Sanitizing railway stations is a relevant issue, primarily due to the recent evolution of the Covid-19 pandemic. In this work, we propose a multi-robot approach to sanitize railway stations based on a distributed Deep Q-Learning technique. The proposed framework relies on anonymous data from existing WiFi networks to dynamically estimate crowded areas within the station and to develop a heatmap of prioritized areas to be sanitized. Such heatmap is then provided to a team of cleaning robots - each endowed with a robot-specific convolutional neural network - that learn how to effectively cooperate and sanitize the station's areas according to the associated priorities. The proposed approach is evaluated in a realistic simulation scenario provided by the Italian largest railways station: Roma Termini. In this setting, we consider different case studies to assess how the approach scales with the number of robots and how the trained system performs with a real dataset retrieved from a one-day data recording of the station's WiFi network.

Stabilising an inverted pendulum on a cart is a well-known control problem. This paper proposes the mechan-ical and control design for solving the oscillation problem of a variable-length flexible beam mounted on a mobile robot. The system under consideration is the robot PovRob, used at the European Organization for Nuclear Research (CERN) for visual and remote inspection tasks of particle accelerators. The flexible beam mounted on the robot houses cameras and sensors. The innovative aspect of the approach concerns the use of actuated masses mounted at the end of the rod, which induces an impulsive moment due to their inertia and angular acceleration. The modelling of the flexible rod has been suitably simplified in a lumped-parameter system, with dynamic parameters related to the rod's flexibility. A linearisation of the dynamic model allows a linear-quadratic control to stabilise the system. Experimental results support the identification and the validation of the dynamic model, while simulation results evaluate the performances of the designed control law.

This paper presents a novel control approach, based on the interconnection and damping-assignment passivity-based control (IDA-PBC), to achieve stable and periodic walking for underactuated planar biped robots with one degree of underactuation. The system's physical structure is preserved by assigning a target port-Hamiltonian dynamics to the closed-loop system, which also ensures passivity. The control design ensures that the tracking error to the desired periodic gait converges exponentially to zero, and the convergence rate can be adjusted via gain tuning. Besides, through the hybrid zero dynamics, the stability of the full-order system can be retrieved from the stability of the orbit created in a lower-dimensional manifold. The proposed approach is the first example of a tracking controller based on the IDA-PBC applied to underactuated biped robots. Numerical simulations on a five-link planar biped robot with unactuated ankles validate the approach and show the performance of the closed-loop system.

A compass-like biped robot can go down a gentle slope without the need of actuation through a proper choice of its dynamic parameter and starting from a suitable initial condition. Addition of control actions is requested to generate additional gaits and robustify the existing one. This paper designs an interconnection and damping assignment passivity-based control, rooted within the port-Hamiltonian framework, to generate further gaits with respect to state-of-the-art methodologies, enlarge the basin of attraction of existing gaits, and further robustify the system against controller discretization and parametric uncertainties. The performance of the proposed algorithm is validated through numerical simulations and comparison with existing passivity-based techniques.

Cobots (Collaborative robots) are increasingly employed in industrial workplaces, assisting human operators in decreasing the weight and the repetitiveness of their activities. In order to promote the acceptance of cobots, novel interaction paradigms enabling intuitive collaboration between humans and robots are needed. In this work, we propose a variable admittance control framework based on virtual fixtures, in which the damping and the stiffness of the admittance controller are on-line adjusted to increase the effectiveness of the co-manipulation task. Virtual paths are generated to support and guide the operator during collaborative task execution, providing him/her an over-responsive compliant system when the robot is far from the targets and a precise heavy tool in their proximity. The proposed approach has been compared with a fixed admittance controller in an industrial use case consisting of a human operator interacting with a Kuka LBR IIWA arm. The collected results demonstrate the effectiveness of the proposed approach.

