Advances in Service and Industrial Robotics

In this paper, a multilayer neural network based approach is proposed for the human-robot collisions detection during the motions of a 2-DoF robot. One neural network is designed and trained by Levenberg-Marquardt algorithm to the coupled dynamics of the manipulator joints with and without external contacts to detect unwanted collisions of the human operator with the robot and the link that collided using only the proprietary joint position and joint torque sensors of the manipulator. The proposed method is evaluated experimentally with the KUKA LWR manipulator using two joints in planar horizontal motion and the results illustrate that the developed system is efficient and very fast in detecting the collisions as well as the collided link.

Virtual guides are used in human-robot cooperation to support a human performing manipulation tasks. They can act as guidance constrains to assist the user to move in the preferred direction or along desired path, or as forbidden-region constraint which prevent him to move into restricted region of the robot workspace. In this paper we proposed a novel framework that unifies virtual guides using virtual robot approach, which is represented with the admittance control, where a broad class of virtual guides and constraints can be implemented. The dynamic properties and the constraints of the virtual robot can be defined using three sets of parameters and variables: desired motion variables, dynamic parameters (stiffness, damping and inertia) and dead-zones. To validate the approach we implemented it on a KUKA LWR robot for the Buzz-Wire tasks, where the goal is to move a ring along a curved wire.

Introducing compliance in stiff robots arises from different design paradigms, such as achieving explosive motion or enhancing cyclic motion. However, a great emphasis is laying on the safety aspect especially in collaborative robots (Cobots) design. Herein, we present a novel design V2SOM which stands for Variable Stiffness Safety Oriented Mechanism. This new device, as its name indicates, comes to ensure the safety of physical Human/Robot Interaction (pHRI) as well as to reduce the dynamics’ drawbacks of making robots compliant. Due to its two continuously linked functional modes, high and low stiffness modes. V2SOM presents a high inertia decoupling capacity which is a necessary condition for a safe pHRI without compromising robot’s dynamic performances. The high stiffness mode eases the control without compromising the safety aspect. Once a Human/Robot (HR) collision occurs, a spontaneous smooth shift to low stiffness mode is passively triggered to safely absorb the impact. In this work two complementary safety criteria of pHRI are considered. The impact force (ImpF) criterion and Head Injury Criterion (HIC) for external and internal damage evaluation of blunt shocks, respectively.

Integration of active compliant joints in a robotic system contributes to enhance safety in physical human-robot interactions. In this paper we study the behavior of a variable stiffness actuator -V2SOM-, proposed in the context of the SISCob project. We describe the working principle and design of the proposed system. Furthermore, the dynamic modeling, including its nonlinear stiffness law, is presented. A preliminary study is conducted to evaluate the safety performance of a collaborative robot using the proposed mechanism. Thus, a dynamic simulator has been developed to integrate the compliant behavior of V2SOM in two joints of a Kuka iiwa robot. In order to evaluate the safety performance of the system, some following-trajectory tasks are executed in a planar workspace whereas unexpected collisions are produced with an unknown environment. The results presented in this paper show the effectiveness of V2SOM, enhancing the safety performance of the robot.

Motion transfer from a human to a robot implies accurate tracking of desired, demonstrated trajectories. However, direct imitation of joint position trajectories might not result in similar behavior of the robot and the human, because they have different kinematic and dynamic properties, i.e., different embodiment. To avoid the correspondence problem, the demonstrated trajectories need to be somehow adapted. In this paper we go beyond simple imitation, but we show how the torque profiles that should execute the demonstrated position trajectories are being learned in a manner that preserves the correspondence. Thus, position trajectories are modified from the demonstration and, furthermore, the robot executes the motion that preserves correspondence in a compliant manner. Because it is compliant, the robot is safer for the nearby person or environment, as potential unforeseen collisions will result in lower impact forces. We show the results of motion transfer of squatting from a human to a simulated CoMaN humanoid robot.

Progressive automation aims at quick and easy programming of repetitive tasks by means of kinesthetic demonstration from an operator. In this paper we address the general case, where the operator demonstrates a task involving motion of both the robot’s wrist position and its orientation, so that the robot progressively transitions from manual into autonomous operation based on the agreement level between consecutive demonstrations. Utilizing unit quaternions, we propose a method that combines dynamic movement primitives for encoding the task space motion with variable stiffness control for continuous role allocation, enabling an operator to easily program task by demonstration. We verify experimentally that a pick and place task with changes in the position and the orientation can be programmed with the proposed method in less than a minute.

Human personality has always been an esoteric topic in the field of social science. The five-factor model dealing personality traits plays a critical role in the field of communication studies, psychology and philosophy. Judging or recognizing personality trait is undoubtedly a cognitive aspect which requires intelligence. This judgmental process is extremely fuzzy as there are so many facets ingrained in every human. Interestingly, there is no quantitative standard to judge the severity of each facet. Intuitive and perceptive skills do the trick for the person judging personality of another person. In this work, an approach has been presented that uses only nonverbal cues to recognize extroversion-introversion personality. Major facets, e.g., human posture, facial expression, speech duration, rapid body movements, etc. are considered to recognize extroversion-introversion trait. A job interview scenario has been created in which a robot interacts with a candidate and judges his personality trait. Experimental studies validate the approach which is a promising basis for the development of computing approaches capable of predicting a specific personality trait.

This paper presents an implementation and experimental validation of an algorithm for collision detection for industrial robots performing repetitive tasks. Collisions are detected by using two decision rules, one of which compares current-based estimated torques with previously calculated reference limits, while the second rule detects changes in torque dynamics. Reference limits represent experimentally acquired torque values processed in order to determine measurement tolerances based on dynamics of the signal, as well as to adapt to different sampling times. The main contribution of the paper is the algorithm which is entirely implemented on robot’s controller with closed control architecture, and only requires PC for initial offline processing of reference torques. It can be adapted for use on various brands of industrial robots, and it is universal to different robot configurations.

