Tendon-driven Systems Research Articles

Multiparameter Remote Contact Force Sensor With Embedded Bend Sensing for Tendon-Driven Hand Robots

Integrating multiple sensors without increasing the structural complexity and their original form factor is a major concern in robotic applications, particularly in tendon-driven systems. This study proposes a multiparameter contact force sensor with bend sensing by utilizing an improved Bowden cable design. The proposed sensor consists of an end tip with an elastomer membrane, a Bowden cable consisting of a sheath and dual inner wires, and a remotely separated sensing component. The electrical wirings are decoupled from the mechanical Bowden cable system. The displacement difference between the force-sensing and bend-sensing wires is transmitted through the sheath, indicating the contact force on the end tip. The modeling and experiments verify that the contact force applied on the end tip is linearly related to the sensor output signal and is reliable after repeated measurements. The proposed sensor, utilizing the flexible Bowden cable as its sensing modality, can be integrated into a diverse range of tendon-driven robotic applications. These applications include a robotic hand with tactile and angle sensing capabilities, as well as a soft wearable robot for the hand (Exo-glove Poly II). The compact form factor of the proposed sensor enables geometric sensing without increasing structural complexity for multiparameter sensing in robotic applications.

Development and Evaluation of a Friction Model for Tendon-Driven Soft Robotic Devices

The Capstan formula is a common theoretical model that has been widely used to characterize friction between tendons and sheaths in tendon-driven transmission systems. Although several factors affect the friction in these systems, only two factors, the friction coefficient and the curvature angle of the sheaths, are taken into account in this theoretical model. Thus, understanding friction behavior still remains a significant limitation of control system performance for robotic systems that use tendon-driven mechanisms. This study aims to develop an improved friction model to more accurately determine the friction in tendon-driven systems. It considers the physical properties of the tendons and the sheaths by calculating the contact area and the adhesion force between them. The proposed friction model was verified by simulation and benchtop experiments, and compared with the Capstan formula. The results demonstrate that the error is reduced between 45% and 95% depending on the tendon angle and the sheath curvature. Thus, the proposed friction model can be used to characterize the friction between the tendons and sheaths in tendon-driven wearable devices, which could result in improved accuracy and better control of these devices.

Human-Like Endtip Stiffness Modulation Inspires Dexterous Manipulation With Robotic Hands.

We present a novel method for biomechanically inspired mechanical and control design by quantifying stable manipulation regions in 3D space for tendon-driven systems. Using this method, we present an analysis of the stiffness properties for a human-like index finger and thumb. Although some studies have previously evaluated biomechanical stiffness for grasping and manipulation, no prior works have evaluated the effect of anatomical stiffness parameters throughout the reachable workspace of the index finger or thumb. The passive stiffness model of biomechanically accurate tendon-driven human-like fingers enables analysis of conservatively passive stable regions. The passive stiffness model of the index finger shows that the greatest stiffness ellipsoid volume is aligned to efficiently oppose the anatomical thumb. The thumb model reveals that the greatest stiffness aligns with abduction/adduction near the index finger and shifts to align with the flexion axes for more efficient opposition of the ring or little fingers. Based on these models, biomechanically inspired stiffness controllers that efficiently utilize the underlying stiffness properties while maximizing task criteria can be developed. Trajectory tracking tasks are experimentally tested on the index finger to show the effect of stiffness and stability boundaries on performance.

Elastic Structure Preserving Impedance Control for Nonlinearly Coupled Tendon-Driven Systems

Elastic Structure Preserving Impedance Control for Nonlinearly Coupled Tendon-Driven Systems

InsideOut: A Bio-Inspired Machine Learning Approach to Estimating Posture in Robots Driven by Compliant Tendons.