This paper presents a new procedure to design a control law using the classical interconnection and damping assignment technique within the passivity-based port-Hamiltonian framework. The sought goal is to reduce the complexity of solving the so-called matching equations. The proposed approach is applied to two case studies of planar rolling nonprehensile manipulation, namely, the ball-and-beam and the eccentric disk-on-disk. The performance of the resulting controllers is illustrated through both simulations and experimental results, showing the applicability of the design in a real setup.

This paper presents a method for tracking multiple interacting deformable objects undergoing rigid motions, elastic deformations and contacts, using image and point cloud data provided by an RGB-D sensor. A joint registration frame-work is proposed, based on physical Finite Element Method (FEM) elastic and interaction models. It first relies on a visual segmentation of the considered objects in the RGB images. The different segmented point clouds are then processed to estimate rigid transformations with on an ICP algorithm, and to determine geometrical point-to-point correspondences with the meshes. External forces resulting from these correspondences and between the current and the rigidly transformed mesh can then be derived. It provides both non-rigid and rigid data cues. A classical collision detection and response model is also integrated, giving contact forces between the objects. The deformations of the objects are estimated by solving a dynamic system balancing these external and contact forces with the internal or regularization forces computed through the FEM elastic model. This approach has been here tested on different scenarios involving two or three interacting deformable objects of various shapes, with promising results.

We present a novel multimodal interaction framework supporting robust human-robot communication. We consider a scenario where a human operator can exploit multiple communication channels to interact with one or more robots in order to accomplish shared tasks. Moreover, we assume that the human is not fully dedicated to the robot control, but also involved in other activities, hence only able to interact with the robotic system in a sparse and incomplete manner. In this context, several human or environmental factors could cause errors, noise and wrong interpretations of the commands. The main goal of this work is to improve the robustness of human-robot interaction systems in similar situations. In particular, we propose a multimodal fusion method based on the following steps: for each communication channel, unimodal classifiers are firstly deployed in order to generate unimodal interpretations of the human inputs; the unimodal outcomes are then grouped into different multimodal recognition lines, each representing a possible interpretation of a sequence of multimodal inputs; these lines are finally assessed in order to recognize the human commands. We discuss the system at work in a case study in which a human rescuer interacts with a team of flying robots during Search & Rescue missions. In this scenario, we present and discuss real world experiments to demonstrate the effectiveness of the proposed framework.

In this paper, we present the design of a robust interconnection and damping assignment controller for a rolling-balancing system known as the disk-on-disk. The underactuation feature of this system hampers the control design, and since we consider matched disturbances, the problem becomes even more challenging. To overcome this difficulty, we propose to design first a controller to stabilize the desired equilibrium of the case where the disturbance is not present, and then we robustify this controller by adding a nonlinear PID outer loop that compensates the disturbance. Finally, we evaluate the practical applicability of the control design by implementing the controllers on a real hardware for the disk-on-disk system.

This paper presents an optimal trajectory planner for a robotic batting task . The specific case of a table tennis game performed by a robot is considered. Given an estimation of the trajectory of the ball during the free flight, the method addresses the determination of the paddle configuration (pose and velocity) to return the ball at a desired position with a desired spin. The implemented algorithm takes into account the hybrid dynamic model of the ball in free flight as well as the state transition at the impact (the reset map). An optimal trajectory that minimizes the acceleration functional is generated for the paddle to reach the desired impact position, velocity and orientation. Simulations of different case studies further bolster the approach along with a comparison with state-of-the-art methods.

In this work, we demonstrate how autonomous pizza tossing and catching can be achieved. Under the assumption that the pizza dough is grasped by a number of fingers with soft contact, we formulate the grasp constraints and use them to derive the individual and combined Euler-Lagrange dynamic equations of motion of the robotic manipulator and the dough. In particular, the dynamics of the dough is a modified version of the rigid-body dynamics, taking into account the change of inertia due to its deformation. Armed with these mathematical models, we tackle the two control problems of tossing and catching. For the tossing phase, we derive an exponentially convergent controller that stabilizes a desired velocity of the dough as it is let go. On the other hand, so as to catch the dough, we generate an optimal trajectory for the end-effector of the robotic manipulator. Finally, we derive control laws to make the optimal trajectory exponentially attractive. We demonstrate the developed theory with an elaborate simulation of the tossing and catching phases.

