An official website of the United States government
Creating soft robots with sophisticated, autonomous capabilities requires these systems to possess reliable, on-line proprioception of 3D configuration through integrated soft sensors. We present a framework for predicting a soft robot’s 3D configuration via deep learning using feedback from a soft, proprioceptive sensor skin. Our framework introduces a kirigami-enabled strategy for rapidly sensorizing soft robots using off-the-shelf materials, a general kinematic description for soft robot geometry, and an investigation of neural network designs for predicting soft robot configuration. Even with hysteretic, non-monotonic feedback from the piezoresistive sensors, recurrent neural networks show potential for predicting our new kinematic parameters and, thus, the robot’s configuration. One trained neural network closely predicts steady-state configuration during operation, though complete dynamic behavior is not fully captured. We validate our methods on a trunk-like arm with 12 discrete actuators and 12 proprioceptive sensors. As an essential advance in soft robotic perception, we anticipate our framework will open new avenues towards closed loop control in soft robotics.
more »
« less
Deep reinforcement learning achieves multifunctional morphing airfoil control
Smooth camber morphing aircraft offer increased control authority and improved aerodynamic efficiency. Smart material actuators have become a popular driving force for shape changes, capable of adhering to weight and size constraints and allowing for simplicity in mechanical design. As a step towards creating uncrewed aerial vehicles (UAVs) capable of autonomously responding to flow conditions, this work examines a multifunctional morphing airfoil’s ability to follow commands in various flows. We integrated an airfoil with a morphing trailing edge consisting of an antagonistic pair of macro fiber composites (MFCs), serving as both skin and actuator, and internal piezoelectric flex sensors to form a closed loop composite system. Closed loop feedback control is necessary to accurately follow deflection commands due to the hysteretic behavior of MFCs. Here we used a deep reinforcement learning algorithm, Proximal Policy Optimization, to control the morphing airfoil. Two neural controllers were trained in a simulation developed through time series modeling on long short-term memory recurrent neural networks. The learned controllers were then tested on the composite wing using two state inference methods in still air and in a wind tunnel at various flow speeds. We compared the performance of our neural controllers to one using traditional position-derivative feedback control methods. Our experimental results validate that the autonomous neural controllers were faster and more accurate than traditional methods. This research shows that deep learning methods can overcome common obstacles for achieving sufficient modeling and control when implementing smart composite actuators in an autonomous aerospace environment.
more »
« less
- Award ID(s):
- 1935216
- PAR ID:
- 10378422
- Publisher / Repository:
- SAGE Publications
- Date Published:
- Journal Name:
- Journal of Composite Materials
- Volume:
- 57
- Issue:
- 4
- ISSN:
- 0021-9983
- Format(s):
- Medium: X Size: p. 721-736
- Size(s):
- p. 721-736
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
null (Ed.)A closed-loop control algorithm for the reduction of turbulent flow separation over NACA 0015 airfoil equipped with leading-edge synthetic jet actuators (SJAs) is presented. A system identification approach based on Nonlinear Auto-Regressive Moving Average with eXogenous inputs (NARMAX) technique was used to predict nonlinear dynamics of the fluid flow and for the design of the controller system. Numerical simulations based on URANS equations are performed at Reynolds number of 106 for various airfoil incidences with and without closed-loop control. The NARMAX model for flow over an airfoil is based on the static pressure data, and the synthetic jet actuator is developed using an incompressible flow model. The corresponding NARMAX identification model developed for the pressure data is nonlinear; therefore, the describing function technique is used to linearize the system within its frequency range. Low-pass filtering is used to obtain quasi-linear state values, which assist in the application of linear control techniques. The reference signal signifies the condition of a fully re-attached flow, and it is determined based on the linearization of the original signal during open-loop control. The controller design follows the standard proportional-integral (PI) technique for the single-input single-output system. The resulting closed-loop response tracks the reference value and leads to significant improvements in the transient response over the open-loop system. The NARMAX controller enhances the lift coefficient from 0.787 for the uncontrolled case to 1.315 for the controlled case with an increase of 67.1%.more » « less
-
Abstract Uncrewed aerial vehicles are integral to a smart city framework, but the dynamic environments above and within urban settings are dangerous for autonomous flight. Wind gusts caused by the uneven landscape jeopardize safe and effective aircraft operation. Birds rapidly reject gusts by changing their wing shape, but current gust alleviation methods for aircraft still use discrete control surfaces. Additionally, modern gust alleviation controllers challenge small uncrewed aerial vehicle power constraints by relying on extensive sensing networks and computationally expensive modeling. Here we show end-to-end deep reinforcement learning forgoing state inference to efficiently alleviate gusts on a smart material camber-morphing wing. In a series of wind tunnel gust experiments at the University of Michigan, trained controllers reduced gust impact by 84% from on-board pressure signals. Notably, gust alleviation using signals from only three pressure taps was statistically indistinguishable from using six pressure tap signals. By efficiently rejecting environmental perturbations, reduced-sensor fly-by-feel controllers open the door to small uncrewed aerial vehicle missions in cities.more » « less
-
Time has become an essential aspect of many computing systems where temporal correctness is as important as functional correctness. Autonomous vehicles, Industry 4.0, and smart grids are a few examples of time-sensitive systems. As time-sensitive applications become large, complex, and distributed, traditional methods fall short of achieving the desired orchestration among components. In this vision article, we first propose a standard to maintain an accurate notion of time among all components of the system, i.e., sensors, computing platforms, and actuators. Then, we propose explicit-time state estimation and closed-loop control algorithms that can tolerate large delays while achieving reasonable performance, and an integrated fail-safe mechanism that achieves a high level of robustness when timing failures happen.more » « less
-
A bio-inspired, passively deployable flap attached to an airfoil by a torsional spring of fixed stiffness can provide significant lift improvements at post-stall angles of attack. In this work, we describe a hybrid active–passive variant to this purely passive flow control paradigm, where the stiffness of the hinge is actively varied in time to yield passive fluid–structure interaction of greater aerodynamic benefit than the fixed-stiffness case. This hybrid active–passive flow control strategy could potentially be implemented using variable-stiffness actuators with less expense compared with actively prescribing the flap motion. The hinge stiffness is varied via a reinforcement-learning-trained closed-loop feedback controller. A physics-based penalty and a long–short-term training strategy for enabling fast training of the hybrid controller are introduced. The hybrid controller is shown to provide lift improvements as high as 136 % and 85 % with respect to the flapless airfoil and the best fixed-stiffness case, respectively. These lift improvements are achieved due to large-amplitude flap oscillations as the stiffness varies over four orders of magnitude, whose interplay with the flow is analysed in detail.more » « less