The current work addresses the problem of learning the spatio-temporal characteristics of human motions through observation. Learned actions can be subsequently invoked in the context of complex Human-Robot Interaction scenarios. Unlike previous Learning from Demonstration (LfD) methods that cope only with the spatial features of an action, the formulated approach effectively encompasses spatial and temporal aspects. The latter are compactly depicted in a latent space representation of human motions. Learned actions are reproduced in the studied scenarios under the high-level control of a time-informed task planner. During the implementation of a given scenario, temporal and physical constraints may impose speed adaptations in the reproduced actions. The employed latent space representation readily supports such variations, giving rise to novel actions in the temporal domain. Experimental results demonstrate the effectiveness of the proposed formulation, as well as the proper execution of more involved scenarios.

This paper deals with the identification of upper limb motion specifications for robot assisted exercising. Motion capture system has been used to track the upper limb motion of a subject with a specific protocol. Two movements have been investigated based on the procedure done by the medical specialists on patients. The experimental analysis focused on the shoulder, leading to range of motion, has been presented and used to reconstruct the upper limb kinematic model. The ability of a cable-driven robot LAWEX, which has been designed and built at LARM to replicate the upper limb motion is discussed.

This paper analyzes the use of a McKibben pneumatic artificial muscle in a device for hand rehabilitation. This kind of actuator is widely implemented in several contexts related to human-machine mechanical interaction, such as robotics and rehabilitation, but a quantitative assessment of its suitability for such applications is often lacking. The structure of a system for retrieval of extension in a rigid finger is outlined, as well as the definition of operating characteristics like the finger workspace and the fingertip trajectory. A mathematical model for the whole system is presented, then the kinematic and static analyses are performed, leading to the calculation of the torques at the joints of the patient’s finger. Finally, the results of the simulations are discussed. The work demonstrates that the use of a pneumatic muscle can lead to the realization of an effective and well controlled rehabilitation system.

The development of manipulative systems oriented to concrete orthopaedic manipulations is an advanced tendency aiming maximally simplification of robot mechanical system. As a result so called Handheld Robotized Systems appear. They have the accuracy like the universal multifunctional robots and in the same time they are much cheaper, user-friendly and easy for maintenance. In this paper a new handheld bone drilling robotized system ODRO is considered. The purpose is to compare its functional characteristics with the known in the literature similar systems some of which are also available on the market.

The purpose of our study is to research the usability of a tele-controlled service robot by performing real tests of the robot with the elderly. Robot control is based on a multichannel system for data distribution from external devices, such as – joystick, virtual joystick, microphone, Kinect, and Leap Motion to provide effective assistance of the elderly for their different needs. In the paper, there are described the functionalities of the robot ROBCO 17 and the principles of the performing of the experiments. Finally, we present the results of the impressions of the elderly and their recommendations for future upgrades of the service robot.

Mobility impairment is a common problem for the elderly population which relates to difficulties in performing Activities of Daily Living (ADLs) and consequently leads to restrictions and the degradation of the living standards of the elders. When designing a user-friendly assistive device for mobility constrained people, the variable spectrum of disabilities is a factor that should affect the design process, since people with different impairments have different needs to be covered by the device, thus an adaptive behavior of those systems is necessary. Also, the performance of bathing activities includes several challenges for the elderly people, since such tasks require body flexibility. In this paper, we present current frameworks and solutions for intelligent robotic systems for assistive living involving human robot interaction in a natural interface. Our aim is to build such systems, in order to increase the independence and safety of these procedures. To achieve human - robot interaction in a natural way, we have to adapt the expertise of carers regarding bathing motions and walking assistance. The main goal of this work is to present recent research results towards the development of two real-life use cases incorporating intelligent robotic systems, aiming to support mobility and bathing activities for the elderly in order to provide context-aware and user-adaptive assistance.

This paper presents a mechatronic evaluation of forest fire monitoring systems based on UAV. To begin with, a mapping of the requirements to the mechatronic abilities, which should be embodied by these systems, is presented. The enabling technologies that support these abilities are briefly reported. The evaluation of these systems’ architectural schemes is accomplished with the discrete Choquet integral. As a result, UAV based schemes are found to be better than other proposed schemes for forest fire monitoring.

In recent years the term “industry 4.0” has been consolidated, indicating a trend towards the automation of industrial processes with the introduction of new technologies to support the production of goods and services. Connected to this new type of industry, in this paper, the authors describe the implementation of a solution based on the automatic analysis of data extracted from a photogrammetric survey. In particular, a low-cost UAV system has been used and, in this paper, all the analyses involved in the planning of the data acquisition phase, their management and processing within a GIS system are described. During the acquisition phase, different data sets were acquired using different techniques, recording 4 k video and acquiring high resolution images. The place chosen to perform this test was the Serrù dam, a plant located in Ceresole Reale, near Turin, Piedmont (Italy). A 3D model was made of this structure on which a wall integrity check was carried out.

In this paper a new UAV video stitching method which is robust even in the presence of forest fire is proposed. The proposed method exploits in a complementary way both optical and thermal cameras. In particular, the visibility of the optical cameras in the presence of smoke is improved by exploiting the transparency of the thermal cameras to smoke. In addition, a new system based on the ECC image alignment algorithm that is capable in solving the video stabilization and stitching problems is proposed. Preliminary results obtained from the application of the proposed system in UAV captured videos are very promising.

Micro Aerial Vehicles (MAVs) are platforms that received great attention during the last decade. Recently, the mining industry has been considering the usage of aerial autonomous platforms in their processes. This article initially investigates potential application scenarios for this technology in mining. Moreover, one of the main tasks refer to surveillance and maintenance of infrastructure assets. Employing these robots for underground surveillance processes of areas like shafts, tunnels or large voids after blasting, requires among others the development of elaborate navigation modules. This paper proposes a method to assist the navigation capabilities of MAVs in challenging mine environments, like tunnels and vertical shafts. The proposed method considers the use of Potential Fields method, tailored to implement a sense-and-avoid system using a minimal ultrasound-based sensory system. Simulation results demonstrate the effectiveness of the proposed strategy.

