Musculoskeletal Robot Research Articles

Musculoskeletal Robots: Scalability in Neural Control

Anthropomimetic robots are robots that sense, behave, interact and feel like humans. By this definition, anthropomimetic robots require human-like physical hardware and actuation, but also brain-like control and sensing. The most self-evident realization to meet those requirements would be a human-like musculoskeletal robot with a brain-like neural controller. While both musculoskeletal robotic hardware and neural control software have existed for decades, a scalable approach that could be used to build and control an anthropomimetic human-scale robot has not been demonstrated yet. Combining Myorobotics, a framework for musculoskeletal robot development, with SpiNNaker, a neuromorphic computing platform, we present the proof-of-principle of a system that can scale to dozens of neurally-controlled, physically compliant joints. At its core, it implements a closed-loop cerebellar model which provides real-time low-level neural control at minimal power consumption and maximal extensibility: higher-order (e.g., cortical) neural networks and neuromorphic sensors like silicon-retinae or -cochleae can naturally be incorporated.

Musculoskeletal lower-limb robot driven by multifilament muscles

This paper presents a redundant musculoskeletal robot using thin McKibben muscles that is based on human anatomy. The purpose of this robot is to achieve motions and characteristics that are very similar to a human body. We use a thin McKibben muscle, which is compliant and flexible, as the actuator of a musculoskeletal robot. Using a bundle of thin McKibben muscles, we develop a multifilament muscle that has characteristics similar to those of human muscles. In contrast, the actuators of conventional musculoskeletal robots are very heavy, not densely attached and have poor backdrivability. Because multifilament muscles are light and can be densely attached, we can attach them to the musculoskeletal robot as skeletal muscle and achieve a redundant system that is equivalent to a human drive mechanism. In this paper, we report a method for fabricating multifilament muscles that imitate various muscles, the development of a lower-limb muscle mechanism for the redundant musculoskeletal robot with thin McKibben muscles and experimental results showing that the proposed musculoskeletal robot achieves humanlike motions that have not yet been reported for other robots.

A new biarticular actuator design facilitates control of leg function in BioBiped3

Bioinspired legged locomotion comprises different aspects, such as (i) benefiting from reduced complexity control approaches as observed in humans/animals, (ii) combining embodiment with the controllers and (iii) reflecting neural control mechanisms. One of the most important lessons learned from nature is the significant role of compliance in simplifying control, enhancing energy efficiency and robustness against perturbations for legged locomotion. In this research, we investigate how body morphology in combination with actuator design may facilitate motor control of leg function. Inspired by the human leg muscular system, we show that biarticular muscles have a key role in balancing the upper body, joint coordination and swing leg control. Appropriate adjustment of biarticular spring rest length and stiffness can simplify the control and also reduce energy consumption. In order to test these findings, the BioBiped3 robot was developed as a new version of BioBiped series of biologically inspired, compliant musculoskeletal robots. In this robot, three-segmented legs actuated by mono- and biarticular series elastic actuators mimic the nine major human leg muscle groups. With the new biarticular actuators in BioBiped3, novel simplified control concepts for postural balance and for joint coordination in rebounding movements (drop jumps) were demonstrated and approved.

Design and control of a pneumatic musculoskeletal biped robot.

Pneumatic artificial muscles are quite promising actuators for humanoid robots owing to their similar characteristics with human muscles. Moreover, biologically inspired musculoskeletal systems are particularly important for humanoid robots to perform versatile dynamic tasks. This study aims to develop a pneumatic musculoskeletal biped robot, and its controller, to realize human-like walking. According to the simplified musculoskeletal structure of human lower limbs, each leg of the biped robot is driven by nine muscles, including three pairs of monoarticular muscles which are arranged in the flexor-extensor form, as well as three biarticular muscles which span two joints. To lower cost, high-speed on/off solenoid valves rather than proportional valves are used to control the muscles. The joint trajectory tracking controller based on PID control method is designed to achieve the desired motion. Considering the complex characteristics of pneumatic artificial muscles, the control model is obtained through parameter identification experiments. Preliminary experimental results demonstrate that the biped robot is able to walk with this control strategy. The proposed musculoskeletal structure and control strategy are effective for the biped robot to achieve human-like walking.

