We propose a practical state-of-the-art method to develop a machine-learning-based humanoid robot that can work as a production line worker. The proposed approach provides an intuitive way to collect data and exhibits the following characteristics: task performing capability, task reiteration ability, generalizability, and easy applicability. The proposed approach utilizes a real-time user interface with a monitor and provides a first-person perspective using a head-mounted display. Through this interface, teleoperation is used for collecting task operating data, especially for tasks that are difficult to be applied with a conventional method. A two-phase deep learning model is also utilized in the proposed approach. A deep convolutional autoencoder extracts images features and reconstructs images, and a fully connected deep time delay neural network learns the dynamics of a robot task process from the extracted image features and motion angle signals. The “Nextage Open” humanoid robot is used as an experimental platform to evaluate the proposed model. The object folding task utilizing with 35 trained and 5 untrained sensory motor sequences for test. Testing the trained model with online generation demonstrates a 77.8% success rate for the object folding task.

In this paper, we propose a sophisticated design of human symbiotic robots that provide physical supports to the elderly such as attendant care with high-power and kitchen supports with dexterity while securing contact safety even if physical contact occurs with them. First of all, we made clear functional requirements for such a new generation robot, amounting to fifteen items to consolidate five significant functions such as ldquosafetyrdquo, ldquofriendlinessrdquo, ldquodexterityrdquo, ldquohigh-powerrdquo and ldquomobilityrdquo. In addition, we set task scenes in daily life where support by robot is useful for old women living alone, in order to deduce specifications for the robot. Based on them, we successfully developed a new generation of human symbiotic robot, TWENDY-ONE that has a head, trunk, dual arms with a compact passive mechanism, anthropomorphic dual hands with mechanical softness in joints and skins and an omni-wheeled vehicle. Evaluation experiments focusing on attendant care and kitchen supports using TWENDY-ONE indicate that this new robot will be extremely useful to enhance quality of life for the elderly in the near future where human and robot co-exist.

Distributed tactile sensing is crucial to perform stable, subtle, and precise manipulation so that a robot can recognize and handle objects properly. However, currently existing skin sensors still have common problems such as complex and expensive production or are difficult to integrate into robot hands. In particular, a practical distributed soft skin sensor system that can cover various parts of the robot hand, measure force in 3-axis, with a subcentimeter spatial density, and digital output at the same time does not exist yet. This letter discusses uSkin, a soft, distributed, 3-axis force sensor for robot hands and presents its implementation for multicurved fingertips. The sensor is low-cost, easy to manufacture, and can measure normal and shear forces. The experimental results revealed that this sensor has 10% hysteresis for perpendicular force with a maximum range of 6 N. The Signal to Noise Ratio (SNR) value of 54 dB for 0.4 N load was achieved, which constitutes the state of the art for this kind of sensors. Evaluation experiments also showed that the distributed 3-axis load cells could produce vectors that represent the shape of objects. This opens the possibility that the sensor can be used for classifying different shapes. Furthermore, the fingertip sensor was installed on the Allegro hand and the changing force measurements when the robot is grasping an object are presented.

Tactile sensing is one important element that can enable robots to interact with an unstructured world. By having tactile perception, a robot can explore its environment by touching objects. Like human skin, a tactile sensor that can provide rich information such as distributed normal and shear forces with high density can help the robot to recognize objects. In previous work we introduced uSkin, a soft skin with distributed 3-axis force sensitive elements, and a center-to-center distance between the 3-axis load cells of 4.7 mm for the flat version. The current paper presents a new structure for the distributed soft force transducer which reduces the crosstalk between the components of the 3-axis force measurements. 3D printing the silicon structure eased the prototype production. However, the 3D printed material has a higher hysteresis than the previously used Ecoflex. Microcontroller boards originally developed for the skin of iCub were implemented for uSkin, increasing the readout frequency and reducing the space requirements and number of wires. The sensor was installed on iCub and successfully used for shape exploration.

This paper presents an easy means to produce a 3-axis Hall effect–based skin sensor for robotic applications. It uses an off-the-shelf chip and is physically small and provides digital output. Furthermore, the sensor has a soft exterior for safe interactions with the environment; in particular it uses soft silicone with about an 8 mm thickness. Tests were performed to evaluate the drift due to temperature changes, and a compensation using the integral temperature sensor was implemented. Furthermore, the hysteresis and the crosstalk between the 3-axis measurements were evaluated. The sensor is able to detect minimal forces of about 1 gf. The sensor was calibrated and results with total forces up to 1450 gf in the normal and tangential directions of the sensor are presented. The test revealed that the sensor is able to measure the different components of the force vector.

Recognizing grasped objects with tactile sensors is beneficial in many situations, as other sensor information like vision is not always reliable. In this paper, we aim for multimodal object recognition by power grasping of objects with an unknown orientation and position relation to the hand. Few robots have the necessary tactile sensors to reliably recognize objects: in this study the multifingered hand of TWENDY-ONE is used, which has distributed skin sensors covering most of the hand, 6 axis F/T sensors in each fingertip, and provides information about the joint angles. Moreover, the hand is compliant. When using tactile sensors, it is not clear what kinds of features are useful for object recognition. Recently, deep learning has shown promising results. Nevertheless, deep learning has rarely been used in robotics and to our best knowledge never for tactile sensing, probably because it is difficult to gather many samples with tactile sensors. Our results show a clear improvement when using a denoising autoencoder with dropout compared to traditional neural networks. Nevertheless, a higher number of layers did not prove to be beneficial.