We present a system suitable for human multi-robot interaction that supports the operator in the robot selection process. The proposed framework allows a human to issue commands to a robotic team without an explicit robot selection in so enabling a fluent interaction. This work is framed in the operative context of the SHERPA project [1], which proposes the deployment of a robotic platform for Search & Rescue in an alpine scenario and assumes the presence of a human rescuer that can orchestrate the robots operations with multimodal commands. In this context, implicit robot selection is mainly motivated by fast communication and the difficulties to distinguish different robots of similar shape in a hazardous environment and in adverse weather conditions. In the proposed approach, each robot of the team can evaluate the probability to be referred in an incomplete command, considering its actual capabilities along with geometrical and contextual information. We describe the overall system architecture focusing on the human intention recognition process. The proposed framework is trained and evaluated in a simulated case study.

We present a multimodal attentional interface suitable for a human operator that monitors and controls the activities of a team of drones during search and rescue missions. We consider a scenario where the operator is a component of the rescue team, hence not fully dedicated to the robots, but only able to interact with them with sparse and incomplete commands. In this context, an adaptive interface is needed to support the user situation awareness and to enable an effective interaction with the drones. In this work, we propose a multimodal attention-based interface designed for this domain. This framework is to filter the information flow towards the operator selecting and adapting the communication mode according to the context and the human state. We illustrate the features of the adaptive system along with an initial assessment in a simulated scenario.

This paper introduces a method able to track in real-time a 3D elastic deformable objects which undergo fractures, using the point cloud data provided by an RGB-D sensor. Our framework relies on a prior visual segmentation of the object in the image. The segmented point cloud is registered by non-rigidly fitting the mesh, based on the Finite Element Method to physically model elasticity, and on geometrical point-to-point correspondences to compute external forces exerted on the mesh. Fractures are handled by processing the stress tensors computed on the mesh of the FEM model, in order to detect fracturable nodes. Local remeshing around fracturable nodes is then performed to propagate the fracture. The real-time performance of the system is demonstrated on real data involving various deformations and fractures.

This paper proposes a new visual controller to endow flying manipulators with the capability of cooperatively and automatically positioning an assembly part on a visual target. We consider two VToL UAVs each equipped with a robot manipulator and a video camera. The two manipulators are rigidly connected to an assembly part, consisting of bar in this case study, that needs to be automatically placed on its target pose. The control system we propose uses the images relayed from the two onboard cameras to simultaneously control the motions of the UAVs and the manipulators to achieve the task: positioning the bar as well as respecting the closed-chain constraint. The controller evolves in the image space, thus inheriting robustness with respect to calibration errors. It takes into account the under-actuation inherent to rotary-wing aerial vehicles as well as the redundancy of the whole system. Furthermore, by leveraging image moments the 3D relative motion between the carried assembly part and the environment is controlled, thus yielding to natural and practically-relevant movements. Numerical simulations have been carried out to verify the validity of the proposed approach.

A multilayer architecture to control multirotor UAVs equipped with a servo robot arm is proposed in this paper. The main purpose is to control the aerial platform taking into account the presence of the moving manipulator. Three layers are considered in this work. First, a novel mechanism is proposed considering a moving battery to counterweight the statics of the robotic arm. Then, in order to overcome the mechanical limitations of the previous layer, the residual of the arm static effects on the UAV is computed and compensated through the given control thrust and torques. Finally, an estimator of external forces and moments acting on the aerial vehicle is considered and the estimations are fed back to the controller to compensate neglected aerodynamic effects and the arm dynamics. The performance of the proposed architecture has been experimentally evaluated.

