Posts from category "réalisation"

Stream Analysis and Filtering for Reliability and Post-processing of Sensor Big data

Context Presentation

Anomaly detection in real fluid distribution applications is a difficult task, especially, when we seek to accurately detect different types of anomalies and possible sensor failures. Resolving this problem is increasingly important in building management and supervision applications for analysis and supervision. Our case study is based on a real context: sensor data from the SGE (Rangueil campus management and operation service in Toulouse).

We propose CoRP” Composition of Remarkable Points” a configurable approach based on pattern modelling, for the simultaneous detection of multiple anomalies. CoRP evaluates a set of patterns that are defined by users, in order to tag the remarkable points using labels, then detects among them the anomalies by composition of labels. CoRP is evaluated on real datasets of SGE and on state of the art datasets and is compared to classical approaches.

image044

Figure 1: « Anomaly Detection in Sensor Networks »

Scientific Goals

- Detect different types of anomalies observed in real deployment

- Improve the supervision of sensor networks

- Use learning methods for anomaly detection on static and continuous data

 

Keywords

neOCampus, Sensor Data, Univariate Time Series, Anomaly Detection, Pattern-based Method

Contacts

Ines.ben-kraiem@irit.frOlivier.teste@irit.fr, andre.peninou@irit.fr

Real-time distributed optimization of energy management in smart grids

Context Presentation

RennesGrid is an energy transition project in the Ker Lann business park in Bruz. In particular, this project focuses on self-consumption integrating photovoltaic panels, storage devices and energy data collection. As part of this project, this thesis aims to implement a multi-agent system managing the consumption of flexible loads, particularly electric vehicles, and the production of power sources (photovoltaics).The smart grid concept is driving an explosion in the number of controllable units (flexible loads, decentralized producers, storage units, etc.). In addition, issues related to energy management in the smart grid, whether local (voltage control at bus level, congestion control) or global such as managing the balance between consumption and production, make the problem strongly linked.

The flexibility of adaptive multi-agent systems is relevant to this issue. Indeed, it enables to manage a dynamic environment (consumption, production, power grid...). It is also open and robust. Thus, it is able to adapt to the ever-increasing energy demand and the need to keep the power grid in service, particularly when an incident happens.

image043

Figure 1: Smart grid concept

Scientific Goals

- The realization of a micro grid simulator and a scenario generator

- The design and evaluation of an adaptive multi-agent system managing a micro grid

Keywords

neOCampus, smart grid, optimization, multi-agent systems

Contacts

Jean-Baptiste.Blanc-Rouchosse@irit.fr

Impact of spatial strategies of bees on colony performance

Context Presentation

Foraging for food to substantiate one’s needs is of great importance for every species. In the case of bees, who are a social species, only a small selection of individuals has the task to bring the food for the whole colony, and thus has to take into account the needs of the entire population in terms of nutrients. As central place foragers, bees will explore and exploit flowers around their nest, where different species provide bees with different amounts and qualities of nectar. Bees are as a result faced with a complex problem: finding flowers that are not already exploited by other bees, which provide the nutrients in the right amount (either by foraging on a single species of flowers with a balanced diet, or on multiple species with unbalanced but complementary diets), and create a stabilized exploitation route between them. Following each individual bee in its foraging trip has been a technological challenge. However, today, as different tracking technologies (radars, camera tracking) are being developed, assisted with colony monitoring systems (connected hives), we can finally get some insights on these complex behaviors. As data are still scarce and only available in limited, simplified situations, building theoretical models that successfully replicate the spatial strategies of bees will allow us to make predictions on more complex and ecologically relevant scenarios.

Scientific Goals

- Conduct experimental tests for the fundamental hypotheses of the behavior.

- Build a new model based on experimental tests of simple situations and theoretical knowledge of bee foraging behavior.

- Test the model’s predictions in complex environmental situations.

Keywords

Spatial strategy, foraging behavior, nutritional geometry, connected hive

Contacts

thibault.dubois@univ-tlse3.frmathieu.lihoreau@univ-tlse3.fr

RECOVAC: conditions for REtaking COntrol by self-observation of situations within a connected autonomous vehicle

Context Presentation

Connected autonomous vehicles of level 3 are vehicles in which the human driver delegates driving control in specific situations. During these situations, it may be necessary for the human to regain control of the driving activity. The main objective of this thesis is to develop a system for the safe and efficient transition of two-way control between the human and the autonomous vehicle.

