uses is generated to effectively reduce the search space. Next, we establish a weight model based on grid contribution and motion sequence similarity to extract location and temporal features with stronger user orientation. Finally, according to the weight model, the weighted cluster center distance of trajectories is used to calculate the similarity between two different user trajectories. The user with the highest similarity is selected from the candidate subset to complete the user linkage. The extensive experiments are conducted on six public datasets containing 115 866 trajectories and a self-built dataset with 5358 trajectories. The results show the following: compared with the four existing typical location-based methods $k$-BCT, GS, TF-IDF, and TF-IWF, the accuracy Acc@1 is improved by an average of 33%, 44.94%, 15.2%, and 14.55%, respectively, and the accuracy Acc@3 is improved by 30.52%, 34.67%, 13.84%, and 13.19%, respectively.

Secure and Efficient k-anonymous Trajectory Privacy Protection Method based on Differential Privacy

The Location-based service scheme have already involved in every aspect of People's daily life and are increasingly used in various industries. Aiming at the problem of the security and efficiency of mobile terminal users’ trajectory privacy protection in location-based service, we propose a k-anonymous trajectory privacy protection scheme based on differential privacy. This scheme adopts differential privacy technology to add Laplace noise to the user's trajectory many times to generate 2k noise trajectory, and then according to the trajectory similarity to determine k-1 noise users whose trajectory are similar to the user trajectory, and sets them and the real user as an anonymous user group, and then uses the anonymous user group to request LBS services. Security analysis shows that the scheme satisfies the security features of anonymity, unforgeability, and anti-counterfeiting attack. The simulation results show that the scheme not only guarantees the similarity between the false trajectory and the real trajectory but also has higher execution efficiency.

Do cyclists disregard ‘priority-to-the-right’ more often than motorists?

The integration of autonomous driving into mixed traffic environments poses unique challenges, especially at unsignalized intersections where communication and cooperation with human road users is necessary. This study addresses a critical research gap by investigating the adherence to the “priority-to-the-right” rule at an urban unsignalized intersection, comparing behavioral patterns of cyclists and motorists. Utilizing stationary mounted cameras, a 12-day traffic observation of an urban T-intersection in Braunschweig, Germany, was conducted. The class and the trajectory of road users as well as their scenarios were identified. In 202 cases, cars appearing from the right (ego, with priority) encountered cars or bikes coming from the left (foe, without priority). The study analyzed the impact of variables, ego’s direction, foe’s class and lateral position and their arrival time on their passing order through descriptive statistics and logistic regression. The findings reveal that cyclists disregard the “priority-to-the-right” rule more often than motorists. Additionally, road users who have the priority are more likely to yield when turning right, arriving at the intersection later, and encountering an opposing road user who is close to the center of the road. This study highlights the importance of implicit communication in traffic and provides essential benchmarks for developing more human-like autonomous driving systems, capable of interpreting and responding to nuanced road user interactions at unsignalized intersections.

Design and Validation of an Ambulatory User Support Gait Rehabilitation Robot: NIMBLE

Relearning to walk requires progressive training in real scenarios—overground—along with assistance in basic tasks, such as balancing. In addition, user ability must be maximized through compliant robotic assistance as needed. Despite decades of research, gait rehabilitation robotic devices yield controversial results. This article presents the conceptual design of a novel walking assistance and rehabilitation robot, the NIMBLE robot, aimed at providing ambulatory, bodyweight-supported gait training, assisting the user’s center of mass trajectory to aid weight transfer and dynamic balance during walking. NIMBLE consists of a robotic mobile frame, a partial bodyweight support (PBWS) system, an ambulatory lower-limb exoskeleton (Exo-H3) and a cable-driven pelvis-assisting robot. Designed as a modular structure, it differentiates hierarchical communication levels through a Robot Operating System (ROS) 2 network. We present the mechatronic design and experimental results assessing the impact of the mechatronic coupling between the robotic modules on the walking kinematics and the frame movement control performance. The robotic frame hardly affects the walking kinematics up to 2 degrees in both the sagittal and frontal planes, making it feasible for lateral balance and weight translation training. Moreover, it successfully tracks and follows user trajectories. The NIMBLE robotic frame assessment shows promising results for ambulatory gait rehabilitation.