In this paper, we present a novel framework for horizon line (HL) detection that can be effectively used for Unmanned Air Vehicle (UAV) navigation. Our scheme is based on a Canny edge and a Hough detector along with an optimization step performed by a Particle Swarm Optimization (PSO) algorithm. The PSO’s objective function is based on a variation of the Bag of Words (BOW) method to effectively consider multiple image descriptors and facilitate efficient computation times. More specifically, the image descriptors employed are $$L*a*b$$ color features, texture features, and SIFT features. We demonstrate the effectiveness and robustness of the proposed novel horizon line detector in multiple image sets captured under real world conditions. First, we experimentally compare the proposed scheme with the Hough HL detector and a deep learning HL estimator, a prominent example of line detection, and demonstrate a significant boost in accuracy. Furthermore, since from the horizon line the UAV roll and pitch angles can be derived, this scheme can be used for UAV navigation. To this end, to further validate our approach, we compare the horizon computed roll and pitch angles to the IMU ones obtained with a complementary filter.

The paper treats the control problem of the locomotion phases in a jumping cycle of a robot. The mechanical architecture of the leg with elastic lower segment (ATHLETE MODEL) is discussed. The dynamic model of the motion in stance phase is determined. The touch-down sequence, when the elastic foot hits the ground, is discussed and the control system is analyzed for two cases: actuator as passive damper system and actuator as semi-active damper system with ground-hook damper model. The transmissibility characteristics are analyzed. An active damper with ER fluids actuator and a skyhook viscosity controller is proposed to avoid the vibrations in the mechanical structure that can disturb the evolution of the robot. The control parameters are determined by using the circle criterion. The take-off sequence conditions in the stance phase are inferred by using the energy concept.

This paper describes a hybrid omniwheeled-legged hexapod robot, named as Cassino Hexapod III, which has been designed and built at LARM in Cassino. Direct and inverse kinematics are outlined and implemented in a proper Arduino code for achieving a user interface with an Android application, which can manage robot operation in a simple user-friendly manner. Laboratory experiences have been carried out to demonstrate the suitable operation modes of Cassino Hexapod III in terms of wheeled and legged gaits. Specific attention is addressed to the use of omniwheels for achieving complex turning paths.

This work deals with the kinematic analysis of the gait cycle of a subject fitted with jumping stilts. A motion capture analysis was conducted by recording the trajectory of the subject’s free-flying hoof during a complete gait cycle. A six-degrees of freedom serial chain was used to simulate the subject’s legs kinematic on the sagittal plane. The positions and velocities of the subject’s free-flying hoof obtained from the photo-frames of the video and from the kinematic model was used to validate the model.

ATRV-mini was a popular, 2000’s commercially available, outdoors robot. The successful upgrade procedure of a decommissioned ATRV-mini is presented in this paper. Its robust chassis construction, skid steering ability, and optional wifi connectivity were the major reasons for its commercial success, mainly for educational and research purposes. However the advances in electronics, microcontrollers and software during the last decades were not followed by the robot’s manufacturer. As a result, the robot became obsolete and practically useless despite its good characteristics. The upgrade used up to date, off the shelf components and open source software tools. There was a major enhancement at robot’s processing power, energy consumption, weight and autonomy time. Experimental testing proved the upgraded robot’s operational integrity and capability of undertaking educational, research and other typical robotic tasks.

The paper presents an innovative design for wireless sensor networks (WSNs) which can increase their performances in the precision agriculture (PA). The authors argue on the reasons why presently WSNs are not integrated in PA on a larger scale. To this end, the paper describes the state of art in WSNs used in PA, and develops the proposed solution on the basis of the identified conjectures. The conceived solution proves how nodes can have small dimensions without mitigating the communication range or energy autonomy. Small dimensions bring benefits also to costs and operations. Such results can be achieved by overcoming the misalignment between advances in unmanned agricultural vehicles (AVs) and advances in WSN. The fixed nodes and mobile nodes (AVs), cooperate to fulfil sensor and communication coverage. Endowing the WSN nodes only with necessary sensors together with an energy awareness communication algorithm adopted in resonance with PA requirements proves to be a new competitive approach. Some simulations to validate these concepts are also provided.

In this paper, an electric stair-climbing wheelchair, named wheelchair.q05, able to move on flat ground and to climb stairs, is presented. The proposed solution has been developed through a series of studies and designs, all based on a smart hybrid triple leg-wheel locomotion unit. The stability of the device is guaranteed by a rear support of a pair of pivoting wheels during motion on flat ground, and by the support of an idle track, when climbing on or going down stairs. By means of mechanisms and actuators, it is possible to change the configuration of the wheelchair, from the flat ground motion to the stair climbing configuration.

A mobile robot employed in Search & Rescue tasks has to carry a wide set of sensors to get the most possible information from the environment and at the same time has to traverse hard uneven terrain, facing difficult stability conditions that could lead the robot to tip over and compromise definitively the mission. In this paper a stability assessment approach is applied to a chained UGV with manipulator driving along a determined path on uneven terrain. The stability from the beginning to the end of the path is predicted using prior knowledge of the terrain. The center of gravity of the whole robot is calculated statically for a finite set of positions along the path. The direct kinematics of the manipulator is used to find the contribution of the manipulator’s masses to the overall center of gravity. The polygon described by the contact points between the robot and the ground plays a critical role to determine how close is the vehicle to tip over.

Applications in the field of mobile robotics have high demands on flexibility and maneuverability of mobile platforms. Especially in logistics, vehicles have limited space for the movement. This paper presents the analysis and control of a mobile platform which uses a novel steering principle. The considered vehicle is able to perform a steered forward movement, lateral motion or pure rotation without the need of steering motors. The kinematics is analyzed and a kinematic model is derived. For simulation a dynamic model formulated in terms of redundant coordinates is used. A control scenario where reference values are commanded by a joystick is presented. For feedforward control design, a reduced dynamic model based on minimal velocities and the kinematic model are used. For feedback control, a cascaded structure with an inner velocity loop for the wheels and a superimposed steering control is used. The efficiency of the presented control approach is demonstrated by simulation results.

