Tokyo Robotics has unveiled its advanced humanoid robot, Torobo, in a recent video demonstrating the robot’s ability to precisely hammer a nail into a piece of wood.
Standing at 1660 mm with a reach of 680 mm, Torobo closely resembles the size and reach of a human. According to Tokyo Robotics, each arm of the robot can handle a payload of 8 kg in the worst-case holding posture and up to 20 kg at the instantaneous maximum holding force, making it well-suited for a variety of robotics research and development tasks aimed at replacing human labor.
“The instantaneous rebound force from the hammer is absorbed through a combination of the elasticity of the rubber material securing the hammer, the deflection in torque sensors and harmonic gears, backdrivability, and impedance control,” the company explained.
Equipped with torque sensors at all joints of its arms and waist, Torobo’s joint torque control allows for safe contact stops and force-controlled task executions. This makes it capable of performing tasks such as assembly, cooking, and physical interaction with humans more safely and skillfully. The robot’s design features 7-axis dual arms, a 3-axis waist (pitch, pitch, yaw), a 3-axis neck (yaw, pitch, roll), and a 4-axis undercarriage with an omnidirectional mobile base, providing it with a human-like range of motion suitable for living and working environments.
Developed to advance industrial applications of full-body humanoid robots, Torobo serves as a versatile platform for research into task automation involving active interaction with people, the environment, and objects. It supports next-generation force-controllable dual-armed robots and applies machine learning techniques to improve its functionality.
Torobo’s software leverages the Robot Operating System (ROS), enabling state visualization with RViz, trajectory planning using MoveIt!, and comprehensive logging and saving of sensor information such as camera images, joint angles, and joint torques. The robot also utilizes Noitom’s Perception Neuron PRO motion capture system for motion teaching, machine learning, and remote control applications.
The Perception Neuron PRO system provides essential features such as singularity avoidance, self-interference protection, and torque-limiting functions when the robot encounters physical constraints, such as pushing against a desk.
Moreover, it also allows independent control of the robot’s left and right arms, grippers, neck, waist, and mobile base, enhancing its versatility for a wide range of applications.
