Modèle IoT de gestion du feu en temps réel dans les villes intelligentes
Bibliographic record
Abstract
RÉSUMÉ: Chaque année, nous observons des milliers d'incendies résidentiels dans nos villes. Ces incendies causent de nombreux décès et dommages matériels. Au Québec, il y a en moyenne 16 500 incendies par année dont au moins 5 600 sont des incendies résidentiels. Entre 2015 et 2017, les incendies ont causé 131 décès et 540 millions de dollars CAD de dommage en moyenne par année. En 2018, il y avait 11 046 incendies avec 91 décès et 836.1 millions de dollars CAD de dommages en Ontario. Les personnes chargées d'intervenir en cas d'incendies sont les sapeurs-pompiers. Ces derniers disposent d'un temps très étroit (i.e., 5 à 10 minutes) pour sauver le maximum de personnes et arrêter la propagation des incendies. Pour aider les sapeurs-pompiers dans leurs tâches, de nombreuses solutions utilisant les Technologies de l'Information et de la Communication » (TIC) ont été mises en place. L'avènement des villes intelligentes et des nouvelles technologies, tels que le réseau 5G, les objets connectés « Internet of Things » (IoT) et la technologie « Building Information Model » (BIM), ont permis la création de solutions innovantes. Cependant, ces solutions présentent quelques lacunes comme l'incapacité de fonctionner en temps réel et de s'adapter à la plupart des situations d'incendie. Dans ce mémoire, nous proposons un nouveau modèle de gestion des incendies en temps réel. Ce modèle exploite les systèmes distribués et l'architecture microservice pour fournir un système robuste, adaptable et évolutif. Le modèle proposé est composé de quatre modules : le réseau de capteurs sans fil « Fire Emergency Sensor Network » (FESNet), l'application distribuée, les applications clients et un serveur Web qui agit comme intermédiaire entre l'application distribuée et les applications clients. Afin de valider notre modèle, un prototype capable du traitement de l'information en temps réel et de l'affichage des bâtiments en 3D a été implémenté. Nous avons effectué trois simulations en utilisant l'algorithme « Parallel Merge Sort » pour le triage des données. ABSTRACT: Every year, we see thousands of residential fires in our cities. These fires cause many deaths and property damage. In Quebec, there are an average of 16,500 fires per year of which at least 5,600 are residential fires. Between 2015 and 2017, fires caused 131 deaths and CAD 540 million in damage on average per year. In 2018, there were 11,046 fires with 91 deaths and CAD 836.1 million in damage in Ontario. The people responsible for responding to fires are the firefighters. Firefighters have a very limited amount of time (i.e., 5 to 10 minutes) to save as many people as possible and to stop the fire prorogation. To help firefighters in their tasks, many solutions using Information and Communication Technologies (ICT) have been implemented. The advent of smart cities and new technologies, such as the 5G network, the Internet of Things (IoT) and the Building Information Model (BIM), have allowed the creation of innovative solutions. However, these solutions have some shortcomings such as the inability to operate in real time and to adapt to most fire situations. In this dissertation, we propose a new model for real-time fire management. This model exploits distributed systems and microservices architecture to provide a robust, adaptable and scalable system. The proposed model is composed of four modules: The Fire Emergency Sensor Network (FESNet), the distributed application, the client applications and a Web server that acts as an intermediary between the distributed application and the client applications. To validate our model, a prototype able to process the information in real-time and to display the buildings in 3D, was implemented. Moreover, we perform three simulations using the Parallel Merge Sort algorithm.
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 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 | 0.001 | 0.002 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.001 | 0.001 |
| Scholarly communication | 0.003 | 0.002 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.002 | 0.001 |
| Insufficient payload (model declined to judge) | 0.008 | 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; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
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".