Planification (intelligence artificielle)

Résumé

vignette|Exemple de plan pour un robot qui déplace des blocs : prendre B, poser B sur la table, prendre C, poser C sur A.|alt=|257x257px En intelligence artificielle, la planification automatique (automated planning en anglais) ou plus simplement planification, vise à développer des algorithmes pour produire des plans typiquement pour l'exécution par un robot ou tout autre agent. Les logiciels qui incorporent ces algorithmes s'appellent des planificateurs. La difficulté du problème de planification dépend des hypothèses de simplification qu'on tient pour acquises, par exemple un temps atomique, un temps déterministe, une observabilité complète, etc. La compétition IPC (international planning competition) a lieu tous les ans durant le congrès ICAPS (International Conference on Planning and Scheduling). Principe Un planificateur typique manipule trois entrées décrites dans un langage formel (tel que STRIPS ou PDDL) qui utilise des prédicats logiques :

une description de

Source officielle

https://fr.wikipedia.org/wiki/Planification_(intelligence_artificielle)

À propos de ce résultat

Cette page est générée automatiquement et peut contenir des informations qui ne sont pas correctes, complètes, à jour ou pertinentes par rapport à votre recherche. Il en va de même pour toutes les autres pages de ce site. Veillez à vérifier les informations auprès des sources officielles de l'EPFL.

Publications associées (14)

Publications associées

Chargement

Personnes associées (2)

Personnes associées

Chargement

Unités associées (1)

Unités associées

Chargement

Concepts associés (18)

Concepts associés

Chargement

Cours associés (2)

Cours associés

Chargement

Séances de cours associées (4)

Afficher plus

Séances de cours associées

Chargement

Discrete and Rhythmic Motor Primitives for the Control of Humanoid Robots

Sarah Dégallier Rochat

Controlling robots with multiple degrees of freedom (DOFs) is still an open and challenging issue, notably because planning complex, multidimensional trajectories in time-varying environments is a laborious and costly process. In this dissertation, we propose a control architecture where the planning consists in defining the key characteristics of the desired movement (e.g., the target for reaching), while the generation of the complete trajectory is based on predefined dynamics. More precisely, the generation of joint trajectories is decoupled from high-level planning (i.e., the definition of the task) through the use of a combination of discrete and rhythmic motor primitives, that is, movements with predefined dynamics. The concept of motor primitives is inspired by the study of the motor system of vertebrates: animals are capable not only of performing highly complex tasks in a robust way but also of rapidly adapting to changes or uncertainties in the environment. Interestingly, the planning of movements and the actual generation of trajectories are most likely decoupled in vertebrates: the actual spatio-temporal sequence of activation of the muscles is produced at the spinal level through neural networks called central pattern generators (CPGs). These networks are activated by simple, non-patterned control signals from the brain, that is, only the key parameters of the movement seems to be needed from the brain for a task to be completed. Here we develop a control architecture for the generation of both discrete and rhythmic movements based on motor primitives that has the following attributes (i) the planning phase is simplified thanks to the motor primitives, in the sense that the control commands that are required are reduced to the key characteristics of the movement, (ii) the implementation intrinsically ensures smooth transitions between different tasks (discrete and rhythmic) (iii) the dynamics of the motor primitives can be modulated by sensory feedback in order to have fast adaptive responses and (iv) several DOFs can be coupled together to ensure coordinated behaviors. In addition, this method has a low computational cost and is well-fitted for applications requiring fast control loops. We illustrate the efficiency of the architecture through two applications: (a) interactive drumming with the Hoap2 and the iCub and (b) infant-like crawling and reaching with the iCub.

EPFL2010

Chargement

Gate Optimization: Schedule Planning and Day of Operations

Prem Kumar

This talk focuses on the multi-objective scenario of optimizing airport gates to incoming and outgoing flights of an airline

2010

Chargement

Big-Data streaming applications are used in several domains such as social media analysis, financial analysis, video annotation, surveillance, medical services and traffic prediction. These applications, running on different types of platforms from mobile devices to servers, are characterized by a highly-variable stochastic input data stream, stringent delay constraints and complex task graphs. Several software and hardware optimization techniques have been proposed to maximize the execution quality and the throughput of these applications and to minimize their energy consumption on many-core platforms. By analyzing the existing techniques, one can observe that most solutions classify as a hardware-based task-specific optimization, or as an operating system scheduler optimization, or yet as a load shedding mechanism at the application level. Each of these categories is limited in scope and can be blind to the nature of the program, the data being processed or the characteristics of the hardware. Big-Data streaming applications, due to their wide range of host hardware and content and dynamically-changing input streams, expose the fragmentation of the optimization techniques and create a clear need for a better approach. In this thesis, I propose a suite of energy-efficient hardware-software co-design techniques to bridge the gap between modern Big-Data streaming applications and existing many-core platforms. I choose to model the task graph of the class of applications I consider here by a direct acyclic graph (DAG). First, at the application layer, I propose a unified DAG monitoring solution to process online the general DAG model of the application and provide a set of relevant information that is leveraged at run time by a connected scheduler. At the operating system layer, I propose three different online scheduling solutions for many-core platforms which leverage the feedback from both the application and the hardware layers. The first scheduler addresses the problem of minimizing the energy consumption and the deadlines miss rates of multimedia applications. It takes advantage of the output of the DAG monitoring solution to adapt the mapping of the tasks of multimedia applications to the hardware according to the detected performance and targeted quality of service. The second and third schedulers address the problem of maximizing the quality and throughput and minimizing the energy consumption of Big-Data stream mining applications with single and multiple streams originating from different sources. To cope with the dynamically-changing Big-Data characteristics, the schedulers integrate machine-learning techniques to learn the environment dynamics and the application requirements and adapt the scheduling policy to the desired quality of service. I show that the proposed schedulers are able to scale the execution of data-mining applications to the system capability even in the presence of concept drift. Last, at the hardware layer, to address existing system architectures limitations and to increase even more the throughput, I propose a novel low-power many-core architecture for modern Big-Data stream mining applications that integrates a novel flexible memory hierarchy able to adapt to the dynamic data-driven nature of the input data stream.

EPFL2015

Chargement

EPFL2015

Discrete and Rhythmic Motor Primitives for the Control of Humanoid Robots

Sarah Dégallier Rochat

EPFL2010