Soft robots can revolutionize several applications with high demands on dexterity and safety. When operating these systems, real-time estimation and control require fast and accurate models. However, prediction with first-principles (FP) models is slow, and learned black-box models have poor generalizability. Physics-informed machine learning offers excellent advantages here, but it is currently limited to simple, often simulated systems without considering changes after training. We propose physics-informed neural networks (PINNs) for articulated soft robots (ASRs) with a focus on data efficiency. The amount of expensive real-world training data is reduced to a minimum — one dataset in one system domain. Two hours of data in different domains are used for a comparison against two gold-standard approaches: In contrast to a recurrent neural network, the PINN provides a high generalizability. The prediction speed of an accurate FP model is exceeded with the PINN by up to a factor of 467 at slightly reduced accuracy. This enables nonlinear model predictive control (MPC) of a pneumatic ASR. Accurate position tracking with the MPC running at 47 Hz is achieved in six dynamic experiments.
1 September 2016 in Current Directions in Biomedical Engineering
One of the greatest challenges in cardiovascular tissue engineering is to develop vascular grafts with properties similar to autologous vessels. A promising approach is the fabrication of scaffolds from biodegradable polymers by electrospinning. Unstructured vascular subs possess a weak dimensional stability resulting in lumen collapse when subjected to bending stress. In order to examine different structured grafts, a standardised test method is required. A manual test method, designed in a former study, was adopted in terms of standardisation and automation. Therefore, a control system was programmed to regulate the required electronics. The electronic circuit was then developed and put into service. To fix samples into the test bench a new sample holder and a new collector for electrospinning were designed. Subsequently, a validation showed the new systems’ improved functionality compared to the former test bench. The samples were manufactured with the new collector. They could be fixed to the sample holder with high repeatability. The demand for vascular grafts with biological and mechanical properties similar to autologous vessels requires a standardised test method to examine bendability. The new test system enables the scaffolds to be examined regarding bendability with low personal expense and a simultaneously high degree of reproducibility. In addition, the new collector geometry can be easily adapted to higher or lower inner diameters. Hence, a new sample geometry was developed within this work.
Positioning objects in industrial handling applications is often compromised by elasticity-induced oscillations reducing the possible motion time and thereby the performance and profitability of the automation solution. Existing approaches for oscillation reduction mostly focus on the elasticity of the handling system itself, i.e. the robot structure. Depending on the task, elastic parts or elastic grippers like suction cups strongly influence the oscillation and prevent faster positioning. In this paper, the problem is investigated exemplarily with a typical handling robot and an additional end effector setup representing the elastic load. The handling object is modeled as a base-excited spring and mass, making the proposed approach independent from the robot structure. A model-based feed-forward control based on differential flatness and a machine-learning method are used to reduce oscillations solely with a modification of the end effector trajectory of the robot. Both methods achieve a reduction of oscillation amplitudes of 85% for the test setup, promising a significant increase in performance. Further investigations on the uncertainty of the parameterization prove the applicability of the not yet widely-used learning approach in the field of oscillation reduction.
Borescope inspection is essential for assessing the airworthiness of aircraft gas turbines. Yet, current procedures remain highly manual, operator-dependent, and inconsistent, which limits the reliability of subsequent image-based damage analysis. This paper introduces a systematic design approach for an ultra-slender continuum robot (CR) tailored to the geometric and operational constraints of aero-engine inspection. We formalize the design space, compare actuation concepts, and select a tendon-driven architecture based on a structured evaluation. Dimensional synthesis is formulated as an optimization problem that maximizes the visible blade surface, yielding segment lengths that ensure high inspection coverage. We detail design, material, and cable choices, the actuation unit, and two variants of the manipulator: A fully actuated (FA) version and a hybrid version with a passive carrier (PC). Evaluation in a high-pressure compressor mock-up reveals distinct strengths in stiffness, controllability, friction, pose observability, and system complexity between the two systems. Based on these findings, future work should focus on advancing a hybrid solution that combines the benefits of both approaches. Moreover, the presented methodology is not limited to high-pressure compressor inspection but can be applied to any section of the engine, significantly broadening its scope of application.