The paper presents the third version of the hybrid leg-wheel ground mobile robot Mantis, a small-scale platform designed for inspection and surveillance tasks. The locomotion system is based on the cooperating action of a couple of actuated front legs and wheels, along with a passive rear carriage. The system performs wheeled locomotion on even grounds and hybrid locomotion in case of terrain irregularities or obstacles. This architecture combines high speed, energy efficiency, maneuverability and stable camera vision on flat terrains with good motion capabilities in unstructured environments. In the embodiment design presented hereafter, referred to as Mantis 3.0, the rear carriage has been equipped with four passive wheels, instead of two as in the previous versions, in order to improve the stability during steep stair climbing maneuvers; moreover, the legs, the main body and the rear carriage have been significantly redesigned in order to be realized by additive manufacturing techniques, with the final aim of obtaining a low-cost device suitable for Open Source distribution.

This paper proposes a design concept for a novel robotic hand with a metamorphic palm mechanism. A design inspiration is elaborated from laboratory experiences on human grasping. The proposed design is presented with concepts and kinematic design that is based on linkages. A design model has been implemented in SolidWorks environment for a performance evaluation. Results are presented to show the feasibility and efficiency of the proposed solution.

Our pseudo-linear variable-lever variable stiffness actuator (PLVL-VSA) allows almost linear torque deflection characteristics in a compact form. However, in its original form, this advantage of the PLVL-VSA comes at a cost of increased stiffness variation torque, which limits the passive deflection range. In this work we propose a modification of our PLVL-VSA to solve this drawback, and thus extend its workspace, i.e., deflection range, while keeping the overall parameters of the actuators and the spring the same. We analyze the effect of the modification in a mathematical model and compare it to the original. The new configuration requires less torque and energy to vary the stiffness. The mechanism is also able to store a lot more elastic energy. As a drawback, the modification both introduces a discrete nonlinearity and reduces the pseudo-linearity of the torque-deflection graph. It also disables the use of very low stiffness setups.

This paper describes the mechatronic design of a gyroscopically stabilized bicycle. This bicycle is equipped with two flywheels rotating at constant speed in opposite direction. The flywheels can be rotated about the vertical axis of the bicycle by an additional drive unit, and due to the balance of angular momentum, a torque about the horizontal axis is generated that can be used for stabilization of the system. For the dimensioning of the overall system the equations of motion for a non-moving bicycle are used and an LQR controller for the stabilization is designed. Experimental results for this task are presented. Additionally the electrical design for this system is presented.

The integral part of any advanced robotic system is an effector dedicated to performing the specific tasks and the sensor system monitoring interaction with an environment. This complex mechatronic system determines basic characteristics of the robotic agent and the possible class of tasks (range of objects, kind and capacity/intensity of action) that will be able to perform. This paper briefly surveys some task related end effectors equipped with internal sensory systems providing supervision of states are used for aerial operations by UAV. The second part of the paper indicates a modular approach with exchangeable tools how to expand possibilities and capabilities towards making more complex tasks by multi-robot/agent systems. One approach to the design of the multi-purpose effector is described. The base of this concept is using the compliant tool changer/holding mechanisms that enable an exchange of various tools/detectors.

The monolithic design of serial robot arms has become widely accepted in industrial applications. One might note that these manipulators are designed in such a way that they can implement as many applications as possible at the same time and therefore do usually not have an application-specific structure. In this article we describe the construction and modeling of a kinematic chain in which a mechanical modification is possible. As a result, the structure of a serial robot can be adapted in a wide range. For this purpose, the links of the kinematic chain are modified in such a way that the robot structure matches the desired Denavit-Hartenberg parameter set. This means also that optimizations with different kinematic goals can be realized in order to flexibly adapt the robot system to a current task.

The design of innovative robotic devices reflects the behavior of the modern competitive markets, where the customer’s expectations dramatically increased. Many researches aim to apply innovative solutions to obtain high machining performances and quality of products. This paper deals to present a 3-DOFs parallel robotic device for miniaturized object manipulation and machining. The proposed device has multi-purpose applications and it consists of an actuation system based on preloaded piezoactuators with high dynamics and force, that automatically recover the workpiece distortions or positioning errors due to fixture set-up or machining vibrations. The study presents the main concept of the robotic device, highlighting the main functionalities and components, that impact on the system performance. Moreover, an analytical model is presented and validated using FE analysis, confirming the effectiveness of the adopted approach to control the device.

The paper presents the study of a hub+bearing system for tire pressure monitoring and control. The tire pressure control is of importance to improve vehicle safety and dynamic behavior, and to reduce fuel consumption and emissions. This feature becomes fundamental in the development of fully automated self-driving cars, which recently some important Companies started to study and test. Firstly, a commercial version of such a system was studied experimentally to evaluate the air-flow vs. pressure curves. Then, a 2D numerical model of the system, by using the commercial CFD software Fluent, was developed and a good match was found between the numerical and the experimental results. Subsequently, taking into account the most important geometric parameters that could influence the conductance and also considering the constraints connected to the bearing dimensions, different configurations of the hub+bearing system were studied numerically. Finally, to validate the CFD results different modular hub+bearing systems were built and experimentally tested. In this way, the critical zones in terms of pressure or velocity were detected and guidelines for designing a new system were obtained.

This paper presents the design of a new 2 degrees of freedom parallel manipulator and its theoretical implementation as a precise solar tracker. The cinematic mathematical model of the mechanism is studied in detail and a robust control algorithm based on sliding mode control is introduced to perform the desired angular positions and velocities in the presence of disturbances/uncertainties. The numerical simulations are carried out to verify the behavior of the suggested controller.

Having developed a hybrid optimization methodology for mechatronics system movements in space, the task is to efficiently control of robotic system with 6 degrees of freedom. In this work, hybrid methodology is compared with a simple genetic algorithm. It is applied in a simulation environment of a 6-degrees of freedom robotic arm and the results are compared with the mathematical model developed to support this methodology. The results of the research show that the optimization with the hybrid method compared to the simple GA, which is calculated by the mathematical model, is confirmed by more than 90% of the robotic arm simulation model examples.

