Search

To be seamlessly integrated in human-centered environments, robots are expected to have intelligent social capabilities on top of their physical abilities. To this end, research in artifi-cial intelligence and human-robot interaction face two major challenges. Firstly, robots need to cope with uncertainty during interaction, especially when dealing with factors that are not fully observable and hard to infer (latent variables) such as the states representing the dynamic environment and human behavior (e.g., intents, goals, preferences). Secondly, robots need to communicate their behaviors to agents (humans and other robots in the environment) in a clear and understandable manner. Therefore, robots need to be context-aware: being able to perceive and understand their surroundings, and adapt their functionalities accordingly.

The human gestures occur spontaneously and usually they are aligned with speech, which leads to a natural and expressive interaction. Speech-driven gesture generation is important in order to enable a social robot to exhibit social cues and conduct a successful human-robot interaction. In this paper, the generation process involves mapping acoustic speech representation to the corresponding gestures for a humanoid robot. The paper proposes a new GAN (Generative Adversarial Network) architecture for speech to gesture generation. Instead of the fixed mapping from one speech to one gesture pattern, our end-to-end GAN structure can generate multiple mapped gestures patterns from one speech (with multiple noises) just like humans do. The generated gestures can be applied to social robots with arms. The evaluation result shows the effectiveness of our generative model for speech-driven robot gesture generation.

Many Advanced Driver Assistance Systems (ADAS) and chassis systems can be found within the same passenger car. Most of these systems are controlled separately using standalone controllers. Several studies show the benefits of coordinating several systems especially in simultaneous operations. However, most of the studies tend to ignore the time-delays that exist in the control loop. This time-delays can destabilize the vehicle in severe situations. This paper provides a solution to face time-delays in the context of integrated ADAS and chassis systems. Results show the benefits of the control logic proposed in terms of stability and performance. This brings one step closer towards standardization of global vehicle motion control.

International audience; Most of recent researches on the automotive field focus on autonomous vehicles. These vehicles are equipped with conventional chassis systems. The goal is to control the vehicle's traction, brakes, and front steering. This paper discusses the importance of advanced chassis systems, as driving/braking torque vectoring, for both autonomous and non-autonomous vehicles, especially in a race mode. Reliable co-simulation results show that expanding the vehicle's potential leads to high performances and safety with respect to severe situations when optimal control allocation is ensured. Therefore, future passenger cars shall not only be equipped with additional sensors, but also by advanced systems along with adequate control algorithms.

International audience; Several systems are integrated in passenger cars. Some of them are just redundant systems due to safety requirements. Others, are completely different and can interact with each other as long as they are operating inside the same vehicle. Control allocation methods have been successfully implemented in advanced aircrafts to avoid conflicts, especially in the context of redundant systems. In this paper, we will rather focus on coordinating non-redundant advanced chassis systems with different dynamics. This difference in dynamics can be especially problematic when systems exhibit different communication delays. Model Predictive Control Allocation (MPCA) methods are therefore investigated in order to activate the right system at the right moment. Results show that particularly when the most effective system is saturated, another system with a different time delay can be activated few steps before saturation to instantly take over the maneuver. With good knowledge of actuator dynamics and higher computation power, MPCA methods are able to solve complex problems in severe situations.

International audience; Many car manufacturers are equipping their vehicles with advanced systems to win the race towards autonomous vehicles. In order to ensure multiple objectives, different embedded systems influencing differently the same physical variable are often implemented in the same vehicle. Most of car manufacturers tend to activate one system at a time, especially when the low-level control of actuators is developed by different equipment suppliers, making the inner dynamics uncertain. However, this limits the potential of the vehicle. This paper discusses the feasibility of coordinating these systems when activated at the same time, while ensuring an acceptable robustness regarding the uncertain dynamics. Results showed good performance in severe maneuvers when combining robust control synthesis and optimization-based control allocation algorithms. Therefore, several embedded systems may be activated at the same time to expand the overall potential of the vehicle and deal with more difficult situations.

International audience; Autonomous vehicles are round the corner. Most of car manufacturers are racing to become a leader in this promising market. In this paper, we focus on one of the remaining challenges to be overcome, namely, the control of vehicle motion when friction varies. Researchers tend to estimate the friction coefficient using complex observers based on simple mathematical equation. We believe that by taken into account some formerly simplified dynamics, as roll dynamics, we can reduce the estimations' errors, and therefore use simple observers. Co-simulations of the control logic with a high-fidelity vehicle model proved the relevance of this method. Results showed that the vehicle can be controlled in severe conditions while the friction changes using only today's passenger cars sensors. Vertical dynamics should therefore be taken into account even for plane motions.

