Robust control of a silicone soft robot using neural networks

Zheng, Gang; Zhou, Yuan; Ju, Mingda

doi:10.1016/j.isatra.2019.12.004

Cited by 34 publications

(15 citation statements)

References 21 publications

(40 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…[145] SOFA, in combination with an ad hoc plugin for soft robotics, [146] allows the real-time simulation of soft systems, and the inverse kinematics calculation, accounting also the contact with external objects. Recently, it has also been coupled with AI algorithms [147] for design optimization and higher-level control. While effective in many scenarios, the computational approach has an intrinsic limitation to deliver only a case-by-case solution, not providing a general modeling framework.…”

Section: Modeling Of Soft Systemsmentioning

confidence: 99%

“…Among them, a feedforward neural network and a Jacobian-based method [149] are compared and tested experimentally in a non-constant curvature soft arm. In another work, [150] a dynamic controller is implemented for a soft robot manipulator utilizing a deep neural network, while in [147] the robust control of a four-cable driven soft robot is addressed by using artificial neural networks. While computationally efficient, the main disadvantage of ML based techniques is the need for a training phase that can be timeconsuming for the acquisition of enough data to cover the whole workspace, and that can be used only for that specific system.…”

Section: Modeling Of Soft Systemsmentioning

confidence: 99%

See 1 more Smart Citation

3D Printing Materials for Soft Robotics

et al. 2020

View full text Add to dashboard Cite

that uses the sensor-generated information to accurately regulate and enhance the dynamic performance.Soft robots (and more in general soft material-based systems) can be built with a plethora of materials and composites having very different mechanical properties, spanning from fluids (i.e., liquid metals, etc.) and hydrogels, to stiff solids (i.e., acrylonitrile butadiene styrene (ABS), epoxy resins, etc.). Also, in most cases, the constituent materials have many degrees of freedom and are highly non-linear, with relevant hysteresis and viscoelastic behavior. Hence, modeling soft systems is a very challenging task, and it still misses a comprehensive theoretical foundation. [19] As a result, the modeling phase is often overlooked in the design process, which in most cases relies on trial and error experimentation. The most commonly used materials in the design of soft robotics are silicone-based elastomers, such as polydimethylsiloxane (PDMS). They are used as the flexible material in pneumatic actuators, [4,5,14] in magnetic actuators, [11] and in electrically driven actuators as well, usually along with a conductive layer, commonly a carbon-based material. [15,16] Besides, other flexible polymers such as polyimides, [8] polyurethanes (PU), [12,20] or hydrogels [9] are used. So far, most devices are fabricated by casting or molding of liquid monomers, followed by hardening by UV light or heat.3D printing is an additive manufacturing technique, in which a required material is deposited, layer by layer, to form a 3D object, that is, it is a sequential 2D printing. The main advantages of 3D printing over subtractive manufacturing methods, are the simplicity of the process and the ability to form complex and hollow structures. There is a variety of 3D printing technologies that are commercially available, as described in a diversity of review reports. [21][22][23] Therefore, we will present here only a concise description of the printing technologies that are most relevant to soft robotics, all of them are based on ink compositions which are liquid during the printing processes.The most widely used methods are based on extrusion of materials, called fused deposition modeling (FDM) [24,25] and direct ink writing (DIW). [26,27] Both methods rely on a material that is extruded through a nozzle onto a substrate, and differ by the materials being used, the nozzle's nature, and the fixation mechanism. In FDM, a thermoplastic filament is extruded through a heated nozzle, followed by solidification upon cooling on the substrate, to form a solid 2D layer. This technique is simple and not expensive, but is limited to thermoplastic polymers and has relatively low resolution. In DIW, a viscous Soft robotics is a growing field of research, focusing on constructing motorless robots from highly compliant materials, some are similar to those found in living organisms. Soft robotics has a high potential for applications in various fields such as soft grippers, actuators, and biomedical devices. 3D printing of soft robotics present...

show abstract

Section: Modeling Of Soft Systemsmentioning

confidence: 99%

Section: Modeling Of Soft Systemsmentioning

confidence: 99%

3D Printing Materials for Soft Robotics

et al. 2020

View full text Add to dashboard Cite

show abstract

“…These relevant research results support the possibility of realizing the proposed adaptive 4D-printed systems. Additionally, there have been several studies on the design and fabrication of 4D-printed compliant mechanisms promoting controlled self-management and self-actuation without needing sensors [34,56,63,[152][153][154]. Further investigations on equipping such mechanisms with controllers to provide more flexibility and adaptability are envisaged, allowing for their wider ranges of autonomous operations.…”

Section: Adaptive 4d-printed Systems Designmentioning

confidence: 99%

Control-Based 4D Printing: Adaptive 4D-Printed Systems

et al. 2020

View full text Add to dashboard Cite

Building on the recent progress of four-dimensional (4D) printing to produce dynamic structures, this study aimed to bring this technology to the next level by introducing control-based 4D printing to develop adaptive 4D-printed systems with highly versatile multi-disciplinary applications, including medicine, in the form of assisted soft robots, smart textiles as wearable electronics and other industries such as agriculture and microfluidics. This study introduced and analysed adaptive 4D-printed systems with an advanced manufacturing approach for developing stimuli-responsive constructs that organically adapted to environmental dynamic situations and uncertainties as nature does. The adaptive 4D-printed systems incorporated synergic integration of three-dimensional (3D)-printed sensors into 4D-printing and control units, which could be assembled and programmed to transform their shapes based on the assigned tasks and environmental stimuli. This paper demonstrates the adaptivity of these systems via a combination of proprioceptive sensory feedback, modeling and controllers, as well as the challenges and future opportunities they present.

show abstract

“…To date, the aforementioned approaches have relied on purposefully built sensor structures within the soft robot and may require the use of machine learning to fully utilize the sensor responses. In a similar vein, with the non-linear behavior surrounding the materials and unstructured nature of soft robotics, various facets of soft robotics research have turned to the use of data-driven or deep learning based methods for tasks such as design and fabrication [17], sensing [18,19], state estimation [20,21] and control [22][23][24].…”

Section: Introductionmentioning

confidence: 99%

Sensuator: A Hybrid Sensor–Actuator Approach to Soft Robotic Proprioception Using Recurrent Neural Networks

Preechayasomboon

2021

Actuators

View full text Add to dashboard Cite

Soft robotic actuators are now being used in practical applications; however, they are often limited to open-loop control that relies on the inherent compliance of the actuator. Achieving human-like manipulation and grasping with soft robotic actuators requires at least some form of sensing, which often comes at the cost of complex fabrication and purposefully built sensor structures. In this paper, we utilize the actuating fluid itself as a sensing medium to achieve high-fidelity proprioception in a soft actuator. As our sensors are somewhat unstructured, their readings are difficult to interpret using linear models. We therefore present a proof of concept of a method for deriving the pose of the soft actuator using recurrent neural networks. We present the experimental setup and our learned state estimator to show that our method is viable for achieving proprioception and is also robust to common sensor failures.

show abstract

Robust control of a silicone soft robot using neural networks

Cited by 34 publications

References 21 publications

3D Printing Materials for Soft Robotics

3D Printing Materials for Soft Robotics

Control-Based 4D Printing: Adaptive 4D-Printed Systems

Sensuator: A Hybrid Sensor–Actuator Approach to Soft Robotic Proprioception Using Recurrent Neural Networks

Contact Info

Product

Resources

About