Roberto Bortoletto

Virtual modelling of a real exoskeleton constrained to a human musculoskeletal model

Exoskeletons represent one of the most important examples of human-oriented robotic devices. This paper describes an existing lower-limb exoskeleton designed to assist people with lower extremity paralysis or weakness during the movements of standing up and walking. Starting from the analysis of a real system developed about seven years ago, a virtual multibody model was realized in order to deeply understand how the device worked and find out some potential improvements in the actuators control and in the kinematic design. The virtual device was properly constrained to a human musculoskeletal model in order to simulate a real operating condition. The analysis of the simulation results suggested a kinematic modification of the system and a new dynamic model was developed in order to test the new design through the comparison of four different models.

Socio-economic impact of medical lower-limb Exoskeletons

A constructive debate is ongoing among experts and academics about the social and economic impacts of advanced robotics. Exoskeleton robotic suits represent one of the most significant examples of what Human-Oriented Robotics is. After recent technological advances, the range of application fields of these devices has widened with respect to the first applications about teleoperation and power amplification. The aim of this paper is to contribute to the ongoing discussion by offering a vision of the possible future developments in terms of socio-economic impacts, resulting from the increasing use of Exoskeleton Robots, especially with regard to their applications in lower limb medical rehabilitation. In order to provide a concrete contribution to the current state-of-the-art, we are working on an alternative exoskeleton design approach to overcome the identified limits to the diffusion of this new technology. The achieved results are presented in the final part.

GMM-Based Single-Joint Angle Estimation Using EMG Signals

This paper aims to explore the possibility to use Electromyography (EMG) to train a Gaussian Mixture Model (GMM) in order to estimate the bending angle of a single human joint. In particular, EMG signals from eight leg muscles and the knee joint angle are acquired during a kick task from three different subjects. GMM is validated on new unseen data and the classification performances are compared with respect to the number of EMG channels and the number of collected trials used during the training phase. Achieved results show that our framework is able to obtain high performances even using few EMG channels and with a small training dataset (Normalized Mean Square Error: 0.96, 0.98, 0.98 for the three subjects, respectively), opening new and interesting perspectives for the hybrid control of humanoid robots and exoskeletons.

Lower Limb Stiffness Estimation during Running: The Effect of Using Kinematic Constraints in Muscle Force Optimization Algorithms

The focus of this paper is on the effect of muscle force optimization algorithms on the human lower limb stiffness estimation. By using a forward dynamic neuromusculoskeletal model coupled with a muscle short-range stiffness model we computed the human joint stiffness of the lower limb during running. The joint stiffness values are calculated using two different muscle force optimization procedures, namely: Toque-based and Torque/Kinematic-based algorithm. A comparison between the processed EMG signal and the corresponding estimated muscle forces with the two optimization algorithms is provided. We found that the two stiffness estimates are strongly influenced by the adopted algorithm. We observed different magnitude and timing of both the estimated muscle forces and joint stiffness time profile with respect to each gait phase, as function of the optimization algorithm used.

WorkCellSimulator: a 3d simulator for intelligent manufacturing

This paper presents WorkCellSimulator, a software platform that allows to manage an environment for the simulation of robot tasks. It uses the most advanced artificial intelligence algorithms in order to define the production process, by controlling one or more robot manipulators and machineries present in the work cell. The main goal of this software is to assist the user in defining customized production processes which involve specific automated cells. It has been developed by IT+Robotics, a spin-off company of the University of Padua, founded in 2005 from the collaboration between young researchers in the field of Robotics and a group of professors from the Department of Information Engineering, University of Padua.

Modeling and simulating compliant movements in a musculoskeletal bipedal robot

This paper describes the modeling and the simulation of a novel Elastic Bipedal Robot based on Human Musculoskeletal modeling. The geometrical organization of the robot artificial muscles is based on the organization of human muscles. In this paper we study how the robot active and passive elastic actuation structures develop force during selected motor tasks, and how we can model the contact between feet and ground. We then compare the robot dynamics to that of the human during the same motor tasks. The motivation behind this study is to reduce the development time by using a simulation environment for the purpose of developing a bipedal robot that takes advantage of the mechanisms underlying the human musculoskeletal dynamics for the generation of natural movement.

Simulating an Elastic Bipedal Robot Based on Musculoskeletal Modeling

Many of the processes involved into the synthesis of human motion have much in common with problems found in robotics research. This paper describes the modeling and the simulation of a novel bipedal robot based on Series Elastic Actuators (SEAs) [1]. The robot model takes inspiration from the human musculoskeletal organization. The geometrical organization of the robot artificial muscles is based on the organization of human muscles. In this paper we study how the robot active and passive elastic actuation structures develop force during selected motor tasks. We then compare the robot dynamics to that of the human during the same motor tasks. The motivation behind this study is to translate the mechanisms underlying the human musculoskeletal dynamics to the robot design stage for the purpose of developing machines with better motor abilities and energy saving performances.

