Controlling soft robots with precision is a challenge due to the difficulty of constructing models that are amenable model-based control design techniques. Koopman operator theory offers way construct explicit dynamical and them using established methods. This approach data driven, yet yields an control-oriented model rather than just "black-box" input-output mapping. work describes Koopman-based system identification method its application predictive (MPC) for robots. Three MPC controllers...

Robotic ankle exoskeletons can provide assistance to users and reduce metabolic power during walking. Our research group has investigated the use of proportional myoelectric control for controlling robotic exoskeletons. Previously, these controllers have relied on a constant gain map user's muscle activity actuation signals. A may act as constraint user, so we designed controller that dynamically adapts amplitude. We hypothesized an adaptive would energy expenditure compared walking with...

This paper introduces StarlETH, a compliant quadrupedal robot that is designed to study fast, ecient, and versatile locomotion. The platform fully actuated with high series elastic actuation, making the system torque controllable at same time well suited for highly dynamic maneuvers. We additionally emphasize key elements of powerful real control simulation environment. work concluded number experiments demonstrate performance presented hardware controllers.

Controlling soft robots with precision is a challenge due in large part to the difficulty of constructing models that are amenable model-based control design techniques.Koopman operator theory offers way construct explicit linear dynamical and them using established methods.This method data-driven, yet unlike other data-driven such as neural networks, it yields an control-oriented model rather than just "black-box" input-output mapping.This work describes this Koopman-based system...

This paper demonstrates methods for the online optimization of assistive robotic devices such as powered prostheses, orthoses and exoskeletons. Our algorithms estimate value a physiological objective in real-time (with body "in-the-loop") use this information to identify optimal device parameters. To handle sensor data that are noisy dynamically delayed, we rely on combination dynamic estimation response surface identification. We evaluated three (Steady-State Cost Mapping, Instantaneous...

Physiological measures, such as pain, anxiety, effort, or energy consumption, play a crucial role in the evaluation and development of assistive robotic devices.Physiological data are collected analyzed by researchers clinicians, often used to inform an iterative tuning process device its controller.Currently, these then offline that they only evaluated after experiment has ended.This makes any design tedious time consuming since must be done on subject-by-subject basis for variety tasks is...

Soft robots are challenging to model due in large part the nonlinear properties of soft materials. Fortunately, this softness makes it possible safely observe their behavior under random control inputs, making them amenable large-scale data collection and system identification. This paper implements evaluates a identification method based on Koopman operator theory which models dynamical systems constructed via linear regression observed by exploiting fact that every has representation...

Controlling soft continuum manipulator arms is difficult due to their infinite degrees of freedom, nonlinear material properties, and large deflections under loading. This letter presents a data-driven approach identifying models that enables consistent control variable loading conditions. accomplished by incorporating loads into linear Koopman operator model as states estimating values online via an observer within the loop. Using this approach, autonomous pneumatically actuated achieved....

User preference is a promising objective for the control of robotic exoskeletons because it may capture multifactorial nature exoskeleton use. However, to use it, we must first understand its characteristics in context control. Here, systematically measured preferences individuals wearing bilateral ankle during walking. We investigated users’ repeatability identifying their and how changes with walking speed, device exposure, between different technical backgrounds. Twelve naive 12...

In this paper, we are presenting a method to estimate terrain properties (such as small-scale geometry or surface friction) improve the assessment of stability and guiding foot placement legged robots in rough terrain. Haptic feedback, expressed through joint motor currents ground contact force measurements that arises when prescribing predefined motion was collected for variety samples (four different shapes four properties). Features were extracted from data used training classification by...

SLIP models are generally known as one of the best and simplest abstractions describing spring-like leg behavior found in human animal running, have thus been subject to exhaustive investigation. To exploit these findings real robots, we utilize an operational space controller that projects model onto dynamics actual segmented robotic leg. Additionally, introduce a method compensate for energetic losses at impact collisions, which not accounted simplified assumptions. This allows direct...