Estimates of limb posture are critical for controlling robotic systems. This is generally accomplished with angle sensors at individual joints that simplify control but can complicate mechanical design and robustness. Limb posture should be derivable from each joint's actuator shaft angle but this is problematic for compliant tendon-driven systems where (i) motors are not placed at the joints and (ii) nonlinear tendon stiffness decouples the relationship between motor and joint angles. Here we propose a novel machine learning algorithm to accurately estimate joint posture during dynamic tasks by limited training of an artificial neural network (ANN) receiving motor angles and tendon tensions, analogous to biological muscle and tendon mechanoreceptors. Simulating an inverted pendulum—antagonistically-driven by motors and nonlinearly-elastic tendons—we compare how accurately ANNs estimate joint angles when trained with different sets of non-collocated sensory information generated via random motor-babbling. Cross-validating with new movements, we find that ANNs trained with motor angles and tendon tension data predict joint angles more accurately than ANNs trained without tendon tension. Furthermore, these results are robust to changes in network/mechanical hyper-parameters. We conclude that regardless of the tendon properties, actuator behavior, or movement demands, tendon tension information invariably improves joint angle estimates from non-collocated sensory signals.

Sampling-Based Nonlinear Stochastic Optimal Control for Neuromechanical Systems.

Determining how the nervous system controls tendon-driven bodies remains an open question. Stochastic optimal control (SOC) has been proposed as a plausible analogy in the neuroscience community. SOC relies on solving the Hamilton-Jacobi-Bellman equation, which seeks to minimize a desired cost function for a given task with noisy controls. We evaluate and compare three SOC methodologies to produce tapping by a simulated planar 3-joint human index finger: iterative Linear Quadratic Gaussian (iLQG), Model-Predictive Path Integral Control (MPPI), and Deep Forward-Backward Stochastic Differential Equations (FBSDE). We show that averaged over 128 repeats these methodologies can place the fingertip at the desired final joint angles but-because of kinematic redundancy and the presence of noise-they each have joint trajectories and final postures with different means and variances. iLQG in particular, had the largest kinematic variance and departure from the final desired joint angles. We demonstrate that MPPI and FBSDE have superior performance for such nonlinear, tendon-driven systems with noisy controls.Clinical relevance- The mathematical framework provided by MPPI and FBSDE may be best suited for tendon-driven anthropomorphic robots, exoskeletons, and prostheses for amputees.

A supertough electro-tendon based on spider silk composites

Compared to transmission systems based on shafts and gears, tendon-driven systems offer a simpler and more dexterous way to transmit actuation force in robotic hands. However, current tendon fibers have low toughness and suffer from large friction, limiting the further development of tendon-driven robotic hands. Here, we report a super tough electro-tendon based on spider silk which has a toughness of 420 MJ/m3 and conductivity of 1,077 S/cm. The electro-tendon, mechanically toughened by single-wall carbon nanotubes (SWCNTs) and electrically enhanced by PEDOT:PSS, can withstand more than 40,000 bending-stretching cycles without changes in conductivity. Because the electro-tendon can simultaneously transmit signals and force from the sensing and actuating systems, we use it to replace the single functional tendon in humanoid robotic hand to perform grasping functions without additional wiring and circuit components. This material is expected to pave the way for the development of robots and various applications in advanced manufacturing and engineering.

Force Sensing With 1 mm Fiber Bragg Gratings for Flexible Endoscopic Surgical Robots

This article presents a novel force sensor to detect the distal force of tendon-sheath mechanisms (TSMs) in flexible endoscopic surgical robots. We propose to measure the compression force on the sheath at the distal end so that the tension force on the tendon, which equals the compression force on the sheath, can be obtained. With this approach, a new force sensor made up of a 1 mm fiber Bragg grating attached to a 3 mm long nitinol tube was developed to measure the compression force exerted on the sheath. Mechanics analysis and verification tests were conducted to characterize the relationship between tension and compression on a TSM. Force calibrations, hysteresis study, and temperature compensation verification tests on the sensor were carried out. The force sensor has a measurement error of 0.178 N and a sensitivity of 34.14 pm/N. Applications of the sensor in a TSM-driven robotic grasper and a tendon-driven continuum robot were demonstrated. This force sensor has salient advantages: it is small, structurally simple, electrically passive, temperature-compensated, easy to assemble and disassemble, flexible, and biocompatible. This proposed approach with the new force sensor can also be applied to both TSM-driven systems and tendon-driven systems such as robotic fingers/hands, wearable devices, surgical catheters, and rehabilitation devices.