The optimization of the task placement as well as the anatomy of a metamorphic manipulator in the presence of obstacles is presented. The rigid links of the metamorphic manipulator and the obstacles are represented using point clouds. An optimization problem is formulated, where the minimum manipulability index is maximized deriving a collision-free path using a genetic algorithm. A 3 DoF metamorphic manipulator is used as a case study.

The optimal positioning of tasks in robot applications is an extremely important step in the design of robotic work cells as it will allow the system to achieve the required high performance given the selected performance measure. In this work the optimal positioning of a robotic task is presented with the aim to minimize the required joint velocities required during task execution, for a 6 D.o.F. manipulator. The method is used to determine the optimal location for a path following task in the workspace of a UR-5 manipulator. Results show that the optimal task placement allows for a significant reduction of joint velocities to maintain a given constant end effector velocity during task execution.

In this paper, an optimization methodology for the structure and control optimization of a 3RRR planar parallel robot is presented. The proposal consists of three stages in cascade: firstly, we optimize the geometry for a maximum workspace. Secondly, the kinematics is used to optimize dexterity for a set of desired paths inside the workspace that is found in the first stage, and, finally, a set of dynamic control gains are optimized for trajectories given by the same paths. The methodology permits to reduce the computational cost for the geometry optimization stages, while optimizing the control gains using high precision numerical simulation using SimWise 4D commercial software, with a reduced number of evaluations of candidate solutions, and as consequence, a reduced computational time. The results demonstrate that the final structure-control optimized design accurately follows the desired trajectories.

The traditional “Receding Horizon Controller (RHC)” is a heuristic approach based on the concept of “Nonlinear Programming (NP)” that in the most general cases applies Lagrange’s “Reduced Gradient (RG)” method. Since its realization requires a huge amount of numerical calculations, in the practice it is often restricted to quadratic cost functions and “Linear Time Invariant (LTI)” approximation of the dynamic model of the controlled system as “Linear Quadratic Regulator (LQR)”. To release these restrictions a novel approach was recently invented that directly drives the gradient of the “auxiliary function” near zero by replacing the RG with a fixed point-based iteration. It was also shown that the same iteration technique allows the introduction of an “Adaptive Receding Horizon Controller (ARHC)”. Since the convergence of the ARHC strongly depends on the structure of the cost contributions in this paper the operation of the classic RG-based RHC is investigated in the control of the “TORA” system that is a popular paradigm for benchmarking purposes. Conclusions are drawn for the allowable or recommended parameter settings for the cost contributions.

Belt driven systems are part of many industrial devices, like cutting machines and 3D-printers. In this paper the dynamic modeling and a nonlinear control approach of a belt driven system containing a flexible beam are presented. Due to the special kinematics, the stiffness of the belt is nonlinear. This together with the elastic beam, leads to nonlinear partial differential equations of motion. By omitting negligible dynamical effects, a nonlinear controller can be designed that uses measurements of accelerations and shear forces of the flexible beam. In detail, an integrator backstepping design for an infinite dimensional system is performed. The proposed control approach is compared with a conventional PD control law and a validation by means of simulation results is shown.

This aim of this paper is to present a hyper-redundant crawling robot movement. The authors present a new physical structure and a control algorithm for movement in narrow labyrinth spaces, by using walls touching. Collisions and contacts are part of the task. Support-points, where the robot touches the walls, are used for its movement. Subsequently, a hybrid force/position controller for hyper-redundant locomotion is presented. Ultimately, the paper presents simulations of locomotion, positioning hyper-redundant evolutions and force tracking errors. An HHR robot used for applying the algorithm is built of identical elements that are interconnected through swivel joints. The curvature of the robotic arm is created with a cable system, actuated by DC motors. The position control is established through a pneumatic system. Using actuation cables and an electro-pneumatic system for blocking the movement of desired elements, the position of the robotic arm is obtained.

The problem of controlling robot manipulators of six joints carrying different objects (loads) is studied using a nonlinear controller with nonlinear static feedback and a nonlinear proportional and derivative precompensator. The system is described in a multi model nonlinear form. To solve the position control problem for all models via the same controller, the design requirement of common I/O decoupling with simultaneous common arbitrary command following is satisfied. Closed loop I/O stability is satisfied and performance bounds are retained.

This paper presents cascade control approach for simultaneous position and stiffness control of antagonistic actuators, which can be easily applied to other types of Variable Stiffness Actuators (VSA). The control design approach presented in this paper has two steps. The first step is tuning of inner loop PIDs for motor position control based on the second order dynamic model. The second step is adaptive controller design for fine tuning of system dynamics in different set points. Therefore, bank of controllers is formed and it is used to tune outer loop controllers’ for shaping position and stiffness references.

This paper deals with an inverted pendulum mounted on a sliding cart pneumatically actuated. Low cost digital pneumatic valves driven by PWM technique are used. The control system uses three nested PID regulators: the inner one controls the force imposed by the pneumatic actuator and the other two control the inclination of the pendulum and the position of the cart. The numerical model of the mechanical and pneumatic system is here shown, with also some first results.

This paper presents a new approach for resolving Jacobian singularities borrowing ideas from Software engineering and the Continuity principle of Leibniz. It considers kinematic motion control of redundant robot arms in case the work operation is assigned in terms of a geometrical path and the motion along it in task space. The proposed visualization of motion allows to better understand the motion of a redundant robot in terms of vector space methods. The Null space of the Jacobian is geometrically visualized in relation to a corresponding Null space of configurations. This geometrical representation resolves the velocity of the robot arm into two components. One of them is responsible for reconfiguring the robot arm motion in the Null spaces of configurations, while the other component moves the end effector along the prescribed path in task space. The visualization shows also that Jacobian singularities serve as “gates” between Null spaces of non-singular configurations. It enables the formulation of a numerical procedure for identifying and resolving Jacobian singularities. Unlike existing approaches this procedure employs the Principle of continuity to “predict” the velocity in the joints in a singular configuration and pass through such a configuration instead of avoiding it.