This paper presents a method to track in real-time a 3D textureless object which undergoes large deformations such as elastic ones, and rigid motions, using the point cloud data provided by an RGB-D sensor. This solution is expected to be useful for enhanced manipulation of humanoid robotic systems. Our framework relies on a prior visual segmentation of the object in the image. The segmented point cloud is registered first in a rigid manner and then by non-rigidly fitting the mesh, based on the Finite Element Method to model elasticity, and on geometrical point-to-point correspondences to compute external forces exerted on the mesh. The real-time performance of the system is demonstrated on synthetic and real data involving challenging deformations and motions.

In this paper, we propose a nonlinear controller that stabilizes unmanned aerial vehicles in GPS-denied environments with respect to visual targets by using only onboard sensing. The translational velocity of the vehicle is estimated online with a nonlinear observer, which exploits spherical visual features as the main source of information. With the proposed solution, only four visual features have shown to be enough for the observer to operate in a real scenario. In addition, the observer is computationally light with constant numerical complexity, involving small-dimension matrices. The observer output is then exploited in a nonlinear controller designed with an integral backstepping approach, thus yielding a novel robust control system. By means of Lyapunov analysis, the stability of the closed-loop system is proved. Extensive simulation and experimental tests with a quadrotor are carried out to verify the validity and robustness of the proposed approach. The control system runs fully onboard on a standard processor, and only a low-cost sensing suite is employed. Tracking of a target whose speed exceeds 2 $\mathrm{m/s}$ is also considered in the real-hardware experiments.

An estimator of external generalized forces (force plus moments) acting on aerial platforms, and based on the momentum of the mechanical system, is proposed for the control of VToL UAVs together with a hierarchical architecture separating the translational and rotational dynamics of the vehicle. The closed-loop system equations are shaped as mechanical impedances, programmable through the controller gains, and forced by the residuals given by the estimation error. This arrangement allows the VToL UAVs to perform hovering and tracking tasks without a precise knowledge of the vehicle dynamics and in presence of external disturbances and unmodeled aerodynamic effects. Experiments are presented to evaluate the performance of the proposed control design.

We present an approach to mixed initiative control for unmanned aerial vehicles (UAVs) where sliding autonomy is supported by mixed-initiative planning and haptic feedback. In the proposed framework, we assume that an autonomous system can plan and execute robotic tasks while a human operator can provide interventions when necessary receiving a force feedback. The haptic feedback is associated with the sensation about how the system is diverging from the planned operations. We tested the system at work in virtual and real environments considering simple navigation tasks. We compared the performance of human operators with or without the assistance of the force feedback. The collected results support the hypothesis that the proposed approach enables effective and intuitive mixed-initiative control.

A backstepping approach is proposed in this paper to cope with the failure of a quadrotor propeller. The presented methodology supposes to turn off also the motor which is opposite to the broken one. In this way, a birotor configuration with fixed propellers is achieved. The birotor is controlled to follow a planned emergency landing trajectory. Theory shows that the birotor can reach any point in the Cartesian space losing the possibility to control the yaw angle. Simulation tests are employed to validate the proposed controller design.

Manipulation tasks carried out with aerial platforms composed of a UAV and a robotic arm involve cross-coupled dynamics between these subsystems. This paper proposes a new controller for this class of aerial robotic systems that allows regulating on velocity commands generated by an outer image-based visual-servo scheme. The controller, that considers the full dynamics of the system, is designed based on the integral backstepping approach. Visual feedback provided by an onboard camera is employed into a new visual servo scheme to simultaneously generate velocity commands for the UAV and the manipulator so that a visual target is reached. The control system takes into account the under-actuation related to rotary-wing vehicles, while at the same time it exploits the functionality system redundancy to achieve the task. Simulation results validate the proposed control system, as well as its robustness to large modeling error and measurement noise.

This video presents experiments conducted within the final review meeting demonstration session of the AIRobots project. AIRobots started at 2010 and the final review meeting took place on 22 of March, 2013. The presented experiments cover a wide area of the challenges related with aerial industrial inspection. In particular, multiple test-cases related with both vision-based and contact-based inspection and in general physical interaction are shown. It is highlighted that these experiments were recorded live during the project demonstration and evaluation process.

