Analysis of User-Based Insurance Telematics data via Smartphone sensors in Driver Behavior and Safety
Bibliographic record
Abstract
RÉSUMÉ: «RÉSUMÉ: Les accidents de la route sont la troisième cause de mortalité au Canada, posant d'importants défis financiers et sanitaires pour les individus et les compagnies d'assurance. En réponse, l'industrie de l'assurance automobile a vu une croissance rapide de l'Assurance Basée sur l'Utilisation (UBI), un segment qui utilise les données de comportement de conduite pour des tarifs d'assurance personnalisés. Historiquement, la recherche sur l'UBI s'est concentrée sur l'analyse de la variabilité de l'accélération, de la vitesse et des angles des véhicules à l'aide de capteurs embarqués et de données de positionnement global. Cependant, l'avènement des smartphones a révolutionné la collecte de données dans la sécurité routière, réduisant onsidérablement les coûts grâce à leur disponibilité généralisée et à leurs capacités de capteurs avancées. Malgré leur potentiel, les défis liés à la qualité des données, à l'incertitude et à la variabilité entravent leur adoption complète. Cette recherche se concentre sur l'exploitation des données cinématiques issues des programmes UBI à Montréal et Ottawa, en utilisant des capteurs de smartphones tels que les accéléromètres, GNSS, magnétomètres et gyroscopes. Notre méthodologie s'inspire des algorithmes de l’Android pour l'estimation de l'azimut, adaptée pour le jeu de données UBI. Nous avons mené des expériences contrôlées avec un iPhone 13 pour assurer un alignement précis du téléphone avec le véhicule.» ABSTRACT: «ABSTRACT: In recent years, smartphones have become popular for collecting sensor measurements and traffic safety events, owing to their ability to reduce data collection costs substantially. The quality, uncertainty, and variability of these measurements pose barriers to widespread adoption. This research aims to leverage the kinematic data collected in the context of the Usage-based insurance (UBI) program from two cities in Canada, Montreal, and Ottawa. The first goal of this thesis is to explore the alignment of smartphones with vehicles. Specifically, we utilize sensors such as the accelerometer, GNSS, and magnetometer. Our method is inspired by Android's algorithms for estimating azimuth from magnetic values combined with acceleration. This algorithm has been modified in response to the UBI dataset. We have also designed experiments using an iPhone 13 in a controlled environment. This ensures that the position of the phone in relation to the vehicle is known. A crucial component of this algorithm is determining the yaw - the rotation of the phone so that it aligns with the motion of the vehicle. Our approach to determining this angle involves calculating the difference between the heading (obtained from GNSS recordings that indicate the direction of the car) and the smartphone's orientation. Subsequently, the Euler method is employed to rotate the accelerometer readings on the x, y, and z axes. The final step in our process is to verify that the new accelerometer values accurately represent the GNSS speed, braking, and deviation of GNSS speed, ensuring they are all aligned in the same direction as the vehicle.»
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
Direct model labels (unvalidated)
Per-model category and study-design labels from the labeling rounds. They are machine output, unvalidated, and the disagreement between models ships as data. No study design here is MEDLINE-validated yet.
| Model arm | Categories | Study design | Confidence |
|---|---|---|---|
| gemma | no category Domain: not available · Genre: Empirical About the Canadian research system: no · About a Canadian topic: no | Observational | low |
| gpt | no category Domain: not available · Genre: Empirical About the Canadian research system: no · About a Canadian topic: no | Observational | high |
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.000 | 0.003 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.002 | 0.002 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.001 | 0.000 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.002 | 0.001 |
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, unvalidatedLabeled directly by 2 models reading the full record.
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".