In our loboratory, we are researching the following themes that we briefly introduce in this page. In addition, we introduce our previous finished researches.

Violin-playing robot
Four-legged robot
Bipedal robot
Underwater robot inspered by sperm whale
Soft Tactile Sensor
Brain machine interface
Finished researches

Violin-playing robot

Professional violinists have to coordinate their motions of arms and fingers when playing the violin, while they express their emotions through the produced sounds. If we improve those abilities of a robot, the powers of expression also greatly improve, which will lead to better human-robot interaction. Therefore, there are two goals for this research project: building a robot with adequate ability for playing the violin, and improve the powers of expression.

The first goal is simple but difficult. The right arm have to control the bow speed, the bow force and the sounding point (bow-bridge distance), which are three main parameters. The left hand have to control the position of each fingertip to determine pitch of a sound. Recently we have developed an anthropomorphic violin-playing robot (Figs. 1-4), and we succeeded in playing the first four bars of “Go Tell Aunt Rhody”. The robot arm have seven joints and the left hand has four fingers with two joints.

The second goal is also very difficult to achieve. Usually, musical scores does not contain complete information to play musical instruments and human violinists compensate for it. We believe that constructing the procedure will improve the powers of expressiveness. As the first step of this research, we are trying to build an algorithm to determine three main bowing parameters only from the score information.

Fig. 1 Violin-playing robot Fig. 2 Back view Fig. 3 Close up of right arm Fig. 4 Left hand and fingers

Simple bowing motion
Robot plays the first four bars of "Go Tell Aunt Rhody"

Four-legged robot "Ryuma"

Some researches claim that horses adjust its gait pattern according to their locomotion speed to minimize the energy consumption per unit distance for the locomotion, which means that hoses choose the most efficient gait. We want to prove the hypothesis by using a real four-legged robot. In addition, we believe that their head movements affect the locomotion stability. Thus, the goal of this study is to develop a four-legged robot and investigate its energy efficiency and effects of head movements.

We have developed several robots as shown in Figs.5-10 named “Ryuma.” All joints are rotary joints actuated by D.C. motors. Each robot in Figs. 6-10 has a head. The number of their joint is eleven including neck joint. The head motion change the center of gravity forward and backward. We succeeded in walking, trotting and bounding gait using the robots in Fig. 7. In addition, we built a simulation model using Open Dynamic Engine (ODE) and found that the head motion greatly affects its locomotion stability. The robots in Figs. 8 and 9 have the same ling length ratios as the real horses. We are trying to achieve walking, trotting, bounding, and galloping gait using the robot in Fig. 9.

Fig. 5 Ryuma(2004) Fig. 6 Ryuma(2005) Fig. 7 Ryuma(2006)

Fig. 8 Ryuma(2010) Fig. 9 Ryuma(2014)

Walking gait
Trotting gait
Bounding gait
Simulation on trotting gait using ODE

Bipedal robot

Traditionally, many walking robots, especially biped robots, have flat soles with several sensors to detect the ground reaction force and the contact timing with the floor from which we calculate the real zero moment point (ZMP) to achieve dynamically stabilized walking. On the contrary, a human foot contains many bones, tendons, and muscles, which form three arches: two longitudinal and one lateral. The three arches play an important role in absorbing shocks at the touch down and controlling the body posture.

We hypothesize that they prevent the rolling motion and adjust the lateral posture of the body. To confirm this idea, we built several mechanical feet with one longitudinal arch or two longitudinal arches (Figs. 10 and 11) and we attached them to a simple leg mechanism using Chebyshev linkages driven by only one motor (Fig. 12). We are now trying to experimentally prove the effectiveness of the foot with arches.

Fig. 10 One-arich foot Fig. 11 Two-arch foot Fig. 12 Leg mechanism with Chebyshev linkages

Underwater robot inspered by sperm whale

The goal of this study is to develop a buoyancy control device based on the sperm oil hypothesis and apply it to an underwater robot or ships. Sperm whales have a spermaceti organ in their head that is filled with sperm oil. Sperm oil is a high quality oil and was used as material for candles, lubricant, and so on. There is a hypothesis on the role of sperm oil that insists that sperm whales melt and coagulate their sperm oil and change the volume of the oil to control their own buoyancy. This hypothesis very attractive for the underwater robot, because no materials for ballasts, such as water and iron, are discarded in the sea.

We developed an underwater robot (Fig. 13) using paraffin wax as a substitute for sperm oil. Paraffin wax is contained in the syringe that are in the lower part of the robot. By heating the wax, we succeeded in making the robot surface. In addition, we developed other volume change mechanisms: taking advantage of rubber stretch (Fig.14) and metal bellows (Fig.15). We succeeded in depth control using the robot shown in Fig. 15.

Fig. 13 Robot with cylinder and piston Fig. 14 Buoyancy control device with rubber membrane Fig. 15 Robot with metal bellows

Surfacing of robot in Fig 13

Soft Tactile Sensor
We have just begun the research on a tactile sensor using soft material such as silicone rubber. While the sensor containes only one kind of a sensor element such as strain guages, it change its sensitivity to stimuli by altering its morphology. We are now analyzing and fablicating several kinds of the tactile sensors.
Brain machine-interface

Our new target is Brain-machine interface, which attracts many researchers from various fields. Particularly, we focus on a commercially available inexpensive device. In spite of the limitations of measuring brain waves (or an electroencephalogram: EEG), we believe that the device is promising because it is easy to equip it to human and they can move without disturbances that the other heavy measurement devices have. We are trying to apply the measured brain waves to swarm robots. However, we are still in preliminary stage and lots of effort have to be done.