For socially assistive robots, being able to recognize basic human actions is an important capability. The sensors, which are frequently mounted on most recent robots, such as RGB-D and thermal cameras, as well as the advances in deep learning have enabled the research on activity recognition to grow. In this paper, we collected our own dataset of actions in a home-like scenario, which contains thermal imagery in addition to RGB-D data and we proposed a method based on Long-term Recurrent Convolutional Networks (LRCN). We showed that our method has an accuracy comparable with the state-of-the-art. We also proved that thermal information can improve the recognition accuracy. Furthermore, we tested the real-time capability of our system and conducted a real-time experiment with a robot (Pepper robot from Softbank Robotics) so as to investigate the effect of a robot enabled with action recognition capability in a human-robot interaction.

Social robots are expected to be part of everyday life of people. This will generate interactions between humans and robots that may have positive or negative effects on the users. In order to minimize the negative effects and increase robot persuasiveness, robots should behave in an appropriate manner by adapting to their users. How to achieve this adaptation remains a challenge. We propose the usage of the Regulatory Focus Theory, user physiological state, and game-performance information in order to detect user stress and adapt the behavior of the robot. We present a longitudinal experiment conducted with 35 participants in a game-like scenario. The robot was trained for adapting to the regulatory focus of the users and decreasing their stress while they were playing the game. For this reason, we trained the robot with 12 participants with Chronic Promotion State and with 12 participants with Chronic Prevention State. We used a Q-Learning algorithm based on the Regulatory Focus of the participants, user stress, and task performance. The model obtained was tested with 2 groups (6 and 5 participants, respectively) according to their Chronic Regulatory Focus. Results show that our system was able to generate a robot behavior capable of increasing robot persuasiveness and reducing user stress, which is of great importance for social robots.

International audience; Redundancy of integrated systems has always been used in order to increase vehicle's safety. Some new technologies are however too expensive to be redundant. This paper proposes a new way to preserve the safety by enabling a kind of complementarity between different chassis systems. Optimization-based control allocation redistribution algorithms are used in order to find the best way to combine conflicted systems. Results have shown that one system can take over another one when a failure occurs making the control logic fault-tolerant and reconfigurable. This suits better future cars as additional chassis systems are intended to be integrated.

International audience; A key factor of a successful Human-Robot Interaction (HRI) is the robot's ability to understand the messages transmitted by a human user, especially their nonverbal and paraverbal components. Our work focuses on the robot's ability to detect if a person is deceitful or honest, based on an evaluation of their physiological state and using only noninvasive techniques, such as thermal and RGB-D imaging. In particular, we currently study and evaluate several physiological manifestations during an experiment consisting of an interrogation conducted by a robot that follows a mock-up crime, then correlate them with the truth values of the participants' answers. The first results are promising.

International audience; — In the context of a worldwide aging population, it is important to find solutions to help the elderly maintain their cognitive functions. This research was done in the context of the ENRICHME. We investigate the outcome of a 5 day intensive cognitive stimulation with an elderly individual. Each day was composed of two sessions (one at 11am, and one at 3pm). During each session the participant played three cognitive games (i.e., digit cancellation, integer matrix task, Stroop game), two of them having two difficulty levels. The mood of the participant was also recorded before and after each interaction session. Evidence was found that even after a few sessions, the performance of the participant increased for all games. The performance of each game and each difficulty level was analyzed based on the interaction time (11am or 3pm) and the interaction day. A detailed analysis of the performance is presented together with a discussion of these results.

International audience; Most of the chassis systems are developed by automotive suppliers to improve a specific vehicle performance. Drivers, and consequently vehicle manufacturers, are more concerned by the overall behaviour of the Vehicle. Many coordination architectures have been proposed in the literature in order to integrate different chassis systems in a single vehicle. In this paper , these architectures are compared and discussed. Two major classes are proposed: Downstream and Upstream Coordination. The purpose of this classification is to help car manufacturers and suppliers standardize an Integrated Vehicle Dynamics Control architecture for faster and more flexible designs.

In order to promote reuse of software in robotics, standardization, benchmarking, and formalization activities in robotics are being undertaken by many technical working groups and independent agencies. Absence of integrated tools is the real barrier that exists between early adopters of such efforts and early majority of research and industrial community. In this paper, we provide a detailed discussion on how frameworks for designing robotic systems can be formally specified and developed using our meta-framework - SafeRobots. We have demonstrated this process using a mobile robot case study.