Stiffness analysis and experimental validation of modular-type hybrid antagonistic tendon-driven joint systems

SUMMARYThis paper proposes a new antagonistic tendon-driven joint (TDJ) that exhibits higher stiffness and larger travel range than conventional types of TDJs. A detailed mathematical analysis of the stiffness of the proposed TDJ is conducted and compared to other TDJs. The effect of the tendon length is taken into consideration to establish a more precise and realistic stiffness model of the proposed TDJ. Thereafter, two hardware prototypes of the proposed TDJ design, developed in the form of a packaged modular structure that integrates two TDJs, are introduced. Using these prototypes, the stiffness characteristics of the proposed TDJs are verified through experimentation. Additionally, experimental results on the stiffness behavior during the mimicked needle insertion tasks are provided. Results show that the proposed TDJs present much higher stiffness than conventional ones and thus give a potential benefit to precision manipulation.

Model-Free Control of Movement in a Tendon-Driven Limb via a Modified Genetic Algorithm.

Tendon-driven systems have many benefits over other actuation strategies such as torque-driven systems; however, their over-determined nature and posture-dependent actuation presents strong constraints on their control. Also, parameters or even exact structure of the model in these systems, especially in the biological ones, are normally not clear to the controller. Here, we propose a modified Genetic Algorithm that provides the tendon excursion values for the limb to follow a desired trajectory. Our results show that the proposed algorithm was able to accurately follow the desired trajectory without the model of the system being exposed to it. We believe that this method can enable biologically inspired tendon-driven mechanisms with variable mechanical structures to autonomously control their movements.

Elongation Modeling and Compensation for the Flexible Tendon--Sheath System

In tendon-driven systems, the elongation of the tendon would result in inaccuracy in the position control of the system. This becomes a critical challenge for those applications, such as surgical robots, which require the tendon-sheath system with flexible and even time-varying configurations but lack of corresponding sensory feedback at the distal end due to spatial restrictions. In this paper, we endeavor to address this problem by modeling the tendon elongation in a flexible tendon-sheath system. Targeting at flexibility in practical scenarios, we first derived a model describing the relationship between the overall tendon elongation and the input tension with arbitrary route configurations. It is shown that changes in the route configuration would significantly affect the tendon elongation. We also proposed a remedy to enhance the system tolerance against potential unmodeled perturbations along the transmission route during operation. A scaling factor S was introduced as a design guideline to determine the scaling effect. A dedicated platform that was able to measure the tensions at both ends and the overall tendon elongation was designed and set up to validate the new findings. Discussions were made on the performance and the future implementation of the proposed models and remedy.

Anthropomorphic tendon-driven robotic hands can exceed human grasping capabilities following optimization

How functional versatility emerges in vertebrate limbs in spite of their anatomical complexity is a longstanding question. In particular, fingers are actuated by numerous muscles pulling on tendons following intricate paths. In contrast, the tendon-driven robotic hands with intuitive tendon routings preferred by roboticists for their ease of analysis and control do not perform at the level of their biological counterparts. Thus there is much debate on whether and how the anatomy of the human hand contributes to grasp capabilities. These parallel questions in biology and robotics arise partly because it is unclear how the number and routing of tendons offer functional benefits. We use a novel computational approach that analyzes tendon-driven systems and quantifies grasp quality to compare the precision grasp capabilities of thousands of robotic index finger and thumb designs vs. the capabilities measured in human hands. Our exhaustive search finds that neither the symmetrical designs sometimes preferred by roboticists nor randomly generated designs approach the grasp capabilities of the human hand (they are on average 73% weaker). However, optimizing for anatomically plausible asymmetry in joint centers, tendon routings, and maximal tendon tensions produces designs that can exceed the human hand by 13–45%, and outperform the preferred robotic designs by up to 435%. Thus, the grasp capabilities of prosthetic or anthropomorphic hands can be greatly improved by judiciously altering design parameters, at times in counter-intuitive ways. Moreover we conclude that, in addition to its other capabilities, the human hand’s anatomy is very advantageous for precision grasp as it greatly outperforms numerous alternative robotic designs.