Study on a Pneumatically Actuated Robot for Simulating Evolutionary Developmental Process of Musculoskeletal Structures

[abstFig src='/00280002/13.jpg' width=""300"" text='A pneumatic musculoskeletal robot' ]Under the effects of surroundings such as gravitational force, ambient temperature, and chemical substances, each animal has acquired an optimized body structure through its evolution. For example, vertebrate land animals have a sophisticated musculoskeletal structure including not only monoarticular muscles but also multiarticular muscles to support their weight against gravitational force. Many researchers have developed musculoskeletal robots with a biarticular muscle mechanism that enables them to execute physical tasks similar to the mimicked animal. However, the developmental process of the musculoskeletal structure has not been examined in detail in past studies. In this study, we developed a musculoskeletal robot with redundant air cylinders to investigate the developmental process of the body structure of the animal. We proposed a switching mechanism between several muscle structures called the actuator network system (ANS). In the ANS, the selection of mutually interconnected, simultaneously activated air cylinders is changed by switching the interconnections. The experimental results indicate that by changing the connection of the cylinders and the inner pressure of the connected cylinders, i.e., the strength of the connection, the response of the robot to external forces can be modified, thus demonstrating the feasibility of our approach.

細径マッキベン人工筋を用いた筋骨格ロボットの歩行動作実現

In recent years, artificial muscles have attracted attention as a flexible and light powerful pneumatic actuator. We have developed very thin McKibben artificial muscle. In our previous researches on a musculoskeletal robot using multifilament pneumatic muscles , we have confirmed good movements of the knee, ankle, elbow, hip and shoulder joints. In this paper, we successfully realize the musculoskeletal robot walking with dynamic and cooperative joint control. Dynamic characteristics of multifilament pneumatic muscle, driving plan of multifilament pneumatic muscle with musculoskeletal simulator and implementation of walking movement by the musculoskeletal robot are discussed.

筋骨格ロボット顎関節の試作研究

By developing a musculoskeletal mastication robot using thin McKibben muscles based on human anatomy, we aim to realize the same motion and characteristics as a human. Conventional mastication robots have a big system and do not have all muscles of human because it is difficult for conventional motors to be attached. Because thin McKibben muscles are small enough to be densely attached, we can overcome these problems by applying them to the musculoskeletal mastication robot as skeletal muscle and can additionally realize a redundant system equivalent to the human drive mechanism. In this paper, we report a prototype of musculoskeletal mastication robot with thin Mckibben muscles.

単一バルブ操作のみで跳躍動作を実現する空気圧駆動筋骨格ロボット

単一バルブ操作のみで跳躍動作を実現する空気圧駆動筋骨格ロボット

Shoulder complex linkage mechanism for humanlike musculoskeletal robot arms

The shoulder complex in the human body consists of the scapula, clavicle, humerus, and thorax and bears the load imposed by arm movements while at the same time realizing a wide range of motions. To mimic and exploit its role, several musculoskeletal robot arms with shoulder complex mechanisms have been developed. However, although many research groups have tried to design the structures using links and joints that faithfully correspond to the bones and joints in the human shoulder complex, its function has not been successfully reproduced because biologically plausible designs seriously compromise engineering plausibility. In this paper, we propose a linkage mechanism that can reproduce complex three-dimensional scapulo movements and considers the trade-off between biological and engineering plausibilities. Subsequently, the design was validated by driving the mechanism using pneumatic artificial muscles (PAMs) placed similarly to muscles in humans. Further, we present experiments in which the robot was controlled by surface electromyographic signals from a human. We show that the proposed design, due to its kinematic similarity with human musculoskeletal systems, eases the conversion between the surface electromyogram signals and the PAMs control inputs.

