Abstract-Multi-person tracking with a robotic platform is one of the cornerstones of human-robot interaction. Challenges arise from occlusions, appearance changes and a time-varying number of people. Furthermore, the final system is constrained by the hardware platform: low computational capacity and limited field-of-view. In this paper, we propose a novel method to simultaneously track a time-varying number of persons in three-dimensions and perform visual servoing. The complementary nature of the tracking and visual servoing enables the system to: (i) track several persons while compensating for large ego-movements and (ii) visually control the robot to keep a selected person of interest within the field of view. We propose a variational Bayesian formulation allowing us to effectively solve the inference problem through the use of closed-form solutions. Importantly, this leads to a computationally efficient procedure that runs at 10 FPS. The experiments on the NAO-MPVS dataset confirm the importance of using visual servoing for tracking multiple persons.

This paper addresses the problem of sound-source localization (SSL) with a robot head, which remains a challenge in real-world environments. In particular we are interested in locating speech sources, as they are of high interest for humanrobot interaction. The microphone-pair response corresponding to the direct-path sound propagation is a function of the source direction. In practice, this response is contaminated by noise and reverberations. The direct-path relative transfer function (DP-RTF) is defined as the ratio between the directpath acoustic transfer function (ATF) of the two microphones, and it is an important feature for SSL. We propose a method to estimate the DP-RTF from noisy and reverberant signals in the short-time Fourier transform (STFT) domain. First, the convolutive transfer function (CTF) approximation is adopted to accurately represent the impulse response of the microphone array, and the first coefficient of the CTF is mainly composed of the direct-path ATF. At each frequency, the frame-wise speech auto-and cross-power spectral density (PSD) are obtained by spectral subtraction. Then a set of linear equations is constructed by the speech auto-and cross-PSD of multiple frames, in which the DP-RTF is an unknown variable, and is estimated by solving the equations. Finally, the estimated DP-RTFs are concatenated across frequencies and used as a feature vector for SSL. Experiments with a robot, placed in various reverberant environments, show that the proposed method outperforms two state-of-the-art methods.

Abstract-When designing agent-based simulation, the choice of a coordination model is a key issue, since one of the difficulties is to link the activation of the agents with their context efficiently. Current solutions separate the activation phase from the action phase of the agents, and each action phase is based on local agent context analysis which is time-expensive. Moreover, because the link between the context and the action is an internal part of the agent, it is more difficult to modify the way the agent reacts to the context without altering the way the agent is implemented. Our proposal, called EASS (Environment as Active Support for Simulation), is a new approach for agent activation, where the context is analysed inside the environment and conditions the activation of the agents. The main result of contextual activation is to simplify the achievement of complex simulations and to decrease run-time. The EASS model has been implemented within the kernel of MadKit, a multi-agent platform, and the first results are given.

Abstract. Indirect interactions have been shown to be of interest in MultiAgent Systems (MAS), in the simulation area as well as in real applications. The environment is also emerging as a first-order abstraction. Intuitively, the environment being a common medium for the agents, it should be a suitable paradigm to provide a support of both direct and indirect interactions. However, it still lacks of a consensus on how the two relate to each other, and how the environment can support effectively notions as communication or awareness. We propose a general and operational model, Environment as Active Support of Interaction, that enables the agents to actively participate in the definition of their perceptions. Then, we show how the model provides a suitable framework for the regulation of the MAS interactions.

International audienceWe propose a design for collaborative support system in distant tangible environments, in the framework of activity theory. We model collaboration as driven by individual-centered and group-centered rules. Context sharing is core to this process, but reveals difficult in the case of distant tangible communication. We propose to model collaboration as a trace-based process in which tangible object traces are stored, analyzed, enriched and shared. We draw on a normative multi-agent approach in which explicit norms are meant to operate at various levels, from the physical to the social level. These norms do not act as a prerequisite, or as a way to place a priori constraints on action. Rather, they result in a set of signs situating the activity. Such design offers novel ways for embedding activity theory in the current trend of socio-physical computing