Conventional additive manufacturing approaches that build objects in horizontal layers using manipulators with 3 degrees of freedom (DOF) usually result in poor object strength in the vertical direction as well as poor surface quality. These issues can be mitigated by including curved layers in the manufacturing process. Additively building an object using curved layers requires manipulators with extra degrees of freedom as well as new path planning algorithms. We present a new approach to curved layer slicing by translating the geometrical problem of slicing to an optimization one, leading to better performance in terms of accuracy and speed than the existing methods. We also developed a new method of path planning on curved surfaces that ensures better inter-track bonding and consequently better strength and object quality compared to previous approaches. Finally, we implement and test our algorithms on a 6-DOF industrial robot.

The increasing complexity of today’s robot control systems brings more and more load on the control hardware. It is tempting, to relieve the classical main processor by shifting to a SoPC. In this paper, we show, how such a step can bring additional advantages, like high reactivity and real-time capability, especially in the use with behavior-based control. For this, we present a local path-planning algorithm, which we implemented on the FPGA-part of a SoPC and analyze its performance for the control of a 5-DOF robot arm.

A programmable platform and plastic hexagonal microparts, both with embedded circular conductive electrodes are studied. Two layouts for the allocation of the electrodes on the hexagons are proposed, according to which, activation algorithms are introduced for the horizontal and vertical displacement of the hexagonal microparts on the platform. The motion of the microparts based on the new actuation algorithms is implemented on a simulated platform and the results of microparts’ velocity are selected and discussed. Finally, the procedure for the assembly of hexagonal microparts is described in detail.

The Iterative Learning Control method (ILC) can compensate trajectory tracking errors caused by imprecise mathematical model of a robotic manipulator. Industrial robotic manipulators have various motion constraints and obstacles within their working space. This research investigates how an ILC method can be successfully applied when the robot’s workspace is constrained and the desired trajectory is planned to pass closely to any workspace limit. It takes advantage of the Bounded Error ILC method (BEILC) and adapts it for precise trajectory tracking with obstacle avoidance. In this case, BEILC enforces strict limits over the output trajectory which leads to a slow convergence rate. This paper proposes a new stop condition which relaxes the restrictions over the output trajectory. The performance of this new stop condition is then verified and evaluated through a computer simulation and experiments on a physical robotic arm manipulator. Those tests proved that the new stop condition improves the convergence rate and extends the application of the ILC methods when the control of industrial robotic manipulators is considered.

The learning process that arises in response to the visual perception of the environment is the starting point for numerous research in the field of applied and cognitive robotics. In this research, we propose a reinforcement learning based action planning algorithm for the assembly of spatial structures with an autonomous robot in an unstructured environment. We have developed an algorithm based on temporal difference learning using linear base functions for the approximation of the state-value-function because of a large number of discrete states that the autonomous robot can encounter. The aim is to find the optimal sequence of actions that the agent (robot) needs to take in order to move objects in a 2D environment until they reach the predefined target state. The algorithm is divided into two parts. In the first part, the goal is to learn the parameters in order to properly approximate the Q function. In the second part of the algorithm, the obtained parameters are used to define the sequence of actions for a UR3 robot arm. We present a preliminary validation of the algorithm in an experimental laboratory scenario.

In order to enable mobile robots to navigate autonomously in an environment shared with humans, special considerations are necessary to ensure both safe and efficient navigation. This work presents a predictive, human-aware motion controller, based on the Robot Operating System (ROS), which optimizes the vehicle trajectory at the control level with a high update rate. Predicting future positions of persons allows the system to optimize a trajectory around those predictions, yielding a sequence of motor controls for a smooth executed motion. The improvements were statistically evaluated using simulation runs in terms of travel duration, path length, and minimum distance to persons along the path. This way, we are able to show that our new motion controller performs significantly better in the presence of humans than a controller without human-awareness.

The Navigation Transformation proposed in [Loizou (2017)] provides a novel solution to the motion and path planning problems, while enabling temporal stabilization up to a set of measure-zero of initial conditions. Since sets of measure zero are explicitly defined for a given workspace, this work proposes an additional control action that steers the system trajectories away from such sets. The provided theoretical results are backed with experimental studies.

An agent based model - SkyBat, based on long-term observation of bats behaviour under fission-fusion dynamics, is presented in this paper. The agents cooperate while searching for specific targets of interest in an unknown area. Although the agents are autonomous, they have an ability to move from one location to another without a group leader and to react to changes in environment.

Underactuated robotic systems need suitable experimental methods able to measure their small and low-weight component dynamics. Depth sensors represent a valuable strategy to develop quantitative approaches to study the behavior of these systems. Here, an experimental application of markerless vision technique is proposed employing the low-cost and low-resolution Kinect depth sensor to compute the kinematics of an underactuated robotic finger.

The use of ultrasound scanning in rheumatology has been shown to be a sensitive and specific modality for assessing joint disease. It is used for both detection of early signs of disease and determination of arthritis activity in established disease. In order to improve the monitoring and treatment of arthritis patients, an automated system to scan the joints of the wrist and hand is developed. In this paper, a vision algorithm used to locate the joints from a camera is proposed. The joints are found by using a global approach combined with a geometric analysis of the image. The algorithm is used on a platform where a robot arm can move an ultrasound probe to the joints. The overall detection rate is $$94.2\%$$ with the majority of variance found along the finger.

The manufacturing industry is looking into more flexible solution for SMEs, in order to cope with few-of-a-kind production. Short set-up time and no need for experts is a requirement, which is why learning of assembly is done by kinesthetic demonstration. Teaching robot in-hand machine vision tasks requires positioning the robot-camera in the appropriate pose for image acquisition. The same camera is normally used for several inspections, consequently a focus distance is set on the lens. The operator must move the robot and the only feedback is the image from the camera. In this paper we proposed an autofocus control mechanism for a robot with a fixed focus camera. We evaluated the accuracy of the focus and time savings by using this method compared to manual positioning by an operator.