Motion estimation is an open research field in control and robotic applications. Sensor fusion algorithms are generally used to achieve an accurate estimation of the vehicle motion by combining heterogeneous sensors measurements with different statistical characteristics. In this paper, a new method that combines measurements provided by an inertial sensor and a vision system is presented. Compared to classical modelbased techniques, the method relies on a Pareto optimization that trades off the statistical properties of the measurements. The proposed technique is evaluated with simulations in terms of computational requirements and estimation accuracy with respect to a classical Kalman filter approach. It is shown that the proposed method gives an improved estimation accuracy at the cost of a slightly increased computational complexity.

The computation of the grasping forces for a multiarm robotic manipulation system (e.g., an anthropomorphic bimanual system) is considered in this paper. This problem is formulated as a convex optimization problem, also considering joint torque constraints. An algorithmic solution suitable for online implementation is presented, which allows a substantial reduction in the computational load by adopting a compact formulation and dynamically decreasing the number of active torque constrains. Moreover, for the case of a bimanual manipulation system, a suboptimal single-hand optimization algorithm is proposed and compared with that providing the optimal solution. Finally, a new algorithm for a valid initial-point evaluation is proposed. The effectiveness of the described methods has been tested in a simulation case study where the grasping forces of a humanoid torso equipped with two five-finger robotic hands are modified online to handle a load with a time-varying mass.

In this paper, a 3D robotic ball catching algorithm which employs only an eye-in-hand monocular visual-system is presented. A partitioned visual servoing control is used in order to generate the robot motion, keeping always the ball in the field of view of the camera. When the ball is detected, the camera mounted on the robot end-effector is commanded to follow a suitable baseline in order to acquire measurements and provide a first possible interception point through a linear estimation process. Thereafter, further visual measures are acquired in order to continuously refine the previous prediction through a non-linear estimation process. Experimental results show the effectiveness of the proposed solution.

The problem of grasping force optimization for a robotic system equipped with multi-fingered hands is considered in this paper. This problem is cast as a convex optimization problem, considering also joint torque constraints. A solution suitable for on-line implementation is proposed, which allows a substantial reduction of the computational load by dynamically decreasing the number of active torque constraints. Moreover, for the case of a bimanual manipulation system, a sub-optimal single-hand optimization algorithm is presented and compared with the optimal one. The effectiveness of the described methods has been tested in a simulation case study.

An incremental clustering technique to partition 3D point clouds into planar regions is presented in this paper. The algorithm works in real-time on unknown and noisy data, without any initial assumption. An iterative cluster growing technique is proposed in order to correctly classify a flow of 3D points and to merge close regions. The computational efficiency of the approach is achieved by using an Incremental Principal Component Analysis (IPCA) technique, and with the adoption of a compact geometrical representation based on the concave-hull computation of each cluster. This solution adds a more realistic representation of the observed environment and reduces the number of points needed to identify the cluster shape. The effectiveness of the proposed algorithm has been validated with both synthetic and real data sets.

A Cartesian impedance control for UAVs equipped with a robotic arm is presented in this paper. A dynamic relationship between generalized external forces acting on the structure and the system motion, which is specified in terms of Cartesian space coordinates, is provided. Through a suitable choice of such variables and with respect to a given task, thanks to the added degrees of freedom given by the robot arm attached to the UAV, it is possible to exploit the redundancy of the system so as to perform some useful subtasks. The hovering control of a quadrotor, equipped with a 3-DOF robotic arm and subject to contact forces and external disturbances acting on some points of the whole structure, is tested in a simulated case study.

An autonomous wall inspection control based on a stereo optical flow, suitable for unmanned aerial vehicles endowed with a stereo vision system, is proposed in this paper. The inspection task consists of simultaneously controlling the inspection velocity along the surface, the relative yaw angle between the vehicle and the observed plane, as well as the orthogonal distance. A virtual spherical camera is considered at the center of gravity of the vehicle. Then, a stereo optical flow, as if it had been acquired by the virtual camera, is generated from the visual data provided by the stereo vision system. The 3D visual measurements are also employed to estimate the relative position and orientation of the observed plane. Hence, the absolute vehicle velocity is estimated by using a robust translational average optical flow by integrating the total stereo flow. Finally, an inspection control and a hovering control are proposed. The effectiveness of the described approach has been demonstrated with a dynamic simulation in an environment composed of two adjacent walls.