Reinforcement Learning and Synergistic Control of the ACT Hand

Tendon-driven systems are ubiquitous in biology and provide considerable advantages for robotic manipulators, but control of these systems is challenging because of the increase in dimensionality and intrinsic nonlinearities. Researchers in biological movement control have suggested that the brain may employ “muscle synergies” to make planning, control, and learning more tractable by expressing the tendon space in a lower dimensional virtual synergistic space. We employ synergies that respect the differing constraints of actuation and sensation, and apply path integral reinforcement learning in the virtual synergistic space as well as the full tendon space. Path integral reinforcement learning has been used successfully on torque-driven systems to learn episodic tasks without using explicit models, which is particularly important for difficult-to-model dynamics like tendon networks and contact transitions. We show that optimizing a small number of trajectories in virtual synergy space can produce comparable performance to optimizing the trajectories of the tendons individually. The six tendons of the index finger and eight tendons of the thumb, each actuating four degrees of joint freedom, are used to slide a switch and turn a knob. The learned control strategies provide a method for discovery of novel task strategies and system phenomena without explicitly modeling the physics of the robot and environment.

Two Tendon-Driven Systems Using Pneumatic Balloons

In recent years, Japanese society has been aging, creating a shortage of young workers. Robots are expected to be useful to perform tasks such as rehabilitation therapy, nursing elderly people and day-to-day work support for elderly people. In particular, robots that are designed for use in the fields of medical care and welfare must be safe for the human environment because they often come into contact with humans. Furthermore, robots must have dexterity that is similar to that of humans. Under these circumstances, tendon-driven systems of two types using flexible silicone rubber materials have been developed. They are lightweight and have characteristics resembling those of human muscle. This paper describes the basic features and mechanisms of two tendon-driven systems that we have developed as a soft actuator. Furthermore, basic properties of the system are evaluated and compared with those of human muscles.

2A2-B13 腱駆動関節における筋長-姿勢相対変位の非線形性に基づく筋配置推定法

Many tendon-driven systems have been developed to realize more natural and various motions, or to lighten the end of a robot arm. However, the control based on the geometrical model, one of control methods the main of recent, is not easy for tendon-driven robots because more information is necessary, for instance muscle layout etc. This paper describes a muscle layout estimation methods for tendon-driven robots. This method based on nonlinearity between relative displacement of muscle length and relative displacement of joint angle. Futhermore, the confirmatory experiment was done using the simulator and the real musculo-skeletal humanoid.

骨格型肘・前腕関節機構の張力きっ抗駆動

In this paper, the control method for a musculoskeletal mechanism simulating a human forearm is discussed. The human muscular system which is basically the multi-feedback system, serves for skillful human motions. The joint's compliance is controllable to regulate the spinal reflex feedback gains of muscle's length and tension. Furthermore, the cooperation of biarticular and uniarticular muscles improve the manipulation ability and the efficiency of energy. Because the tendon-driven system is similar to the human muscular system, we can take these superior characteristics of the human muscular system into our manipulator. Our manipulator has a similar mechanism to the human forearm and has 2 D.O.F motions (flexion, extension and external, internal rotation) by three contending tendon-driven systems. The control method for this manipulator is designed by imitating the human muscular control system based on spinal reflexes. In the controlled system, each wire compliance is adjustable to change the feedback gains of wire length and tension, and the manipulator compliance under the position control can be controlled by the regulation of the wire compliance without measuring the outer force.