Extracting motor synergies from random movements for low-dimensional task-space control of musculoskeletal robots

In the field of human motor control, the motor synergy hypothesis explains how humans simplify body control dimensionality by coordinating groups of muscles, called motor synergies, instead of controlling muscles independently. In most applications of motor synergies to low-dimensional control in robotics, motor synergies are extracted from given optimal control signals. In this paper, we address the problems of how to extract motor synergies without optimal data given, and how to apply motor synergies to achieve low-dimensional task-space tracking control of a human-like robotic arm actuated by redundant muscles, without prior knowledge of the robot. We propose to extract motor synergies from a subset of randomly generated reaching-like movement data. The essence is to first approximate the corresponding optimal control signals, using estimations of the robot's forward dynamics, and to extract the motor synergies subsequently. In order to avoid modeling difficulties, a learning-based control approach is adopted such that control is accomplished via estimations of the robot's inverse dynamics. We present a kernel-based regression formulation to estimate the forward and the inverse dynamics, and a sliding controller in order to cope with estimation error. Numerical evaluations show that the proposed method enables extraction of motor synergies for low-dimensional task-space control.

2A1-Q03 空気圧駆動筋骨格ロボットにおいて一回の二値弁操作で特定方向力を発揮し続ける非円形プーリ設計法

Self-contained pneumatic driven musculoskeletal robots for dynamic motions are often equipped with ON/OFF valves for weight reduction. It is necessary to switch valves frequently since a discontinuous control system with ON/OFF valves causes overshooting. Frequent switching leads discarding kinetic and elastic energy. One-shot valves operation to realize a target motion is desired. Target of this research is to design a musculoskeletal structure not to discard energy in a target motion such as using maximum performance of a robot. We propose a design method of pneumatic driven musculoskeletal robots to realize a target dynamic motion by one-shot valves operation by designed variable moment arms of a musculoskeletal structure.

Estimation of physical interaction between a musculoskeletal robot and its surroundings

Recently, robots are expected to support our daily lives in real environments. In such environments, however, there are a lot of obstacles and the motion of the robot is affected by them. In this research, we develop a musculoskeletal robotic arm and a system identification method for coping with external forces while learning the dynamics of complicated situations, based on Gaussian process regression (GPR). The musculoskeletal robot has the ability to cope with external forces by utilizing a bio-inspired mechanism. GPR is an easy-to-implement method, but can handle complicated prediction tasks. The experimental results show that the behavior of the robot while interacting with its surroundings can be predicted by our method.

Reaching control of a full-torso, modelled musculoskeletal robot using muscle synergies emergent under reinforcement learning

‘Anthropomimetic’ robots mimic both human morphology and internal structure—skeleton, muscles, compliance and high redundancy—thus presenting a formidable challenge to conventional control. Here we derive a novel controller for this class of robot which learns effective reaching actions through the sustained activation of weighted muscle synergies, an approach which draws upon compelling, recent evidence from animal and human studies, but is almost unexplored to date in the musculoskeletal robot literature. Since the effective synergy patterns for a given robot will be unknown, we derive a reinforcement-learning approach intended to allow their emergence, in particular those patterns aiding linearization of control. Using an extensive physics-based model of the anthropomimetic ECCERobot, we find that effective reaching actions can be learned comprising only two sequential motor co-activation patterns, each controlled by just a single common driving signal. Factor analysis shows the emergent muscle co-activations can be largely reconstructed using weighted combinations of only 13 common fragments. Testing these ‘candidate’ synergies as drivable units, the same controller now learns the reaching task both faster and better.

A musculoskeletal bipedal robot designed with angle-dependent moment arm for dynamic motion from multiple states

When robots make smooth transitions in dynamic motions, they must exert large force over widely various postures. To expand the range of postures that robots can take during dynamic motions, we propose that robots be designed with an Angle-Dependent Moment Arm (ADMA) with biased pivot, for which torque characteristics of actuators attached to joints are adjustable. From jumping simulations of robotic legs designed with ADMA, we demonstrate that ADMA improves robustness to postural and motion timing changes by shifts of the optimal posture, which are also observed in jumping experiments using a full-sized, bipedal robot.

2A1-J06 アフリカツメガエルの筋骨格系を再現した生体筋駆動遊泳ロボット(バイオロボティクス)

Generation of adaptive swimming locomotion of Xenopus laevis is greatly affected by hydrodynamic interaction between its musculoskeletal system and water environments. To understand the mechanism of frog locomotion, it is a proper approach to build a hybrid musculoskeletal robot that has similar structure as a Xenopus laevis, and excite living muscles to drive the body like in vivo. We developed a swimming robot that had a similar musculoskeletal structure as a frog driven by living muscles. At first, we built musculoskeletal structure of the hind leg by a 3D printer copying the anatomy of Xenopus laevis. Then, we developed the swimming robot by driven the gastrocnemius muscles, realized kick motion that generated propulsion for swimming locomotion. And also, we checked that the robot's motion is similar to Xenopus laevis's one. Thus, to achieve to understand the swimming locomotion, the robot is expected to be a useful tool.