In this paper, a new algorithm for online grasping force optimization (GFO) of a dextrous robotic hand is presented. The GFO problem is cast in a convex optimization problem, considering also torque joint constraints. The proposed formulation allows to simplify the computational complexity of the problem by dynamically reducing the number of active torque constraints. Moreover, differently from other approaches, it does not require the evaluation of a new initial point at the beginning of each iteration. The effectiveness and the performance of the proposed method have been tested in a simulation case study where the hand manipulates a load with time-varying mass.

In this paper, a kinematic model for motion coordination and control of a redundant robotic dual-arm/hand system is derived, which allows to compute the object pose from the joint variables of each arm and each finger as well as from a suitable set of contact variables. This model is used to design a two-stage control scheme to achieve a desired object motion and maintain desired normal contact forces applied to the object. Several secondary tasks are accomplished through a prioritized task sequencing management of the whole system redundancy. A simulation case study is presented to demonstrate the effectiveness of the proposed approach.

In this paper a method for fast visual grasp of unknown objects with a multi-fingered hand is presented. The algorithm is composed of an object surface reconstruction algorithm and a local grasp planner, evolving in parallel. The former uses an elastic reconstruction surface, whose dimensions are assigned initially by a preshaping process, and which shrinks toward the object until some parts of the surface intercept the object visual hull. The latter moves the fingertips on the current available reconstruction surface towards points which are optimal (in a local sense) with respect to a certain number of indices weighting both the grasp quality and the kinematic configuration of the hand. Experiments are presented, showing the effectiveness of the proposed algorithm.

A points surface reconstruction algorithm of 3D object models from multiple silhouettes is proposed in this paper. Some images of the target object are taken from a circular trajectory by a robot with a camera mounted in an eye-in-hand configuration. The silhouettes of the observed object are evaluated for each view using a blob analysis process, and from those a set of points that sample a reconstruction sphere surrounding the target object are estimated. The sphere sample points are attracted by the object center of mass using a variable step according to the distance from the silhouettes contours. For each point, the iterative process of constriction is stopped when all the back-projections of the point are within the corresponding silhouettes. Moreover, a new method based on a rough estimation of object dimension is proposed to reduce the disturbances due to projection and shadow cones. Simulations and experiments are presented to evaluate the performance of the proposed algorithm.

A new method for fast visual grasp of unknown objects using a camera mounted on a robot in an eye-in-hand configuration is presented. The method is composed of a fast iterative object surface reconstruction algorithm and of a local grasp planner, evolving in a synchronized parallel way. The reconstruction algorithm makes use of images taken by a camera carried by the robot. A reconstruction sphere, virtually placed around the object, is iteratively compressed towards the object visual hull, dragging out the fingers attached to it. Between two steps of the reconstruction process, the planner moves the fingers, floating on the current reconstructed surface, according to suitable quality measures. The fingers keep moving until a local minimum is achieved, then a new object surface estimation provided by the reconstruction process is considered. Quality measures considering both hand and grasp proprieties are adopted. Simulations are presented to show the performance of the proposed algorithm.

In this paper, an approach to interaction control of a robot manipulator with a partially known environment is proposed. The environment is a rigid object of known geometry but of unknown and possibly time-varying position and orientation. An algorithm for online estimation of the object pose is adopted, based on visual data provided by a camera as well as on forces and moments measured during the interaction with the environment. This algorithm is embedded into an impedance control scheme, resulting in a position-based visual impedance control. Experimental results are presented for the case of an industrial robot manipulator in contact with a planar surface.