A typical feature of robotic architectures are its reactivity and self-adaptivity. In practice, this is achieved by context-dependent dynamic invocation of software components in robotic architectures. In this paper, we specifically address how this self-adaptation capability can be formally defined and modeled in an architecture-independent way. We propose a probabilistic approach that facilitates system design and dynamic runtime adaptation satisfying the quality requirements. We also show how such techniques are incorporated in our model-driven framework: Self Adaptive Framework for Robotic Systems.

A common goal of robot control architecture designers is to create systems that are sufficiently generic to be adapted to different robot hardware. Beyond code re-use from a software engineering standpoint, having a common architecture could lead to long-term experiments spanning multiple robots and research groups. This paper presents a first step toward this goal with HBBA, a Hybrid Behavior-Based Architecture first developed on the IRL-1 humanoid robot and integrating an Adaptive Resonance Theory-based episodic memory (EM-ART). This paper presents the first step of the adaptation of this architecture to two different robots, a Meka M-1 and a NAO from Aldebaran, with a simple scenario involving learning and sharing objects' information between both robots. The experiment shows that episodes recorded as sequences of people and objects presented to one robot can be recalled in the future on either robot, enabling event anticipation and sharing of past experiences.

In this paper, a model-based framework called `Self Adaptive Framework for Robotic Systems (SafeRobots)' for designing robotic software architectures is proposed. With the help of an example of designing an architecture for mobile robot mapping system, it will be shown how formal domain knowledge can be used at multiple abstraction levels in a model-driven engineering approach. The solution pool for a particular problem is modeled using Solution Space Modeling Language. An eclipse based tool-suite is developed for modeling the solution space and for deriving a concrete operational model by explicitly defining the non-functional properties of the constituent components.

International audience; In human-robot interaction, a social intelligent robot should be capable of understanding the emotional internal state of the interacting human so as to behave in a proper manner. The main problem towards this approach is that human internal states can't be totally trained on, so the robot should be able to learn and classify emotional states online. This research paper focuses on developing a novel online incremental learning of human emotional states using Takagi-Sugeno (TS) fuzzy model. When new data is present, a decisive criterion decides if the new elements constitute a new cluster or if they confirm one of the previously existing clusters. If the new data is attributed to an existing cluster, the evolving fuzzy rules of the TS model may be updated whether by adding a new rule or by modifying existing rules according to the descriptive potential of the new data elements with respect to the entire existing cluster centers. However, if a new cluster is formed, a corresponding new TS fuzzy model is created and then updated when new data elements get attributed to it. The subtractive clustering algorithm is used to calculate the cluster centers that present the rules of the TS models. Experimental results show the effectiveness of the proposed method.

A social intelligent robot should be capable of observing and understanding the changes in the environment so as to behave in a proper manner. It also needs to take into account user preferences, user disability level, and user profile. This paper presents a research work based on socially assistive robotics (SAR) technology that aims at providing affordable personalized physical and cognitive assistance, motivation, and companionship to users. The work described here tries to validate that a robotic system can adapt its behavior to the user profile.

International audience; The human perception of the external world appears as a natural, immediate and effortless task. It is achieved through a number of "low-level" sensory-motor processes that provide a high-level representation adapted to complex reasoning and decision. Compared to these representations, mobile robots usually provide only low-level obstacle maps that lack such high- level information. We present a mobile robot whose goal is to autonomously explore an unknown indoor environment and to build a semantic map containing high-level information similar to those extracted by humans and that will be rapidly and easily interpreted by users to assess the situation. This robot was developed under the1 Panoramic and Active Camera for Object Mapping (PACOM) project whose goal is to participate2 in a French exploration and mapping contest called CAROTTE . We will detail in particular how we integrated visual object recognition, room detection, semantic mapping, and exploration. We demonstrate the performances of our system in an indoor environment.

International audience; Human-robot interaction (HRI) takes place espe-cially through interfaces. The design of such interfaces is avery delicate and crucial phase because it influences the robotaccessibility and usability by the user. In this paper, we describeand analyze the results of 2 tests conducted so as to understandsome of the optimal features that should characterize the robotvoice and graphical-based user interfaces. Our test platform isan assistive robot developed for the elderly with mild cognitiveimpairments. Therefore, the user interfaces must be clear andsimple. The ambiguities must be eliminated so as to facilitatethe use of the robot and hence not to discourage the elderlypopulation to use new technologies.

This research focuses on studying the possible role of a socially interactive robot as a tool for monitoring and encouraging cognitive activities of the elderly and/or individuals suffering from dementia. One of the aims of this work is to show the benefits of the robot's physical embodiment in human-robot social interactions. The social therapist robot tries to provide customized cognitive stimulation by playing a music game with the user. The results of the 8-month pilot study depict a more efficient, natural, and preferred interaction with the robot rather than with the simulated robot.

