System identification is a key discipline within the field of automation that deals with inferring mathematical models of dynamic systems based on input-output measurements. Conventional identification methods require extensive data generation and are thus not suitable for real-time applications. In this paper, a novel real-time approach for the parametric identification of linear systems using Deep Learning (DL) and the Modified Relay Feedback Test (MRFT) is proposed. The proposed approach requires only a single steady-state cycle of MRFT, and guarantees stability and performance in the identification and control phases. The MRFT output is passed to a trained DL model that identifies the underlying process parameters in milliseconds. A novel modification to the Softmax function is derived to better conform the DL model for the process identification task. Quadrotor Unmanned Aerial Vehicle (UAV) attitude and altitude dynamics were used in simulation and experimentation to verify the presented approach. Results show the effectiveness and real-time capabilities of the proposed approach, which outperforms the conventional Prediction Error Method in terms of accuracy, robustness to biases, computational efficiency and data requirements.

Variable Stiffness Actuators (VSA) have been proposed as an alternative actuation system for manipulators that are utilized for safe physical Human-Robot Interaction (pHRI). However, in the incidents of collision, the need of a fast response in stiffness tuning would rise to ensure safety. In this paper, we present a novel Discrete Variable Stiffness Actuator (DVSA) to be used in a compliant robotic manipulator for safe physical Human-Robot Interaction (pHRI). The novelty of this actuator lies in its design topology which allows the stiffness level to change swiftly among predefined levels without the need of complex stiffness tuning mechanism. Through this topology, three springs in parallel are connected serially between the motor and the link via gear train. The stiffness of the actuator is altered by adding/subtracting the number of involved springs, which can be realized through engagement/disengagement electromagnetic clutches on two of these spring’s shafts. The working principle, and the detailed design of the actuator are illustrated. Moreover, the stiffness model and the dynamic model are presented and discussed thoroughly. In order to validate these mathematical models and achieve optimal control, system identification for the dynamic parameters was performed experimentally on the physical model. Furthermore, the system’s ability of tracking desired trajectory was achieved through the implementation of different control techniques including PID (Proportional-Integral-Derivative), LQR (Linear Quadratic Regulator) and pole placement. The results show the high potential of utilizing the actuator in compliant manipulators. Moreover, DVSA is also characterized for safety in pHRI through Head-Injury Criterion (HIC). Finally, an application of DVSA in human augmentation task (Weight Bearing Task) is presented.

In this paper we present the design, development and experimental validation of a novel Binary-Controlled Variable Stiffness Joint (BpVSJ) towards haptic teleoperation and human interaction manipulators applications. The proposed actuator is a proof of concept of a passive revolute joint, where the working principle is based on the recruitment of series-parallel elastic elements. The novelty of the system lies in its design topology, including the capability to involve an (n) number of series-parallel elastic elements to achieve (2n) levels of stiffness, as compared to current approaches. Accordingly, the level of stiffness can be altered at any position without the need to revert to the initial equilibrium position. The BpVSJ has low energy consumption and short switching time, and is able to rotate freely at zero stiffness without limitations. Further smart features include scalability and relative compactness. This paper details the mathematical stiffness modeling of the proposed actuator mechanism, as well as the experimentally measured performance characteristics. The experimental results matched well with the physical-based modeling in terms of stiffness variation levels. Moreover, Psychophysical experiments were also conducted using (20) healthy subjects in order to evaluate the capability of the BpVSJ to display three different levels of stiffness that are cognitively realized by the users. The participants performed two tasks: a relative cognitive task and an absolute cognitive task. The results show that the BpVSJ is capable of rendering stiffness with high average relative accuracy (Relative Cognitive Task relative accuracy is 97.3%, and Absolute Cognitive Task relative accuracy is 83%).

Safety holds the prime importance in direct physical human-robot interaction (pHRI) tasks. Robots should have the ability to handle unexpected collisions in unstructured environments. Collision avoidance based on exteroceptive sensors can work in these scenarios. However, it may not be sufficient, especially considering that the relative motion between robots and humans can be fast and hardly predictable. This highlights the importance of fast and reliable detection and reaction techniques for the collisions. Rapid switching to intrinsic complaint mode upon collisions is a promising solution for this requirement. Recently, we have proposed a new design of the discrete variable stiffness actuator (DVSA), which has the capability of instantaneously switching its stiffness between different predefined levels. We believe that this rapid stiffness switching feature can significantly improve safety during collisions. In this letter, we combined a software-based collision detection method with a hardware-based rapid stiffness switching technique. The proposed strategy has been implemented on a DVSA-based manipulator to evaluate its safety performance in the sudden dynamic collision and static near-singular clamping collision scenarios. The results clearly indicate that the proposed strategy can significantly mitigate the impact of unexpected collisions and improve safety during pHRI.

In this paper, the modeling, design, and characterization of the passive discrete variable stiffness joint (pDVSJ-II) are presented. The pDVSJ-II is an extended proof of concept of a passive revolute joint with discretely controlled variable stiffness. The key motivation behind this design is the need for instantaneous switching between stiffness levels when applied for remote exploration applications where stiffness mapping is required, in addition for the need of low-energy consumption. The novelty of this work lies in the topology used to alter the stiffness of the variable stiffness joint. Altering the stiffness is achieved by selecting the effective length of an elastic cord with hook's springs. This is realized through the novel design of the cord grounding unit (CGU), which is responsible for creating a new grounding point, thus changing the effective length and the involved springs. The main features of CGU are the fast response and the low-energy consumption. Two different levels of stiffness (low, high) can be discretely selected besides the zero stiffness. The proposed physical-based model matched the experimental results of the pDVSJ-II in terms of discrete stiffness variation curves, and the stiffness dependency on the behavior of the springs. Two psychophysiological tests were conducted to validate the capabilities to simulate different levels of stiffness on human user and the results showed high relative accuracy. Furthermore, a qualitative experiment in a teleoperation scenario is presented as a case study to demonstrate the effectiveness of the proposed haptic interface.

Designing a mechanism for elbow self-axis alignment requires the elimination of undesirable joint motion and tissue elasticity. The novelty of this work lies in proposing a double-layered interface using a 3-PRR planar parallel mechanism as a solution to the axis alignment problem. 3-PRR planar parallel mechanisms are suitable candidates to solve this as they can span the desired workspace in a relatively compact size. In this paper, we present the modeling, design, prototyping, and validation of the double-layered elbow exoskeleton interface for axis self-alignment. The desired workspace for the self-axis alignment mechanism is specified based on the estimated maximum possible misalignment between the exoskeleton joint and the human anatomical elbow joint. Kinematic parameters of the 3-PRR planar mechanism are identified by formulating an optimization problem. The goal is to find the smallest mechanism that can span the specified workspace. The orientation angle of the mechanism’s plane addresses the frontal frustum vertex angle of the elbow’s joint, while the translational motion allows the translational offsets between the user’s elbow and the exoskeleton joint. The designed exoskeleton axis can passively rotate around the frontal plane ±15 deg and translate along the workspace 30 mm in the frontal plane. Experimental results (quantitative and qualitative) confirmed the capability of the proposed exoskeleton in addressing the complex elbow motion, user’s satisfaction, and ergonomics.