In this paper a reduced complexity eye tracking technique is introduced. The proposed technique is based on a recently introduced eye localization method [5]. In order to exploit its accuracy and robustness under difficult illumination conditions, the presence of occlusions, shadows and pose variations, and making possible its use in real-time applications, we drastically reduce its computational cost while in the same time its accuracy is increased, by the use of an efficient tracking scheme. We also implement the proposed method in low-level C++ programming, using the OpenCV library, thus reducing even more its computational cost. All experiments we have conducted confirm our claims.

This work deals with the graph-based semantic segmentation of a robot’s traversed environment using the Louvain algorithm. In recent years, semantic segmentation has been the focus of several researchers’ interest and is applied to a variety of robotic applications. The Louvain method for community detection is a novel technique for extracting communities from large networks. The method is a greedy optimization one with a complexity of O(nlogn). We first assessed the Louvain algorithm with the COLD-Freiburg dataset to create the semantic map and compute the communities. We demonstrate that lighting conditions do not affect the system’s ability to categorize places. In particular, we train the system with COLD Freiburg seg1cloudy1 and we query images from seg1sunny1 and seg1night1. The results exhibit that the system is capable of categorizing the evaluated frames into the correct community.

The detection of pre-visited areas in robots’ traversed path, widely known as loop closure detection, is vital for drift and position correction in robotic applications, such as simultaneous localization and mapping. In this paper, we present a sequence based approach for pose estimation, by advancing the well known SeqSLAM algorithm with the usage of Bag of Words (BoW) model. A visual vocabulary is produced in an offline procedure resulting in the system ’s ability to describe the incoming image stream by visual words, at the online process. Image similarity is achieved through BoW histogram comparisons instead of sum of absolute differences metric. Comparative results on several publicly-available datasets show the benefits of the proposed method offering high recall scores at 100% precision against the original one.

A vision system for online recognition of moving objects is considered. The system consists of several problem oriented image processing modules aimed to reveal and recognize characteristic patterns and to track the motion of the selected object. The system is working autonomously in the Search mode trying to offer recognized objects for tracking. In case of unsuccessful recognition, the operator can point out an object on the screen to start the Track mode. A calculation of the Pearson’s Cross Correlation Coefficient between the pattern and the corresponding area of interest under the pattern has been decomposed and efficiently implemented in the Fast Fourier Transform domain in order to localize the tracked object in real time. The aim of the paper is to define an effective normalizing procedure in spatial domain for the cross correlation matrix computed in the transform domain.

A layer extraction method aiming at robotic garment unfolding is presented in this paper. Utilizing depth sensor input, the layer detection is achieved through a depth first algorithm and a simple perceptron. An action space based on the edges detected on the garment and their features restricts the complexity of the algorithm and accelerates the detection. This paper constitutes part of the robotic garment unfolding pipeline and emphasizes on the extraction of the upper layer of clothing articles that result in a half folded, two-layer planar state after robotic manipulations. The presented methodology is independent from the type of the handled garment and can cope with unknown designs. First results seem promising and encourage future work on the field.

This paper presents an analysis of the number and the distribution of industrial robots in the Republic of Croatia. Also, the actual state of industrial robotics in the world is given, with the present and future growth trends, the distribution of robots by countries and manufacturing sectors. The number of robots in Croatia was obtained on the basis of a survey questionnaire sent to 1,500 Croatian companies. Regarding the question of robot ownership, 72 companies answered positively, resulting in a total of 326 active industrial robots in 2017. According to the Croatian Chamber of Economy estimates, the Croatian economy should have at least 2,000 installed robots. The paper gives a prediction for the growth trend in order to achieve 2,000 robots in a reasonable time period (by 2026). For that case, an exponential growth rate of 25.4% is required. Based on the current state of the Croatian economy, such an exponential growth is a huge challenge for the near future. The paper gives a brief critical review of the current state of industrial robotics in Croatia and provides guidelines for stimulating the application of industrial robots in the near future.

The paper extends previous developments of cloud robot services for intelligent manufacturing with new data streaming and machine learning techniques that are used to dynamically reschedule resources and predict future behaviour on the shop floor. Data is obtained in real-time with edge computing solutions that ease the computing effort in the cloud, by moving intelligence to the edge of the manufacturing execution system. Thus, machine learning algorithms can be run in real-time context with re-training on new data; the insights become predictions, enabling real-time decisions for: operations scheduling, robot allocation and real status-based maintenance. Experiments are described.

A new smart cyber-physical system (CPS) specially designed to further improve the flexibility of production systems and to improve the collaborative capabilities of industrial service robots designed for interactive work with humans in an information structured work space is presented in the paper. This system enables quick and easy reconfiguration of the technological production process in accordance with changes caused by the small production batches or frequent changes of the production program. As such, the system is extremely suitable for use in small and medium production enterprises (SMEs). The new cyber-physical system is based on a mobile, dual-arm industrial robotic system with the extended working space and a smart application-specific open-type software interface that can be updated and upgraded with new knowledge and skills from the production line according to the technological needs. The basic functions the mechanical modules and cloud-based control architecture are presented in this paper.

The maintenance of the aircraft finish system is executed completely manually at present, involving a big amount of manual labor for a long time and in a hazardous environment. The automation of the process would be able to dramatically speed it up and to decrease manpower involved. Moreover, costs and environmental risks are expected to reduce with the automation. Several solutions are being developed; nevertheless, a system able to achieve the maintenance process automatically is not yet available. Along this thesis, a preliminary design of an automatic system for aircraft painting and paint removal has been carried out. The work points out that a cost effective solution for this complex problem is possible. As a preliminary study, this is intended to be a starting point for further study on this subject.

Reconfigurable manufacturing systems (RMS) provide means to deal with changes and uncertainties in highly dynamic production processes. They allow for a relatively quick adjustment of various modules within the production line. To further increase the flexibility of such systems, multiple robots can be used within. Multi-robot systems provide a higher degree of flexibility and efficiency compared to single-robot systems. These systems can perform tasks that require a high level of dexterity. However, in order to ensure the robots are able to precisely perform cooperative tasks, it is necessary to have a well calibrated system. In this paper, we present a novel approach for robot base frame calibration by exploiting the collaborative robots’ kinesthetic guidance feature. The developed method is suitable in RMS, as it is more time efficient and intuitive without drawbacks in precision.