In less than 25 years, 37 million people are expected to have dementia and/or Alzheimer’s disease. This significant figure is a result of several factors, including increasing life expectancy and an ageing population in developing countries. Alzheimer’s disease, a form of dementia, is a progressive brain disorder that causes memory loss, and behavior and personality changes. People suffering from this disease become increasingly physically and cognitively impaired; they usually need help with all aspects of daily living. While there is currently no cure for dementia, special therapies (e.g., arts and crafts, music therapy) can help people with dementia to live as good a life as possible. We propose to use socially assistive robots as customized/individualized helpers as a complement to humans. This work proposes a new adaptive socially assistive robotic (SAR) system that aims to provide a customized protocol through motivation, encouragement, and companionship for users suffering from cognitive changes related to aging and/or Alzheimer’s disease. The robot aims to maintain/improve the user’s cognitive attention in a cognitive music game setup.

Currently the 2 percent growth rate for the world's older population exceeds the 1.2 rate for the world's population as a whole. By 2050, the number of individuals over the age 85 is projected to be three times more than there is today. Most of these individuals will need physical, emotional, and cognitive assistance. In this paper, we present a new adaptive robotic system based on the socially assistive robotics (SAR) technology that tries to provide a customized help protocol through motivation, encouragements, and companionship to users suffering from cognitive changes related to aging and/or Alzheimer's disease. Our results show that this approach can engage the patients and keep them interested in interacting with the robot, which, in turn, increases their positive behavior.

The world population is growing older. Global aging is occurring at a rate never seen before, it is expected that in 2040 there will be three times more people over the age 85 than there are today. Even if nursing homes and other care facilities can provide assistance, space and staff shortages are already becoming an issue. One of the aims of this work is to show the benefits of robot's physical embodiment in human-robot social interactions. The social therapist robot tries to provide customized cognitive stimulation by playing a music game with the participant. The results of the pilot study over a 6 month period depict a more efficient, natural, and preferred interaction with the robot rather than with the simulated robot.

IEEE International Conference on Robotics and Automation (ICRA 2008, Pasadena, California, May 19-23, 2008), pp. 538-543., This paper presents a vision-based approach for mobile robot localization. The environmental model is topological. The new approach uses a constellation of different types of affine covariant regions to characterize a place. This type of representation permits a reliable and distinctive environment modeling. The performance of the proposed approach is evaluated using a database of panoramic images from different rooms. Additionally, we compare different combinations of complementary feature region detectors to find the one that achieves the best results. Our experimental results show promising results for this new localization method. Additionally, similarly to what happens with single detectors, different combinations exhibit different strengths and weaknesses depending on the situation, suggesting that a context-aware method to combine the different detectors would improve the localization results., This work was partially supported by USC Women in Science and Engineering (WiSE), the FI grant from the Generalitat de Catalunya, the European Social Fund, and the MID-CBR project grant TIN2006-15140-C03-01 and FEDER funds and the grant 2005-SGR-00093.

This paper describes a hands-off therapist robot that monitors, assists, encourages, and socially interacts with post-stroke users in the process of rehabilitation exercises. We developed a behavior adaptation system that takes advantage of the users introversion-extroversion personality trait and the number of exercises performed in order to adjust its social interaction parameters (e.g., interaction distances/proxemics, speed, and vocal content) toward a customized post-stroke rehabilitation therapy. The experimental results demonstrate the robot's autonomous behavior adaptation to the user's personality and the resulting user improvements of the exercise task performance.

In visual-based robot navigation, panoramic vision emerges as a very attractive candidate for solving the localization task. Unfortunately, current systems rely on specific feature selection processes that do not cover the requirements of general purpose robots. In order to fulfill new requirements of robot versatility and robustness to environmental changes, we propose in this paper to perform the feature selection of a panoramic vision system by means of the saliency-based model of visual attention, a model known for its universality. The first part of the paper describes a localization system combining panoramic vision and visual attention. The second part presents a series of indoor localization experiments using panoramic vision and attention guided feature detection. The results show the feasibility of the approach and illustrate some of its capabilities.

Even today, robot mapping is one of the biggest challenges in mobile robotics. Geometric or topological maps can be used by a robot to navigate in the environment. Automatic creation of such maps is still problematic if the robot tries to map large environments. This paper presents a new method for incremental mapping using fingerprints of places. This type of representation permits a reliable, compact, and distinctive environment-modeling and makes navigation and localization easier for the robot. Experimental results for incremental mapping using a mobile robot equipped with a multi-sensor system composed of two 180/spl deg/ laser range finders and an omni-directional camera are also reported.