Unveiling user interests: A deep user interest exploration network for sequential location recommendation

Sequential location recommendation, also known as next location prediction, is a crucial task aiming to compute the likelihood of a user visiting a certain point of interest (POI). It has significant applications in route planning and location-based advertisements. Despite the utilization of geographical information, existing methods often adopt a rigidly sequential prediction mode, focusing solely on predicting the next destination (i.e., where to go), which limits their ability to effectively model user trajectories in terms of predicting from where the user came. Moreover, these methods struggle to capture the evolving and deep user interests along their trajectory. To deal with these problems, we propose a novel approach called Deep User Interest Exploration Network (DUIEN) for sequential location recommendation. DUIEN incorporates two essential components to enhance recommendation accuracy. First, we employ a bidirectional neural network with the Cloze objective, enabling it to predict random masked POIs within user trajectories. This approach aids the model in gaining a better understanding of the context and dynamics of user trajectories. Second, we introduce a user interest exploration layer that seamlessly integrates into the sequential learning process. This layer effectively utilizes tag information associated with POIs and learns the intricate interplay between user interests and their trajectories using a deep user interest exploration method. In our experiments, DUIEN outperforms state-of-the-art sequential location recommendation models, demonstrating the effectiveness of the designed update mechanisms for deep user interest exploration. This showcases the potential of our approach to significantly enhance the accuracy of sequential location recommendation.

Trajectory privacy preservation model based on LSTM-DCGAN

Rapid scientific and technological development has brought many innovations to electronic devices, which has greatly improved our daily lives. Nowadays, many apps require the permission to access user location information, causing the concern on user privacy and making it an important task to protect user trajectory information. This paper proposes a novel model called LSTM-DCGAN by integrating LSTM (Long Short-Term Memory Network) with DCGAN (Deep Convolution Generative Adversarial Network). LSTM-DCGAN takes the advantages of LSTM to remember attributes in the trajectory data and the generator and the discriminator in DCGAN to generate and discriminate the trajectories. The proposed model is trained using real user trajectory data and the experimental results are validated from the perspectives of both effectiveness and practicality. Results show that the proposed LSTM-DCGAN model outperforms similar methods in generating synthesized trajectories that are similar to real trajectories in terms of the temporal and the spatial characteristics. In addition, various influencing factors are evaluated to investigate ways of further improving and optimizing the model. Overall, the proposed LSTM-DCGAN model can achieve the balance between the effectiveness of privacy protection and the practicality of user trajectory data and can thus be applied to safeguarding user trajectory information.

ViewR: Architectural-Scale Multi-User Mixed Reality With Mobile Head-Mounted Displays.

The emergence of mobile head-mounted displays with robust "inside-out" markerless tracking and video-passthrough permits the creation of novel mixed reality (MR) experiences in which architectural spaces of arbitrary size can be transformed into immersive multi-user visualisation arenas. Here we outline ViewR, an open-source framework for rapidly constructing and deploying architectural-scale multi-user MR experiences. ViewR includes tools for rapid alignment of real and virtual worlds, tracking loss detection and recovery, user trajectory visualisation and world state synchronisation between users with persistence across sessions. ViewR also provides control over the blending of the real and the virtual, specification of site-specific blending zones, and video-passthrough avatars, allowing users to see and interact with one another directly. Using ViewR, we explore the transformation of large architectural structures into immersive arenas by creating a range of experiences in various locations, with a particular focus on architectural affordances such as mezzanines, stairs, gangways and elevators. Our tests reveal that ViewR allows for experiences that would not be possible with pure virtual reality, and indicate that, with certain strategies for recovering from tracking errors, it is possible to construct large scale multi-user MR experiences using contemporary consumer virtual reality head-mounted displays.

Understanding Places Exploration and Visitation via Human Mobility Mining

Spatial-temporal data is widely available because of advancement in location acquisition technologies (GPS, GSM, Wifi, etc.) over past decades. This spatial-temporal data can easily be used to leverage user's trajectories, history, and habits to develop location-based services (LBS). But to leverage user's history based on location is bit challenging, in this paper we used the user's GPS log data to discover trajectories and analyze different life patterns over it. We proposed that a human spends 80% of his life on known places or safe places. We converted GPS patterns into visitation locations, converted them into daily and periodic trends and analyzed them over time to prove our assumption. We used the GPS data collected by Microsoft Research Asia, this data has 182 users with 73 users with their mode of movement information i.e., Taxi, Walk, Train etc.

DeepSneak: User GPS Trajectory Reconstruction from Federated Route Recommendation Models

Decentralized machine learning, such as Federated Learning (FL), is widely adopted in many application domains. Especially in domains like recommendation systems, sharing gradients instead of private data has recently caught the research community’s attention. Personalized travel route recommendation utilizes users’ location data to recommend optimal travel routes. Location data is extremely privacy sensitive, presenting increased risks of exposing behavioural patterns and demographic attributes. FL for route recommendation can mitigate the sharing of location data. However, this paper shows that an adversary can recover the user trajectories used to train the federated recommendation models with high proximity accuracy. To this effect, we propose a novel attack called DeepSneak, which uses shared gradients obtained from global model training in FL to reconstruct private user trajectories. We formulate the attack as a regression problem and train a generative model by minimizing the distance between gradients. We validate the success of DeepSneak on two real-world trajectory datasets. The results show that we can recover the location trajectories of users with reasonable spatial and semantic accuracy.