Determination of an object’s position in a given reference frame is the main purpose of a navigation system. This can be done in many different ways and depending on the chosen method, and measurement equipment, produce more or less accurate measurements. Precise indoor navigation is particularly important due to the ever more dynamic development of autonomous systems in many areas of industry. Unfortunately, the measurement accuracy of indoor navigation systems is reduced due to the influences of walls and other obstacles that interfere with the measurement signals causing so called multipathing. Multipathing is often reduced by creating error maps, which is a labor-intensive task. In this paper, we present a method in which a robot manipulator is used as the reference positioning system to determine such mapping. In the next step the mapping, using a second order polynomial, is determined which maps the measured disturbed object’s position from the triangulation system into real object’s position.

Future use cases for stationary robot manipulators envision shared human-robot workspaces. However, shared workspaces may contain a priori unknown obstacles (e.g. humans). Robots must take these obstacles into account when moving (e.g. through online path planning). To this end, current research suggests real-time workspace monitoring with a calibrated multi-camera system. State-of-art solutions to camera calibration exhibit flaws in the above scenario, including long calibration times, excessive reprojection errors, or extensive per-calibration efforts. In contrast, we contribute an approach to multi-camera calibration that is at once efficient, precise, and convenient: We perform fully-automated calibration of each camera with a robot-mounted calibration object. Subsequent multi-camera optimization equalizes reprojection error over all cameras. After initial setup, experiments attest our contribution minor reprojection errors in few minutes time at one button click. Overall, we thus enable frequent system (re-)calibration (e.g. when moving cameras).

Air pads are often embedded into robotic devices for manufacturing and measuring applications. Investigating the performance of these pads is essential to obtain robot characterized by very accurate motions and positioning. Due to their simplicity, mathematical lumped parameter models can be adopted to evaluate quickly both the static and dynamic performance of such bearings. This paper describes a lumped parameter model to study the behavior of a grooved rectangular air pads. The static and dynamic results of the model are validated by using a purpose-built test bench.

Social Robots is one type of cyber-physical systems, that is the social equivalent of “industry 4.0” technology, in applications involving humans e.g. in businesses of services. Our interest here is in applications of Social Robotics in education. This paper provides a road map regarding commercial social robots currently available in education. Recent literature is included regarding (a) analysis and evaluation of the effectiveness of social robots in education in terms of design specifications such as processors, sensors etc., (b) advantages and drawbacks of various robots currently used in education in terms of cost, impact and usability and (c) future potential directions of interest concerning educational robotics. Our study indicates that an effective design of interactive, educational robots calls for robustness and standardization, of both hardware and software. Novel modeling methodologies might be necessary. Future challenges in the field are also discussed.

One of the paradoxes of the present era is that despite the popularity of computer systems in all its forms, the interest of young people in the study of technical sciences, including programming, is diminishing. Technical universities must make a lot of effort to attract a sufficient number of applicants. For this reason, the MYrobot platform was created, which is presented in this article. MYrobot is a mobile platform that can carry a smartphone which provides a powerful robotic system. The developer of an application can use all smartphone components - an accelerometer, a microphone, a camera, etc. Advanced users can also develop expansion boards with various other sensors. The kit’s programming options have been enhanced by a graphical development system that allows developing of simple applications even by primary school pupils. An interactive form of platform programming has the potential to attract those students who are “afraid” of a flashing cursor.

Extensive use of drones raises many new questions regarding ethics and morality. Questions concern the civilian and military use of drones. Drones can serve as a mobile network that can reach places, where other devices or people can’t get to or can only get to with difficulties. They can efficiently collect data from large and hard-to-reach areas. They are often used for scanning and exploring forest and agricultural areas. Together with advanced image and scene recognition methods, they can greatly reduce hard work and help reduce stress in crop growth and protect forests from infestation. Other application areas are e.g. for archaeologists, when exploring remote areas. The speed and efficiency of the capture of a scenenery is a great benefit of this new technology. Many people also use drones for recreational purposes. Therefore, the ethical and legal problems associated with their widespread use should be emphasized.

In this paper, several kinect-based games which are developed and implemented for the improvement of attention and sensory-motor coordination will be presented. The interface, and difficulty levels of these games are specially designed for the ease of different age groups. The games involve physical activities for the fulfillment of some basic tasks within the Human-Computer Interaction (HCI) game, such as fruit picking and air hockey, with different difficulty levels based on varying parameters of the games. The human action is observed and recognized via Kinect RGB-D sensors. The games are tested with a group of deaf children (3.5–5 years) as a part of the experiments of an ongoing project, to decrease the stress of the children, and increase their enjoyment, attention and sensory-motor coordination before the main tests. Both the game results and the evaluation of the therapists and the pedagogues show that the games have a positive impact on the children. The games are also tested with a group of adults as a control group and the attention levels of the adults were also observed via mobile EEG device. The children were supposed to use the EEG device in the main tests therefore the device was not integrated to their game sessions.

This research explores the impact of educational robotics on children aged 9–15. The research was conducted after a training robotics seminar and, as evidenced by its results, the learners were thrilled by the content of the program, although they initially mistrust the program because of relative ignorance. However, they seemed happy that they met new friends and worked with them without any particular difficulty. The use of computer as well as the programming did not present any difficulty. The research concluded that older children were more familiar with concepts that are directly related to programming, obviously because they are every day involved with technology and less with robotics. Of course, with regard to the construction of various vehicles, the trainees showed their impatience and inventiveness for this and were extremely effective. It is therefore obvious that through educational robotics, children can learn to cooperate more effectively with each other and that the teaching of basic principles of computer science, mathematics, geometry, physics, mechanics, and in general mechatronics, can escape the narrow limits of conventional teaching and to take the form of a game.

The present paper tries to emphasize the importance of STEM education in the primary and secondary school, as well as the use of educational software in robotics taught in high schools and universities. Several European and wide world current trends in educational robotics are reviewed.