This paper deals with the problem of position-based visual servoing in a multiarm robotic cell equipped with a hybrid eye-in-hand/eye-to-hand multicamera system. The proposed approach is based on the real-time estimation of the pose of a target object by using the extended Kalman filter. The data provided by all the cameras are selected by a suitable algorithm on the basis of the prediction of the object self-occlusions, as well as of the mutual occlusions caused by the robot links and tools. Only an optimal subset of image features is considered for feature extraction, thus ensuring high estimation accuracy with a computational cost independent of the number of cameras. A salient feature of the paper is the implementation of the proposed approach to the case of a robotic cell composed of two industrial robot manipulators. Two different case studies are presented to test the effectiveness of the hybrid camera configuration and the robustness of the visual servoing algorithm with respect to the occurrence of occlusions

An algorithm for the estimation of the position and orientation of a moving object using a hybrid eye-in-hand/eye-to-hand multi-camera system is presented. Based on the extended Kalman filter, this approach exploits the data provided by all the cameras without "a priori" discrimination, allowing real-time estimation. The proposed formulation can be used with different kinds of image features and different representations of the object orientation. A simulation case study is reported to test the feasibility and the effectiveness of the proposed technique

The problem of interaction control of a robot manipulator with a partially known environment is considered. The environment is a rigid object of known geometry but of unknown and possibly time varying position and orientation. An algorithm for online estimation of the object pose is presented, based on visual data provided by a camera. The estimation accuracy is improved during the interaction by using also force and joint position measurements. The proposed method can be exploited to implement any kind of interaction control strategy. A simulation case study is considered for the case of impedance control of a robot manipulator in contact with a planar surface

An environment for open real-time control of an industrial robotic cell is presented in this paper. The experimental setup is composed of two industrial robot manipulators equipped with force/torque sensors and pneumatic grippers, a vision system and a belt conveyor. The original industrial robot controllers have been replaced by a single PC with software running under a real-time variant of the Linux operative system. The new control environment allows advanced control schemes to be developed and tested for the single robots and for the dual-arm robotic cell, including force control and visual servoing tasks. An advanced user interface and a simulation environment have been developed, which permit fast, safe and reliable prototyping of planning and control algorithms.

An algorithm for the visual estimation of the pose of a moving object is presented in this paper. The algorithm exploits the prediction capability of the extended Kalman Filter to realize in real time a dynamic optimal selection of the object image features used for pose estimation. The robustness of the system with respect to the measurement noise and modelling errors is enhanced by using an adaptive scheme. Experimental case studies are presented to prove the effectiveness of the proposed approach.

The problem of tracking the position and orientation of a moving object using a stereo camera system is considered in this paper. A robust algorithm based on the extended Kalman filter is adopted, combined with an efficient selection technique of the object image features, based on Binary Space Partitioning tree geometric models. An experimental study is carried out using a vision system of two fixed cameras.

The problem of visual tracking of multiple objects is considered in this paper. Special emphasis is devoted to the case when two or more objects overlap with respect to the visual system causing occlusion. The algorithm is based on the Kalman filtering and binary space partition tree representations of the objects geometry. The real-time implementation of the algorithm is experimentally tested for the case of visual tracking of two objects using two cameras.

In this paper a computationally efficient algorithm for real time estimation of the position and orientation of moving objects from visual measurements of a system of fixed cameras is proposed. The algorithm is based on extended Kalman filtering of the measurements of the position of suitable feature points selected on the target objects. The effectiveness of the algorithm is improved by using a pre-selection method of the feature points which takes advantage of the Kalman filter prediction capability combined with a BSP tree modeling technique of the objects geometry. Computer simulations are presented to test the performance of the estimation process in the presence of noise, different types of lens geometric distortion, quantization and calibration errors.

The problem of real-time pose estimation of moving objects using a stereo video camera system is considered in this paper. A computationally efficient algorithm is proposed based on Kalman filtering of the position measurements of suitable feature points selected on the target objects. The efficiency of the algorithm is improved by adopting a new pre-selection technique of the feature points, based on binary space partition (BSP) trees, which takes advantage of the Kalman filter prediction capability. Computer simulations are presented.