For this, the system must identify by self-observation and in real time situations in which the current driver will no longer be able to ensure driving. He must also provide a context for assessing the criticality of the situation as quickly as possible in order to anticipate and react to it as best as possible. The driving context is composed of indicators that characterize the elements that describes part of the driving process. The system is based on self-adaptive multi-agent learning systems.

image037

Figure 1: Dynamic Learning using self-adaptive multi-agent system

Scientific Goals

- Dynamic learning using multi-agent systems

- Generic approach to supervise the activity of a system

- Insure the acceptability of the system by human driver

Keywords

self-adaptive multi-agent systems, autonomous vehicle

Contacts

kristell.aguilar-alarcon@irit.frmarie-pierre.gleizes@irit.frloic.caroux@univ-tlse2.fr

Fleet of Autonomous and Connected Vehicles by Adaptive Multi-Agent System

Context Presentation

The theme addressed in this thesis concerns the autonomous and connected vehicle. Its reliability must be proven from the technological point of view, which is what most projects focusing on technological developments answer. It is essential, after having made a vehicle reliable, to study how several connected autonomous vehicles will be able to interact in order to maximize the safety of the collective. The American Society of Automotive Engineers (SAE International) has defined 5 levels of autonomy: from level 1, in which the driver performs all the maneuvers, to level 5, in which the driver is completely autonomous. This thesis concerns both vehicles with a level of autonomy 4 in which the vehicle drives, and the human driver supervises and can take over the control, as well as vehicles with a level 5 (total vehicle autonomy).

In this thesis, the aim is to study how each vehicle communicates with its neighboring vehicles and how it behaves/reacts in response to the information provided by its neighborhood. A lock consists in particular in determining which information is relevant to communicate among all the data recovered from the many sensors scattered in the vehicle and the effectors.

image036

Figure 1: « A fleet of autonomous and connected vehicles »

Scientific Goals

- Determine which information is relevant

- How to adapt to each traffic situation

- Take the non-connected users in account

Keywords

Autonomous and connected vehicle, Multi-Agent System, Smart Traffic

Contacts

Guilhem.marcillaud@irit.fr

Life Cycle Assessment (LCA) of Lighting Systems: Environmental, Economics and Human Impacts Analysis

Context Presentation

When it comes to identifying and measuring the quantifiable effects of products or services on the environment, Life Cycle Assessment (LCA) is probably the most powerful and recognized tool. Thanks to a multicriterion and a cradle-to-grave approach, LCA identifies and quantifies, throughout the life of products, the physical flows of matter and energy associated with human activities (extraction of raw materials required for the manufacture of the product, distribution, use, collection and disposal to end-of-life systems and all phases of transport). For each of its flows, there are impact indicators that establish the overall potential impact of the system on our environment.

During past years, smart lighting technologies allowed significant improvements regarding lamp efficiency during use phase (from 19% to 15% of global electricity consumption), nevertheless, there are direct or indirect impacts on our environment, health, well-being or productivity not taken into account into Life Cycle Assessment (LCA) studies, and we can’t no longer neglected them.

image035

Figure 1: Impacts assessment of lighting systems

Scientific Goals

- How to extend LCA methodology in order to determine which lighting system is most performant regarding environmental, economic and social aspect?

- How using phase could impact on lamp overall performance (Light Loss Factor, Mean Time Before Failure and Maintenance Factor)?

- Which criteria should be used to reflect lighting impact on human health or ecosystems during use phase?

Keywords

Lighting systems, Life Cycle Assessment, Circadian effect, Life cycle Cost, Multicriterion analysis.

Contacts

kevbertin@gmail.com – bertin@laplace.univ-tlse.frEncadrants : georges.zissis@laplace.univ-tlse.fr , marc2.mequignon@free.fr

Animal Minds (OpenFeeder)

Context Presentation

- Study the behavior and cognition of titmouse in their natural environment using an electronic feeder, called an Openfeeder.

- Developed by SETE (Station of Theoretical and Experimental Ecology) and SelectDesign.

- System successfully deployed as an island (4 to 8 feeders) on 2 high altitude sites and 3 low altitude sites around SETE (fall 2018).

- A feeder = PIR sensor (detect the presence of a bird), RFID reader (identification), a door controlled by a servomotor. The bird is banded (a transponder), a software with several programmed cognitive task scenarios.

