In order to realize a flexible automation system, we generally introduce several sensors to monitor states in the system and to handle a ``uncertainty'' such as tolerance, obstacles, position error. When we set up the robot system with sensors in a real factory, we have not so much time to calibrate and set up the sensors. Therefore, we should set up sensors in a quite simple and practical way. In this paper, we developped a simple calibration method for estimating precisely actual reaction forces from the sensor outputs. After executing the proposed calibration method, we achieved to eliminate outliers from sensor outputs in a practical short time.

Imitation learning has attracted attention for its ability to learn human manipulation skills and enable robots to adapt to changes in the environment. In particular, bilateral control-based imitation learning has been proven to be effective in tasks that require force adjustment. However, conventional methods do not consider the relationship between the angle and angular velocity of the robot in training the neural network. Robots with inadequate learning of physical relationships may lead to low task success rates and poor generalizability of their movements. In this study, we proposed a learning method that considers the relationship between angles and angular velocities as a loss function in bilateral control-based imitation learning. In experiments, two tasks were conducted by the conventional and the proposed methods, and the effectiveness was verified.

We have developed a compact and high-stretch ratio robot arm that can avoid interference with crops which are grown in synecoculture^TM. This robot arm has pitch axis rotation by parallel drive of convex. Compared with the existing arm, it expanded the stroke from 300[mm] to 1,200[mm]. Also, the Pitch axis angle has been improved from one-way operation from 0[deg] to $-$90[deg] to bi-directional operation from 25[deg] to $-$90[deg].

In recent years, research aimed at enhancing the sense of immersion in VR environments by inducing human sensations within VR spaces has been actively conducted. This study focuses on the sense of falling, examining the pitch angle direction's rotation angle and angular velocity, as well as the impact of the timing of rotations between the VR space and the real world on the sense of falling. Specifically, the study aims to clarify how adjusting the timing of rotations in the real world relative to the timing of falling rotations in the VR space affects the sense of falling.

In the hierarchical imitation learning model, multiple neural networks (NN) with memory can learn long horizon tasks. However, because the upper and lower layers have memories, independently learnable models may have an inconsistency between the phases of operation predicted by the upper and lower layer models, which may have a negative effect on motion. Since the lower layers are given the current observation and the future predictions, we thought the motion generation is predictable without considering past state. In this paper, we examined the effect on motion and learning time when the lower layer have no memory.

To extend the under-ice exploration in the Arctic Ocean, we propose an exploration method using a peristaltic sea ice drilling robot that communicates and relays at multiple points. We are currently developing a propulsion unit that mimics the peristaltic motion. Sea ice, the drilling target, has low fluidity, and the drilled ice hole diameter is difficult to change according to the shape of propulsion unit. Therefore, the drilling robot needs to adapt the drilled ice hole diameter. In this paper, a propulsion unit with a highly compliant air cylinder is developed and the characteristics of the propulsion unit are evaluated.

When humanoid robots move within human living environments, it is crucial for safety that humans can promptly react to the robots' movements. In this study, we investigated the impact of vertical oscillations of a robot's upper body, designed to mimic human walking motion, compared with cosine wave motion and a fixed condition, on human reaction time to the robot's movement initiation and cessation in a virtual environment. The results revealed that vertical oscillation enabled quicker human reactions to the robot's movement initiation and cessation than when there was no oscillation.

In this paper, we propose a peristaltic mixing conveyor for continuous mixing and conveying of powder materials and solid-liquid mixtures in the production of food and pharmaceutical products, which generates complex low-friction flow in axial and radial directions by unsymmetric blockage in the radial direction of the pipe. In conventional devices mix the liquid penetrates the solid by squeeze flow in the axial direction within the device, and there are conditions that prevent mixing from progressing. On the other hand, the proposed device is expected to accelerate mixing by generating a combined axial and circumferential flow.