Scalable Urban Crowdsensing: Data Contributor and Consumer Dynamics
Bibliographic record
Abstract
Dynamic transportation routing and parking management rely on real-time data on traffic conditions and curbside availability. Traditional data collection methods require significant infrastructure investments and processing capacity, making them costly and often impractical for large-scale implementation. In contrast, crowdsensing leverages data from users’ smart devices, offering a cost-effective alternative for collecting real-time information. However, crowdsensed data is often noisy, inaccurate, unstructured, and heavily dependent on voluntary user contributions. This thesis is motivated by the central question: how can we effectively design and manage platforms, such as parking management and routing, that rely on such incomplete yet abundant data? To address this, several key studies are presented. The first study, Leveraging Data Contributors to Enhance Social Welfare Through Crowdsensing, examines how user-generated data can improve decision-making at service facilities such as parking lots. By aggregating reported wait times, the platform provides real-time estimates that help incoming users decide whether to enter. This research determines the optimal fraction of data contributors needed to maximize the collective benefit, social welfare, of facility users. The second study, Route Choice Using Crowd-Generated Travel Time Information, explores how crowdsensed travel times influence traffic assignment. Two critical factors, contribution ratio (the proportion of travelers sharing data) and observation window (the period over which data is collected), must be balanced to minimize network-wide travel times. This study analyzes their effects on perceived travel risks and proposes solutions to maintain their optimal values, ensuring efficiency in routing decisions. The third study, Resource Allocation and Route Generation for Urban Mobile Sensing, focuses on optimizing the number and routes of sensing agents, users carrying data collection devices, to enhance efficiency. By determining the optimal number of agents and their navigation paths, this study minimizes total travel costs while ensuring that each parking spot is revisited within the specified headway constraints. A case study in Toronto demonstrates the practical applicability of the proposed optimization framework for parking occupancy detection. The fourth study, Implementing Parking Occupancy Detection Using Dashcam Footage, develops a sensing system that analyzes dashcam video to detect real-time parking occupancy. By applying advanced video processing techniques, this project provides dynamic and scalable insights into urban parking availability. It aims to overcome the limitations of traditional parking data collection tools, such as stationary cameras and sensors, which face challenges related to installation, maintenance, and regulation. Instead, mobile technologies like dashcams, LiDAR, and ultrasound sensors offer a more scalable solution for capturing on-street parking availability. Together, these studies contribute to improving the efficiency of crowdsensing platforms by analyzing the impact of three types of crowdsensed data: wait times, travel times, and urban monitoring reports. These data types are examined across three distinct domains: service facilities (e.g., parking facilities), transportation networks (e.g., scenarios involving multiple modes of transportation or alternative routes with varying congestion levels), and urban monitoring (e.g., parking availability detection, pedestrian safety, and traffic flow monitoring). The research examines how access to such data influences individual decision-making, particularly in the presence of competing alternatives, as well as its impact on overall system performance. Furthermore, it investigates the factors affecting the accuracy and reliability of crowdsensed data in each domain and explores how optimizing these factors can enhance both user experience and system-wide efficiency.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.000 | 0.000 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.001 | 0.000 |
| Bibliometrics | 0.004 | 0.004 |
| Science and technology studies | 0.001 | 0.001 |
| Scholarly communication | 0.001 | 0.004 |
| Open science | 0.003 | 0.005 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.001 | 0.002 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; both teacher heads agree on what is shown here.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".