Control of soft robots by real-time mechanical computation of their deformable structure

Soft robotics are inspired by nature, the way in which certain living organisms, entirely deformable, move and adapt their shape to their environment. These robots are built from highly flexible materials (silicone or other elastomers), enabling them to perform tasks safely and flexibly. They are particularly suited to fragile environments. Their jointless mechanical design makes them easy to miniaturize. There are many potential applications for these robots in industry and medicine.However, one of the major challenges in this field is the modeling and control of these robots: unlike the rigid articulated case, the mechanical model of the robot can no longer be calculated analytically and quickly for control purposes. In the general case, it will be necessary to use a numerical model, of the FEM type, and the size of the model will be of a completely different order of magnitude. This makes things rather incompatible with the constraints of real-time robot control. It is this problem that will be the main focus of this presentation.This keynote will begin with a brief introduction to the field of soft robots and their applications. We will then detail the control problem, presenting the main difficulties encountered in soft robotics: under-actuation, redundancy, obtaining a direct and inverse kinematic model, stability, etc. Finally, we will outline the solutions proposed by our DEFROST research team, based on FEM methods calculated in real time (notably with fast implementations and reduced model size) and optimization methods (for contact calculation and constrained inverse modeling). We will show that this research enables robots to be used as a generalized, active force sensor. We'll also show how closed-loop control schemes can be applied to these robots. The presentation will be illustrated by videos of experiments, and will aim to show participants of the conference an original application of the computation of deformable structures to robotics.

Motion analysis methods for flexible multibody systems by use of Linear Complementarity Problem

Numerous studies have been conducted with enthusiasm to develop analysis methods for deformable mechanical systems, e.g. nonlinear finite element method, and improving the calculation efficiency is one of the significant subjects in these studies.Our research group has developed analysis methods for deformable multibody systems by using linear and complementarity relations of parameters which govern the motion of the system, which form a set of Linear Complementarity Problem (LCP). By leveraging the solvability of LCP in low-cost calculation, our proposed methods can achieve efficient analysis.The presentation provides an overview and highlights the characteristics of our proposed methods developed for two different types of deformable multibody systems. One is for tethered systems, and the other is for mechanical systems undergoing plastic deformation. Some numerical examples are also demonstrated to illustrates the validity of our proposed method.

Demystifying Minimal Coordinate Dynamics using Spatial Operators

Minimal coordinate approaches in multibody dynamics offer numerous benefits, including the elimination of redundant coordinates, compatibility with ODE integrators rather than DAE solvers, rapid solution algorithms, and suitability for embedded applications in robotics and control. Despite these advantages, minimal coordinates are under-utilized and often perceived as intricate, challenging, and having limited applicability to general multibody dynamics problems. This presentation introduces the Spatial Operator Algebra (SOA) methodology for overcoming these challenges associated with minimal coordinate dynamics.SOA employs a concise set of spatial operators to express and reveal the inherent structure of minimal coordinate dynamics models. These operators generate analytical expressions for critical dynamic quantities such as Jacobians, the mass matrix, its inverse, and others. Notably, the operator expressions and analyses exhibit consistency across the broad spectrum of multibody models ranging from simple serial rigid body systems to complex ones with arbitrary size and topology, and with loop constraints as well as flexible bodies.Moreover, these operator expressions seamlessly and systematically translate into cost-effective computational algorithms including both established methods, and deriving new ones. This unified algorithmic formulation serves as the cornerstone for the DARTS dynamics simulation software, widely employed for real-time and closed-loop dynanics simulations in several robotics and aerospace applications. SOA's capabilities extend to system-level analysis, opening up new avenues for exploring dynamics, such as diagonalized dynamics, sensitivity computations, and tackling novel challenges like the Fixman potential in molecular dynamics.

Rotorcraft Pilot Biomechanics for Pilot-Vehicle Interface Design and Verification

The emergence of unexpected, adverse pilot-vehicle interaction (Rotorcraft-Pilot Couplings, or RPC, in the context of this research, but the concept finds application in several fields) may affect vehicle performance, impact the scope, duration, and cost of development programs, and even represent a threat to safety (e.g., the safety of flight).After briefly defining RPCs and motivating the need for their study, the focus is placed on aeroelastic RPCs, where pilot biomechanics plays a fundamental role.The notion of pilot Biodynamic Feedthrough (BDFT) is introduced, with the main parameters that may influence it, also highlighting how it may depend on muscular activation, which in turn is task-dependent.The importance of biodynamics modeling - via multibody dynamics is discussed, presenting advancements on the topic by our research group and their experimental verification, creating a link between muscular activation, its task dependence, and how they both influence BDFT.

Function-Driven Dynamics of Mechanical Systems: Concepts and Applications

Many practical problems involve complex physical interactions among different parts of mechanical systems.The complete and systematic solution of such problems requires some re-thinking and novel considerations in dynamics of mechanical systems.This presentation will give an overview of related concepts and approaches, and bring illustrative examples and applications from areas such as robotics, vehicle systems, and virtual environments.