Kosta Jovanovic

Virtual laboratories for education in science, technology, and engineering

Within education, concepts such as distance learning, and open universities, are now becoming more widely used for teaching and learning. However, due to the nature of the subject domain, the teaching of Science, Technology, and Engineering are still relatively behind when using new technological approaches (particularly for online distance learning). The reason for this discrepancy lies in the fact that these fields often require laboratory exercises to provide effective skill acquisition and hands-on experience. Often it is difficult to make these laboratories accessible for online access. Either the real lab needs to be enabled for remote access or it needs to be replicated as a fully software-based virtual lab. We argue for the latter concept since it offers some advantages over remotely controlled real labs, which will be elaborated further in this paper.We are now seeing new emerging technologies that can overcome some of the potential difficulties in this area. These include: computer graphics, augmented reality, computational dynamics, and virtual worlds. This paper summarizes the state of the art in virtual laboratories and virtual worlds in the fields of science, technology, and engineering. The main research activity in these fields is discussed but special emphasis is put on the field of robotics due to the maturity of this area within the virtual-education community. This is not a coincidence; starting from its widely multidisciplinary character, robotics is a perfect example where all the other fields of engineering and physics can contribute. Thus, the use of virtual labs for other scientific and non-robotic engineering uses can be seen to share many of the same learning processes. This can include supporting the introduction of new concepts as part of learning about science and technology, and introducing more general engineering knowledge, through to supporting more constructive (and collaborative) education and training activities in a more complex engineering topic such as robotics. The objective of this paper is to outline this problem space in more detail and to create a valuable source of information that can help to define the starting position for future research. State of the art in dynamics-based virtual laboratories.Defining the criteria for critical evaluation of existing technologies.State of the art in virtual worlds.Future advances in the field of virtual-world based laboratories.

Toward anthropomimetic robotics: Development, simulation, and control of a musculoskeletal torso

Anthropomimetic robotics differs from conventional approaches by capitalizing on the replication of the inner structures of the human body, such as muscles, tendons, bones, and joints. Here we present our results of more than three years of research in constructing, simulating, and, most importantly, controlling anthropomimetic robots. We manufactured four physical torsos, each more complex than its predecessor, and developed the tools required to simulate their behavior. Furthermore, six different control approaches, inspired by classical control theory, machine learning, and neuroscience, were developed and evaluated via these simulations or in small-scale setups. While the obtained results are encouraging, we are aware that we have barely exploited the potential of the anthropomimetic design so far. But, with the tools developed, we are confident that this novel approach will contribute to our understanding of morphological computation and human motor control in the future.

The Puller-Follower Control of Compliant and Noncompliant Antagonistic Tendon Drives in Robotic Systems

This paper proposes a new control strategy for noncompliant and compliant antagonistic tendon drives. It is applied to a succession of increasingly complex single-joint systems, starting with a linear and noncompliant system and ending with a revolute, nonlinearly tendon coupled and compliant system. The last configuration mimics the typical human joint structure, used as a model for certain joints of the anthropomimetic robot ECCEROBOT. The control strategy is based on a biologically inspired puller-follower concept, which distinguishes the roles of the agonist and antagonist motors. One actuator, the puller, is considered as being primarily responsible for the motion, while the follower prevents its tendon from becoming slack by maintaining its tendon force at some non-zero level. Certain movements require switching actuator roles; adaptive co-contraction is used to prevent tendons slackening, while maintaining energetic efficiency. The single-joint control strategy is then evaluated in a multi-joint system. Dealing with the gravitational and dynamic effects arising from the coupling in a multi-joint system, a robust control design has to be applied with on-line gravity compensation. Finally, an experiment corresponding to object grasping is presented to show the controllers robustness to external disturbances.

Virtual Mechatronic/Robotic laboratory - A step further in distance learning

The implementation of the distance learning and e-learning in technical disciplines (like Mechanical and Electrical Engineering) is still far behind the grown practice in narrative disciplines (like Economy, management, etc.). This comes out from the fact that education in technical disciplines inevitably involves laboratory exercises and this fact drastically increases the complexity of a potential e-learning system. New approach and new specific knowledge are needed to develop such a system. We expect to meet the requirements of distance learning by developing the software-based laboratory exercises, i.e., a virtual laboratory. To fully substitute a physical system like laboratory equipment, one must emulate its full dynamics. The mathematical model in the form of differential equations will be applied to calculate dynamics and provide the data that would otherwise be measured on a physical system - this means simulation. To prove the feasibility of the concept and make a step towards full e-learning in technical disciplines, we consider a complex technical field, Mechatronics and more precisely, Robotics being a perfect symbiosis of Mechanical and Electrical Engineering. We present the Virtual Laboratory for Robotics (VLR). It possesses all the necessary features of a virtual laboratory: user interface, simulator, and visualization.