Next location prediction using heterogeneous graph-based fusion network with physical and social awareness

Location prediction based on social media information is highly valuable in human mobility research and has multiple real-life applications. However, existing research methods often ignore social influences, largely ignoring implicit information regarding interactions between users and geographical locations. Additionally, they generally employ single modeling structures, which restricts the effective integration of complex spatiotemporal characteristics and factors influencing user mobility. In this context, we propose a novel network with physical and social awareness that expresses both physical and social influences of user mobility from a global perspective based on a heterogeneous graph constructed using users and spatial locations as nodes and relationships between them as edges. This graph enables the model to leverage information from connected nodes and edges to infer missing or unobserved data. The model predicts future locations of users by effectively integrating the temporal and spatial features of user trajectory series. The proposed model is validated using three social media datasets. The experimental results demonstrate that the proposed method outperforms the state-of-the-art baseline models. This indicates the importance of considering complex interactions between users and locations, as well as the various influences of physical and social spaces.

A Placement Method of the 5G Edge Nodes Based on the Hotspot Distribution of Mobile Users

Due to the emergence of various new applications, such as short videos and online games, higher requirements of their computing and storage capacity are demanded of mobile networks. The traditional cloud computing paradigm has the shortcomings of large latency and high bandwidth demand of the core network. Therefore, how to mine the hotspot distribution of these applications and reasonably configure 5G edge nodes to reduce latency and core network bandwidth are facing great challenges. To address these issues, we designed a placement method for the 5G edge nodes based on mobile hotspots. In this method, we first cluster all locations from the user trajectories to obtain the cluster areas. Further, we extract the features, such as the number of users and duration time in all cluster areas, and extract the hotspots from all cluster areas based on the features of each cluster. Then, we introduce the base station’s high load utilization rate and the core network’s bandwidth reduction rate as the optimization parameters to construct the mathematical model of multi-objective optimization. Finally, we formalize the model into a 0–1 integer programming problem and design a greedy algorithm to solve this model. We also complete a series of experiments to evaluate our proposed methods using the GeoLife dataset. The experimental results show that the high load utilization rate can be increased up to 7.69%, and the bandwidth reduction rate of the core network can be improved up to 6.34%.

WiVelo: Fine-grained Wi-Fi Walking Velocity Estimation

Passive human tracking using Wi-Fi has been researched broadly in the past decade. Besides straightforward anchor point localization, velocity is another vital sign adopted by the existing approaches to infer user trajectory. However, state-of-the-art Wi-Fi velocity estimation relies on Doppler-Frequency-Shift (DFS), which suffers from the inevitable signal noise incurring unbounded velocity errors, further degrading the tracking accuracy. In this article, we present WiVelo, which explores new spatial-temporal signal correlation features observed from different antennas to achieve accurate velocity estimation. First, we use subcarrier shift distribution (SSD) extracted from channel state information (CSI) to define two correlation features for direction and speed estimation, separately. Then, we design a mesh model calculated by the antennas’ locations to enable a fine-grained velocity estimation with bounded direction error. Finally, with the continuously estimated velocity, we develop an end-to-end trajectory recovery algorithm to mitigate velocity outliers with the property of walking velocity continuity. We implement WiVelo on commodity Wi-Fi hardware and extensively evaluate its tracking accuracy in various environments. The experimental results show our median and 90-percentile tracking errors are 0.47 m and 1.06 m, which are half and a quarter of state-of-the-art. The datasets and source codes are published through Github ( https://github.com/research-source/code ).

TUMDOT–MUC: Data Collection and Processing of Multimodal Trajectories Collected by Aerial Drones

Currently available trajectory data sets undoubtedly provide valuable insights into traffic events, the behavior of road users and traffic flow theory, thus enabling a wide range of applications. However, there are still shortcomings that need to be addressed: (i) the continuous temporal recording (ii) of a coherent area covering several intersections (iii) with the detection of all road users, including pedestrians and cyclists. Therefore, this study focuses on the design of a large-scale aerial drone observation in the city of Munich, Germany, as well as the processing steps and the description of the resulting data set. Using twelve camera-equipped, unmanned aerial drones, the observation monitored an inner urban road section with a length of 700 m continuously for several hours during the afternoon peak hours on two working days. The trajectories of all road users were then extracted from the videos and post-processed in order to obtain a coherent and accurate data set. The resulting trajectories contain information on the category, dimensions, location, velocity, acceleration and orientation of each road user at each frame, merged continuously in time and space across several drone observation areas and subsequent time slots. The data, therefore, includes various interactions between different modes of motorized traffic and active mobility users like pedestrians and cyclists. The whole data set and the supporting data are available open source for research purposes to ensure global accessibility.