Neural and Musculoskeletal Modeling: Its Role in Neurorehabilitation

Human NeuroMusculoSkeletal systems (NMS SYs) are very complex and have redundant anatomical degrees of freedom (DOFs) at muscles and joints. These features enable them to easily perform dexterous tasks since the childhood. NMS SYs have attracted many researchers from different scientific domains such as neurophysiology, robotics, biomechanics, and neuro-rehabilitation engineering because of its multi-task functionalities. Humans can perform hundreds of tasks and dynamically interact with external environments in a very efficient way without thinking about the complexity of the motor task. Thinking about twirling a coin or writing tasks, the many complex operations needed to perform such actions rise important questions like “do we really perform very complex computations to control our musculoskeletal system?” or “how do we control our musculoskeletal system to perform such actions?” and “what is the main contribution of our biomechanical structure in the motor control task?”. Recently, scientists have paid more attention not only to the neural commands but also to the biomechanical properties of NMS Sys and their role in simplifying the motor control tasks. Muscles are the main building blocks in our biomechanical systems. They can be continuously co-activated to produce and to coordinate movements maintaining the stability. Muscle-tendon actuators have been physically modeled, based on Hill-Type model, to study their non-linear behaviors and characteristics. Those models were then integrated with neuron models to provide a better understanding of the local control mechanism of a motor unit (e.g. spinal cord motor neuron and muscle-tendon actuator). Motor unit behaviors are observed through the muscle activity: the physiological process of converting an electrical stimulus to a mechanical response. This process is fundamental to muscle physiology, whereby the electrical stimulus is usually an action potential and the mechanical response is contraction. The transformation from Electromyographic (EMG) signal to muscle activation is not trivial and can occur through several steps. Muscle activation dynamics is the physiological process described by those steps. In general, the control of NMS models can be achieved also by combining together the EMG signals to retrieve muscle synergies. Apparently, humans use different motor control strategies to command their actions, some already exist in the Central Nervous System (CNS) with their birth and many others are developed and/or adapted during their life and gained experiences. However, both views of control strategies suggest a task dependency of the neural control. More details on description of muscle co-activation patterns based on the two views of the task dependent motor control strategies are provided in this chapter which will give an insight not only on a higher level of neural control but also at a lower level control of muscles in the CNS. Computational musculoskeletal models can provide an accurate knowledge of the physiological loading conditions on the skeletal system during human movements and allow quantifying factors that affect musculoskeletal functions, thus it can significantly improve clinical treatments in several orthopedics and neurological contexts. Every patient is different and possesses unique anatomical, neurological, and functional characteristics that may significantly affect optimal treatment of the patient. Therefore, personalized computational models of NMS systems can facilitate prediction of patient-specific functional outcome for different treatment designs and provide useful information for clinicians. Personalize computational models can be derived by generic models or subject-specific models with different levels of subject-specific details. In this chapter, we describe NMS systems in a bottom-up fashion. First we provide a deep insight on muscle contraction dynamics and musculoskeletal system properties. Then we discuss how a musculoskeletal system is locally driven by neuromuscular controls. Afterwards, we define how central motor commands are mapped through muscle synergies into low level controls. We discuss the two visions on the motor control strategies that CNS might use to perform motor control tasks and some related aspects inspired from neurorehabilitation studies and motor control experiments. Finally, we describe the importance and application of personalized subject-specific musculoskeletal modeling in neurorehabilitation.

Human Muscle-Tendon Stiffness Estimation During Normal Gait Cycle Based on Gaussian Mixture Model

The aim of this study is to estimate the stiffness of the muscle-tendon unit, of human lower limb, during the execution of a normal gait cycle. Unlike the analytical techniques already widely validated in literature and discussed below, a probabilistic approach based on the Gaussian Mixture Model (GMM) has been adopted here for the computation of the muscle-tendon unit stiffness. The obtained results for the major muscle groups are shown. The effectiveness of the proposed approach has been evaluated by computing the Root Mean Square (RMS) error between the stiffness calculated analytically and those calculated using the GMM, for each subject.

Effects of reserve actuators on optimization solutions: From muscle force to joint stiffness

This work presents a comparison between two optimization methods used to compute the muscle activation levels and corresponding forces that drives a set of generalized coordinates towards a set of desired trajectories. To improve the performance of musculoskeletal optimizations a supplemental set of actuators is often included in addition to the modeled muscles. Given a dynamic musculoskeletal model and five sets of reserve actuators, a series of numerical simulations have been performed using experimental data from a healthy male subject who executes a running movement at three different speeds. This is the first work to investigate the incidence of different reserve actuator sets on muscle activation-to-force optimization solutions, with respect to the estimation of the human lower limb muscle forces and corresponding joint stiffness. The results show significant differences between the obtained estimates, indicating a greater accuracy on the Computed Muscle Control solutions than pure Static Optimization solutions.