Apprentissage de politiques de navigation a partir de demonstration pour de la simulation de traffic

The development of personal robots or service robots has aroused much interest in recent times, witnessing an amazing growth of robotics in different domains. Although sophisticated humanoid robots have been developed, much more effort is needed for improving their cognitive capabilities. Interactions with humans and / or with other agents are still limited and not considered satisfactory. So, the way we store and represent knowledge in a cognitive architecture(CA) is fundamental in order to overcome these limitations and improve human-machine and machine-machine interactions. In this paper we propose a multimedia graph knowledge base using pure formalisms, which can be used by personal robots to represent and share the acquired knowledge, linking it to abstract concepts gifted with semantic relations. Moreover we propose a prototype architecture for sharing the knowledge between two agents following the Knowledge-as-a-service (KaaS) paradigm.

In this letter, a novel morphological band selection method is proposed to obtain the most representative bands from hyperspectral image (HSI) in an unsupervised manner. In order to sufficiently process the HSI, we propose to use only a small set of data instead of using the original full data. For the obtained clusters, the differences of spectral response curves are applied to measure the local discrimination capability of bands, of which the local maximum value point is yielded based on the morphological processing. To verify the performance of the proposed method, the robustness of the parameters has been evaluated, while the effectiveness and superiority have been tested on three popular hyperspectral data sets. The experiment results have shown that the proposed method outperforms other methods.

In order to alleviate the negative effect of curse of dimensionality, band selection is a crucial step for hyperspectral image (HSI) processing. In this letter, we propose a novel unsupervised band selection approach to reduce the dimensionality for hyperspectral imagery. In order to obtain the most representative bands, the correlation matrix computed from the original HSI is used to describe the correlation characteristics among bands, while the block-diagonal structure is measured to segment all bands into a series of subspace. After applying the spectral clustering algorithm, the optimal combination of band is finally selected. To verify the effectiveness and superiority of the proposed band selection method, experiments have been conducted on three widely used real-world hyperspectral data. The results have shown that the proposed method outperforms other methods in HSI classification application.

Artificial awareness is an interesting way of realizing artificial intelligent perception for machines. Since the foreground object can provide more useful information for perception and informative description of the environment than background regions, the informative saliency characteristics of the foreground object can be treated as a important cue of the objectness property. Thus, a sparse reconstruction error based detection approach is proposed in this paper. To be specific, the overcomplete dictionary is trained by using the image features derived from randomly selected background images, while the reconstruction error is computed in several scales to obtain better detection performance. Experiments on popular image dataset are conducted by applying the proposed approach, while comparison tests by using a state of the art visual saliency detection method are demonstrated as well. The experimental results have shown that the proposed approach is able to detect the foreground object which is distinct for awareness, and has better performance in detecting the information salient foreground object for artificial awareness than the state of the art visual saliency method.

In this paper, we present a cognitive system allowing a robot to discover the world, in which it evolves, and to learn autonomously high level knowledge about it from low level sensor data. The learning is accomplished both by observation and by an active interaction with a human, in which the robot is the actor of the learning. We first describe the system in a general manner and we provide experimental results on a humanoid robot in a real-world environment using everyday objects. We show, that our approach allows a humanoid robot to learn successfully without negative input and from small number of samples.

In this paper, we propose an experimental validation of a new control strategy based on the use of intuitive control and neural networks CMAC in order to control the under-actuated robot RABBIT. The training of the neural networks is carried out during the simulation. When the training is finished, these neural networks are used to control the real robot. We also present an evaluation of the robustness of this control strategy.

In this paper, we propose a control strategy allowing us to perform the transition of velocities included in [0 m/s; 1 m/s] for the dynamic walking of an under-actuated robot (RABBIT) without reference trajectories. This strategy of control also enables us to carry out the transition from stop towards walking and the reverse process. The interest of this method resides in the fact that on one hand the intrinsic dynamics of the system are exploited by using a succession of active and passive phases and on the other hand that the control strategy is very easy to implement on-line.