Towards anthropomimetic robotics: Development, simulation, and control of a musculoskeletal torso

Abstract Anthropomimetic robotics differs from conventional approaches by capitalizing on the replication of the inner structures of the human body, such as muscles, tendons, bones, and joints. Here we present our results of more than three years of research in constructing, simulating, and, most importantly, controlling anthropomimetic robots. We manufactured four physical torsos, each more complex than its predecessor, and developed the tools required to simulate their behavior. Furthermore, six different control approaches, inspired by classical control theory, machine learning, and neuroscience, have been developed and evaluated via these simulations or in small-scale setups. While the obtained results are encouraging, we are aware that we have barely exploited the potential of the anthropomimetic design so far. But, with the tools developed, we are confident that this novel approach will contribute to our understanding of morphological computation and human motor control in the future.

Anthropomimetic robot with passive compliance - Contact dynamics and control

This paper develops a dynamic model and designs a controller for a fully anthropomorphic, compliantly driven robot. To imitate muscles, the robots joints are actuated by DC motors antagonistically coupled through tendons. To facilitate safe interaction with humans, the robot exploits passive mechanical compliance, in the form of elastic springs in the tendons. To enable simulation, the paper first derives a mathematical model of the robots dynamics, starting from the “Flier” approach. The control of the antagonistic drives uses a biologically inspired puller-and-follower concept where at any instant the puller is responsible for the joint motion while the follower keeps the inactive tendon from slackening. In designing the controller, it was first necessary to use the advanced theory of nonlinear control for dealing with individual joints, and then to apply robust control theory in order to extend control to the multi-joint robot body.

Biologically Inspired Control of a Compliant Anthropomimetic Robot

Anthropomimetics copying nature in order to construct and control similar mechanism in technical world is an increasingly attractive research topic. Its core objective is achieving human efficiency in the aspects of technical world where engineering still cannot compete biology (diversity of motions, manoeuvrability, etc). This paper attempts to develop the appropriate dynamic model and design control for an anthropomimetic humanoid ECCEROBOT. Robot has a human-like mechanical structure skeleton. To mimic muscles, robot joints are driven by antagonistically coupled DC motors with tendons. Since the robot is expected to work in a human centered environment and in the presence of humans, an important accent is given to the safe interaction with the surrounding. To resolve the safety, robot involves passive mechanical compliance elastic springs in tendons. The paper derives the mathematical model of robot dynamics, designed for simulation. The control of antagonistic drives is based on a biologically inspired puller-and follower concept where the puller is responsible for joint motion while the follower keeps the inactive tendon from slackening. The advanced theory of nonlinear control was used for particular joints and the theory of robustness was necessary for applying the control to multi-joint robot body.

Control of Compliant Anthropomimetic Robot Joint

In this paper we propose a control strategy for a robot joint which fully mimics the typical human joint structure. The joint drive is based on two actuators (dc motors), agonist and antagonist, acting through compliant tendons and forming a nonlinear multi‐input multi‐output (MIMO) system. At any time, we consider one actuator, the puller, as being responsible for motion control, while the role of the other is to keep its tendon force at some appropriate low level. This human‐like and energetically efficient approach requires the control of “switching”, or exchanging roles between actuators. Moreover, an algorithm based on adaptive force reference is used to solve a problem of slacken tendons during the switching and to increase the energy efficiency. This approach was developed and evaluated on increasingly complex robot joint configurations, starting with linear and noncompliant system, and finishing with nonlinear and compliant system.

Tip-over stability examination of a compliant anthropomimetic mobile robot

This paper presents a wheeled humanoid robot as a structure composed by upper human-like body and mobile platform. The cart construction is supported by two driving wheels and one caster wheel and subjected to nonholonomic kinematic constraints. The real system configuration of the upper body and its model are represented as a fully anthropomimetic, compliant robot with antagonistic coupled drives. In order to ensure stability of the robot position, the robust control is evaluated. In this paper the focus is not on synthesis of local controllers, but the goal is examination of the limits of the adopted robot control strategy and robot handling with disturbance following from the cart motion (analysis of tip-over stability). In order to avoid tipping over and relying on the Zero Moment Point calculation and control algorithm, the appropriate construction (dimensions) of the cart is adopted. Finally, some simulations are carried out and the influence of different cart movements on the robot balance is analyzed by comparing different cases.

Dynamic modeling of an anthropomimetic robot in contact tasks

Considerations of energy efficiency and safe human-robot interaction have led to an increase in the exploitation of compliance in robotics, and much of this work has been inspired by biological systems. As a consequence, new analytical tools are now required in order to enable the dynamic analysis of these novel compliant robots, especially in their interactions with the environment. This paper extends the ‘Flier’ approach to show how it could be applied to the dynamic analysis of contact tasks involving a highly compliant biologically inspired robot – in this case an anthropomimetic robot, a humanoid with a human-like skeleton and artificial muscles, in which the joints are actuated by DC motors acting via compliant tendon transmissions. First, a computer-based model of the robot’s dynamics is developed. Various constraints are then introduced to describe the contacts (including impacts) with the ground, and with objects in the environment. Simulation results are presented for two types of interactions with the external world: a grasping task and the case of the robot moving on a mobile base. Graphical Abstract