Addressing environmental stochasticity in reconfigurable intelligent surface aided unmanned aerial vehicle networks: Multi-task deep reinforcement learning based optimization for physical layer security

Deep Reinforcement Learning (DRL) is a powerful tool for optimizing communications in reconfigurable intelligent surface (RIS)-assisted millimeter-wave unmanned aerial vehicle (UAV) systems, particularly for ensuring physical layer security by dynamically maximizing signal strength for legitimate users while minimizing leakage towards eavesdroppers. However, existing approaches often overfit the DRL agent to a non-stochastic environment, where the trajectories of users and eavesdroppers are fixed. This results in a single-task DRL agent that cannot generalize beyond the training environment/task. In response, this study introduces a novel multi-task DRL framework to address this limitation. Firstly, recognizing that not all tasks share the same solution, we proposed a reward-driven clustering method to group similar tasks based on their reward scores. This enables the training of a shared multi-task DRL agent for each cluster. Each cluster can be treated as an expertise, and the cluster-specific multi-task DRL agents are considered the 'experts'. Meanwhile, we leverage federated learning (FL) to speed up the training of the shared multi-task DRL agent. By combining these two techniques, our reward-driven clustered FL autonomously groups similar tasks together rapidly. Simulation results demonstrate a reduction in the number of agents from 50 to 10 specialized 'experts', while achieving a 10X speedup compared to conventional single-task DRL training. Furthermore, the study presents a mixture of experts (MoE) model that unifies these 'experts' into a single model. The MoE is trained to select the most suitable 'expert' when encountering new tasks. Our MoE exhibits robustness by effectively generalizing to 50 novel tasks without fine-tuning.

An Examination of Machine Learning-Based Outlier Identification from Mobile Phone Tracks

In this paper, two machine learning algorithms—local outlier factor (LOF) and density-based spatial clustering of applications with noise (DBSCAN)—that are used to identify outliers in the context of a continuous framework for point of interest (PoI) detection are analyzed. The mobile trajectories of users are continuously and almost instantaneously loaded into this system. These frameworks are still in their infancy, but they are already essential for large-scale sensing deployments, such as Smart City planning deployments, where the anonymous individual mobile user trajectories can be valuable to improve urban planning. There are two contributions made by this paper. First, the functional design of the entire PoI detection architecture is provided by the study. Second, the study evaluates the effectiveness.

Meta-path aware dynamic graph learning for friend recommendation with user mobility

Recently, friend recommendation has gained widespread popularity in location-based social networks (LBSNs), which provides more opportunities for users to forge new friendships. Most existing studies exploit user trajectories or check-ins of Point-Of-Interests (POIs) to predict friendships based on geographic homophily. However, the dynamics of social relationships are left insufficiently considered in modeling user preferences. In this paper, we explore how geographical and social preferences influence each other in a dynamic manner. Specifically, we propose a Meta-path aware Dynamic Graph with Subgraph Inference, named MDyGSI, which models the evolution of user preferences with time-phased sequences of POIs and social relationships for friend recommendation in LBSNs. In each time step of the evolution, geographical and social preferences are modeled through behavior-specific meta-paths in a dynamic heterogeneous graph. The formations of different meta-paths are facilitate by each other to explore mutual influences of dual preferences. To keep the dynamics of social relationships aligned with check-in history, reliable User-User connections are sampled from social graphs based on geographical collaborative filtering in each step, which also avoids noisy social interactions. Furthermore, the dual preferences are concatenated with evolutionary weights measured by a Gated Recurrent Unit for final recommendation. Experimental results on two real-world datasets show significant improvements in MDyGSI over the state-of-the-art methods.

Trajectory-User Linking via Hierarchical Spatio-Temporal Attention Networks

Trajectory-User Linking (TUL) is crucial for human mobility modeling by linking different trajectories to users with the exploration of complex mobility patterns. Existing works mainly rely on the recurrent neural framework to encode the temporal dependencies in trajectories, have fall short in capturing spatial-temporal global context for TUL prediction. To fill this gap, this work presents a new hierarchical spatio-temporal attention neural network, called AttnTUL , to jointly encode the local trajectory transitional patterns and global spatial dependencies for TUL. Specifically, our first model component is built over the graph neural architecture to preserve the local and global context and enhance the representation paradigm of geographical regions and user trajectories. Additionally, a hierarchically structured attention network is designed to simultaneously encode the intra-trajectory and inter-trajectory dependencies, with the integration of the temporal attention mechanism and global elastic attentional encoder. Extensive experiments demonstrate the superiority of our AttnTUL method as compared to state-of-the-art baselines on various trajectory datasets. The source code of our model is available at https://github.com/Onedean/AttnTUL .