Takato Horii

Throwing Skill Optimization through Synchronization and Desynchronization of Degree of Freedom

Humanoid robots have a large number of degrees of freedom (DoFs), therefore motor learning by such robots which explore the optimal parameters of behaviors is one of the most serious issues in humanoid robotics. In contrast, it has been suggested that humans can solve such a problem by synchronizing many body parts in the early stage of learning, and then desynchronizing their movements to optimize a behavior for a task. This is called as ”Freeze and Release.” We hypothesize that heuristic exploration through synchronization and desynchronization of DoFs accelerates motor learning of humanoid robots. In this paper, we applied this heuristic to a throwing skill learning in soccer. First, all motors related to the skill are actuated in a synchronized manner, thus the robot explores optimal timing of releasing a ball in one-dimensional search space. The DoFs are released gradually, which allows to search for the best timing to actuate the motors of all joints. The real robot experiments showed that the exploration method was fast and practical because the solution in low-dimensional subspace was approximately optimum.

Compensation for tactile hysteresis using Gaussian process with sensory Markov property

Flexible tactile sensors have been studied to enable robots to interact with objects in unstructured environments. However, due to nonlinearity caused by the hysteresis of tactile materials, it is difficult to accurately convert sensor signals into task-relevant information such as force and slip. To compensate for the hysteresis of flexible tactile sensors, we propose a model based on a Gaussian process. The key idea of our model is to include the Markov property of sensory input. The proposed model not only uses the current tactile signal, but also its time-series signals, to extract the influence of the past states on the current state. We evaluate the accuracy of force estimation using the proposed model in comparison to the normal Gaussian process model, which does not take the Markov property into account. The experimental results demonstrate that the performance of our model improves on the normal Gaussian process in terms of root mean squared error, correlation coefficient, and absolute maximum error between the actual and the estimated force. We discuss the advantages of accounting for the sensory Markov property and the potential ability of the Gaussian process to internally acquire the representation of the deviation of sensory signals.

Design of an Articulation Mechanism for an Infant-like Vocal Robot “Lingua”

Spoken language is one of the important means for humans to communicate with others. In developmental psychology, it is suggested that an infant develops it through verbal interaction with caregivers by observation experiments [1]. However, what kind of underlying mechanism works for that and how caregiver’s behavior affects on this process has not been fully investigated yet since it is very difficult to control the infant vocalization. On the other hand, there are several constructive approaches to understand the mechanisms by using infant robots with abilities equivalent to those of human infants, as a controllable platform [2].

Mexican-Hat-Like Response in a Flexible Tactile Sensor Using a Magnetorheological Elastomer

A significant challenge in robotics is providing a sense of touch to robots. Even though several types of flexible tactile sensors have been proposed, they still have various technical issues such as a large amount of deformation that fractures the sensing elements, a poor maintainability and a deterioration in the sensitivity caused by the presence of a thick and soft covering. As one solution for these issues, we proposed a flexible tactile sensor composed of a magnet, magnetic transducer and dual-layer elastomer, which consists of a magnetorheological and nonmagnetic elastomer sheet. In this study, we first investigated the sensitivity of the sensor, which was found to be high (approximately 161 mV/N with a signal-to-noise ratio of 42.2 dB); however, the sensor has a speed-dependent hysteresis in its sensor response curve. Then, we investigated the spatial response and observed the following results: (1) the sensor response was a distorted Mexican-hat-like bipolar shape, namely a negative response area was observed around the positive response area; (2) the negative response area disappeared when we used a compressible sponge sheet instead of the incompressible nonmagnetic elastomer. We concluded that the characteristic negative response in the Mexican-hat-like response is derived from the incompressibility of the nonmagnetic elastomer.

Imitation of human expressions based on emotion estimation by mental simulation

Abstract Humans can express their own emotion and estimate the emotional states of others during communication. This paper proposes a unified model that can estimate the emotional states of others and generate emotional self-expressions. The proposed model utilizes a multimodal restricted Boltzmann machine (RBM) —a type of stochastic neural network. RBMs can abstract latent information from input signals and reconstruct the signals from it. We use these two characteristics to rectify issues affecting previously proposed emotion models: constructing an emotional representation for estimation and generation for emotion instead of heuristic features, and actualizing mental simulation to infer the emotion of others from their ambiguous signals. Our experimental results showed that the proposed model can extract features representing the distribution of categories of emotion via self-organized learning. Imitation experiments demonstrated that using our model, a robot can generate expressions better than with a direct mapping mechanism when the expressions of others contain emotional inconsistencies.Moreover, our model can improve the estimated belief in the emotional states of others through the generation of imaginary sensory signals from defective multimodal signals (i.e., mental simulation). These results suggest that these abilities of the proposed model can facilitate emotional human–robot communication in more complex situations.

Decision-Making in Emotion Model

Having emotions is essential for robots to understand and sympathize with the feelings of people. In addition, it may allow the robots to be accepted into human society. The role of emotions in decision-making is another important perspective. In this paper, a model of emotions based on various neurological and psychological findings that are related to empathic communication between humans and robots is proposed. Subsequently, a mechanism of decision-making that is based on affects using convolutional LSTM and deep Q-network is examined.

Contact force estimation from flexible tactile sensor values considering hysteresis by Gaussian process

Flexible tactile sensors are important elements for facilitating the physical interaction between robots and uncertain environments. For instance, tactile information is used by the robot to grasp objects and interact with humans. A model-based approach is one technique for building a relationship between tactile sensor values and task-relevant information such as force, slip, and temperature. However, it is difficult to create models of flexible tactile sensors for converting sensor signals beforehand due to a nonlinear relation between a contact and the deformations of the flexible form caused by its hysteresis [1]. In contrast, machine learning techniques can be adopted to represent these relationships. For example, Tada et al. [2] proposed a model to acquire the relationship between tactile sensor values and slip vibration using a neural network. The purpose of this study was to develop computational models for learning the association between the force applied to a tactile sensor and the sensor value by compensating for the hysteresis in the sensor. We used the tactile sensor of an iCub fingertip in order to apply our models to cognitive studies. This paper first presents our proposed models that consider a Markov property of taxel (tactile sensor elements) values, and then reports experimental results.

How does emphatic emotion emerge via human-robot rhythmic interaction?

Sharing a same rhythm among different agents matures close empathic relationship. However it is still unclear how rhythmic information is transferred to emphatic emotion in our brain. In this paper, we propose human-robot drumming interaction system to investigate how rhythmic physical interaction brings emphatic emotion. In this paper, we introduce the detail of our proposed system. ACM Classication