- Principle of operative conditioning (learning a stimulus/reward combination).

- Data collection by USB stick, OF by OF!

image033image034

Figure 1 : an 8 OpenFeeder station

Scientific Goals

- Synchronize the clocks of the OpenFeeder on each station

- Collect data (logs): centralization on an OF

- Transmit all collected data to the Laboratory (SETE), with RF module and GSM module

- Transmit errors and anomalies in real time by SMS via the GSM module

- Transmit config. (cognitive task scenarios,.ini files)

Keywords

Birdwatch, OpenFeeder, GSM, ALPHA_TRX 433s,

Contacts

kacimi@irit.fr | thiebolt@irit.fr | mcauchoixxx@gmail.comanzilane.mmadi@irit.fr | anzilane.mmadi@univ-tlse3.fr

SDN approach for Pedestrian Protection in Autonomous 5G-VANETs

Context Presentation

The development of self-driving cars is increasing with 5G techniques. One of the biggest challenges posed by this domain is to protect pedestrians and to decrease accidents by detecting them before they occur. That’s why we need to decrease latency, improve reliability, optimize resource allocation and maintain connectivity… In this regard, we have proposed to preview vehicular and pedestrian traffic and send an alert message to warn them of collision risks. To achieve our goals, we started by proposing a network architecture based on an SDN approach, cell-less configuration, and decentralized computing nodes... Then we noticed that if all vehicles and pedestrians are going to communicate with the controller to send their position, the OpenFlow signaling is going to increase in the controller. So, we have simulated the up-link traffic and we have shown the interest of relieving the overload on the controller by sending position messages just in need. We developed an algorithm that estimate the time interval without future collision risks and decide the frequency of sending position messages in the up-link. Concerning the future work, we have to validate the obtained results with simulation.

image031

Figure 1: Proposed SDN architecture

Scientific Goals

- Generate alert messages under low latency- Improve fiability and throughput- Optimize ressources allocation

Keywords

neOCampus, file, presentation, innovation, VANET, 5G, SDN …

Contacts

Chaima.Zoghlami@irit.fr, Rahim.Kacimi@irit.fr, Riadh.dhaou@irit.fr

Endogenous Learning by Cooperation

Context Presentation

The current digital transformation requires the creation of autonomous applications that can be adapted to complex, dynamic, heterogeneous and unpredictable environments. These systems must be equipped with proactive learning capabilities. To this end, Self-Adaptive Multi-Agent principle allows the decentralization and self-observation of the learning process. Each knowledge granule is an autonomous agent that cooperates with its neighbors to improve learning from exogenous and endogenous feedbacks. Detecting and solving concurrences, conflicts and incompetencies leads to active endogenous learning.This work on an adaptive decentralized learning mechanism will be applied on application domains such as robotics, autonomous vehicles and smart cities.

image030image029

Figure 1: « Schema of the Learning

Figure 2: « Implementation Example Multi-Agent System » on an Industrial Robot »

Scientific Goals

- Design a Self-Learning System

- Lifelong and Endogenous Learning

- Genericity and Scalability

Keywords

Self-Adaptive Learning, Endogenous Learning, Adaptive Multi-Agent Systems, Artificial Intelligence

Contacts

bruno.dato@irit.frfrederic.migeon@irit.fr marie-pierre.gleizes@irit.fr

Automatic traffic generation with multi-agent simulation

Context Presentation

Mobility in cities is a crucial question to improve services offered to users and to decrease pollution. Traffic simulation is an interesting tool for decision-makers for city planning. Decision relevance depends on the realism of the simulation. To ensure that, we suppose that we have a map of the terrain (road topology, buildings…) and observation points that provide traffic information (throughput, direction…). They are in limited number and do not cover all the map.This internship focuses on generating car flows on the whole road network using adaptive multi-agent systems. For that, the system determines the location, speed and direction for each vehicle based on its neighborhood. The behavior of each vehicle self-adjusts to match, at the macro-level, traffic information provided by all observation points.

image028

Figure 1: Example of traffic simulationDynameq © INRO

Scientific Goals

- Self-calibration of simulation parameters for realistic traffic flow generation

- Self-adaptation of the simulation to handle traffic dynamics

Keywords

realistic traffic generation, multi-agent simulation, self-adaptation, cooperation

Contacts

damien.vergnet@irit.frelsy.kaddoum@irit.frmarie-pierre.gleizes@irit.fr

Back to Top