Roll-Over Shapes of Musculoskeletal Biped Walker

Abstract In this paper, we investigate how the roll-over characteristic of a passivity-based walking robot with flat feet is created and affected by its musculoskeletal structure, which is easily modeled by an agonistic and antagonistic pair of muscles. We hypothesize that the curvature radius and the center position of the roll-over shape (ROS), defined as the centers of pressures transformed into a shank coordinate, is strongly related to the stiffness and the equilibrium posture of the ankle joint, which are determined by the tonuses of the agonistic and the antagonistic muscles. We built a new musculoskeletal biped walking robot driven by pneumatic actuators and conducted exhaustive walking experiment with various ankle conditions. The hypothesis is verified experimentally by analyzing the ROSs of the walking robot.

2P1-A12 詳細人体模倣筋骨格ヒューマノイド「腱志郎」の上肢設計(ヒューマノイド)

To design a robot with humanlike body structure, this paper presents a design methodology for upper limb of tendon driven system. The robots explained here is the upper limb of human mimetic musculoskeletal humanoid robot "Kenshiro". Kenshiro is designed based on human anatomy, and consists of muscles, bones, and joint structures to move flexibly and dynamically. Design methodology for such robots is not well established, therefore the paper describes how to design such an upper limb with some key mechanical design points that enables the implementations. To show that these mechanisms is effective we demonstrate the range of motion of Kenshiro, and pitching motion as an example of continuous motion using upper limb.

2A2-F08 表面筋電位を用いた筋骨格ロボットアームの制御(バイオロボティクス(1))

In order to realize the intuitive control of robot arms, many researchers have studied methods of using surface electromyography(sEMG). However, difference in the structure of human and robot makes these methods complex and difficult. Thus, we have focused on the musculoskeletal robot arm which have similar structure to the human arm. In this paper, we propose a method to control a musculoskeletal robot arm driven by pneumatic artificial muscles(PAMs) using sEMG. We have developed a method to estimate the inner pressure of the PAMs from sEMG of the muscles when the robot arm takes a similar posture to the human arm. Then we verified the validity of the proposed method by controlling a 1-DOF robot arm driven by two PAMs using sEMG during movement of human wrist.

Advantages of flexible musculoskeletal robot structure in sensory acquisition

Morphological computation is the concept for which a well-designed hardware can bear part of the computational cost required for robot's control and perception. So far, many musculoskeletal robots have been developed by taking inspiration from human's one and shown superior motion performances. The use of pneumatic artificial muscles (PAMs) has been the key to realize these high performance. Additionally, PAMs have the possibility of being used as sensors for environmental information because they are flexible and backdrivable. In this research, we focus on clarifying how PAMs can contribute to morphological computation of robots driven by these actuators. In particular, we propose an analysis method based on transfer entropy and apply this method to the experimental data acquired by a musculoskeletal robot that opens a door.

Biomechanical Approach to Open-Loop Bipedal Running with a Musculoskeletal Athlete Robot

In this study, a musculoskeletal robot is used as a tool to investigate how animals control their complex body. Sprinting is a challenging task that requires maximizing the potential resources of a musculoskeletal structure. Our approach to robotic sprinting is the Athlete Robot — a musculoskeletal robot with elastic blade feet controlled by feedforward motor command. We use a catapult launcher to provide a stable start to a sprint, and then examine the relation between the initial velocity imparted by the launcher and the change in orientation of the robot. We also investigate the influence of the change in elasticity of the blade foot. The results show that acceleration causes anterior inclination after the first step. The elasticity of the foot dominates the duration of the support phase. The musculoskeletal system of the Athlete Robot is modified to suit catapulted running. Based on the results from real robot experiments, we can provide a consistent propelling force using the catapult launcher. We demonstrate the Athlete Robot running for five steps after a catapult launch, using only feedforward command.