This paper introduces the mechanical design and control concept of Series Compliant Articulated Robotic Leg ScarlETH which was developed at ETH Zurich for fast, efficient, versatile locomotion. Inspired by biological systems, we seek to achieve this through large compliances in joints enable natural dynamics, allow temporary energy storage, improve passive adaptability. A sophisticated chain cable pulley minimizes segment masses, places overall CoG close hip joint, maximizes range motion....

The inherent compliance of soft fluidic actuators makes them attractive for use in wearable devices and robotics. Their flexible nature permits to be used without traditional rotational or prismatic joints. Without these joints, however, measuring the motion is challenging. Actuator-level sensors could improve performance continuum robots with compliant multi-degree-of-freedom We make reinforcing braid a pneumatic artificial muscle (PAM McKibben muscle) "smart" by weaving it from conductive,...

This paper explores the benefits of using multiple gaits in a single robot. Inspired by nature, where humans and animals use different to increase their energetic economy, we analyzed how increasing speed affects choice gait, gait influences optimal speed. To this end, used control as tool identify motions that minimize cost transport two detailed models: planar biped quadruped. Both these models are actuated with high compliance series elastic actuators enable rich set natural dynamics....

Drawing inspiration from nature, this paper introduces and compares two compliant robotic legs that are able to perform precise joint torque position control, enable passive adaption the environment, allow for exploitation of natural dynamic motions. We report in detail on design control both prototypes elaborate specifically problem foot placement during flight without sacrifice efficient energy storage stance. This is achieved through an integrated approach incorporates series elastic...

Objective: Individuals with neurological damage (e.g., stroke or cerebral palsy) often experience a significant loss of arm function. Robotic devices that address muscle strength deficits in task-specific manner can assist the recovery function; however, current are typically large, bulky, and expensive to be routinely used clinic at home. This study sought this issue by developing portable planar passive rehabilitation robot, PaRRo. Methods: We designed PaRRo mechanical layout incorporated...

Controllers for assistive robotic devices can be divided into two main categories: controllers using neural signals and mechanically intrinsic signals. Both approaches are prevalent in research devices, but a direct comparison between the could provide insight their relative advantages disadvantages. We studied subjects walking with ankle exoskeletons different control modes: dynamic gain proportional myoelectric based on soleus muscle activity (neural signal), timing-based gait events...

We introduce a detailed numerical simulation and analysis framework to extend the principles of passive dynamic walking quadrupedal locomotion. Non-linear limit cycle methods are used identify possible gaits analyze stability efficiency walking. In doing so, special attention is paid issues that inherent locomotion, such as occurrence simultaneous contact collisions implications phase difference between front back leg pairs. Limit cycles identified within this correspond periodic can be...

This paper presents the actuation system of robotic leg ScarlETH. It was developed specifically for a quadrupedal robot and is designed to achieve fast position control as well accurate joint torque control. introduces strong passive dynamics create an efficient running behavior. High spring compliance with low damping in combination cascaded, motor velocity based, structure successfully tested simulation experiments. Final tests entire demonstrate that can perform hopping motion providing...

This paper introduces a state estimation framework for legged robots that allows estimating the full pose of robot without making any assumptions about geometrical structure its environment.This is achieved by means an Observability Constrained Extended Kalman Filter fuses kinematic encoder data with on-board IMU measurements.By including absolute position all footholds into filter state, simple model equations can be formulated which accurately capture uncertainties associated intermittent...

In this paper, we systematically investigate passive gaits that emerge from the natural mechanical dynamics of a bipedal system. We use an energetically conservative model simple spring-leg biped exhibits well-defined swing leg dynamics. Through targeted continuation periodic motions model, identify different bouncing in place. show these arise along one-dimensional manifolds bifurcate into branches with distinctly motions. The branching is associated repeated breaks symmetry motion. Among...

Body-in-the-loop optimization algorithms have the capability to automatically tune parameters of robotic prostheses and exoskeletons minimize metabolic energy expenditure user. However, current body-in-the-loop rely on indirect calorimetry obtain measurements cost, which are noisy, sparsely sampled, time-delayed, require wearing a respiratory mask. To improve these algorithms, goal this work is predict user's steady-state cost quickly accurately using physiological signals obtained from...