Search

A new method is presented for planning collision-free robot motions in the presence of workspace obstacles, based on the use of artificial potential fields. In previous approaches, obstacles are modeled by means of repulsive potentials, which give rise to local minima in the total field, possibly causing the planning algorithm to jam up. Instead, the core of the proposed technique is to place vortex fields around obstacles, forcing the robot to turn around them and preventing the generation of local minima. In conjunction with this, an effective planning algorithm is introduced, which is valid both for mobile robots and manipulators. Simulation results show the efficiency of the method.

Presents a method for planning motions of three-link planar robots with a passive rotational third joint in the presence of gravity. These underactuated mechanisms can be fully linearized and input-output decoupled by means of a nonlinear dynamic state feedback, provided that a physical singularity is avoided. The linearizing output is the position of the center of percussion of the third link. Based on this, one can plan motions joining any initial and desired final state in finite time; in particular, transfers between inverted equilibria and swing-up maneuvers are easily obtained. Simulation results are reported for a 3R robot.

In this article, we present algorithms for robot path planning via navigation functions on fuzzy maps. In these kinds of bitmaps, a number is associated to each cell, characterizing the possibility that it belongs to an obstacle. Such a representation is useful when unknown environments are reconstructed from sensor measures. Two methods are devised; both are based on a proper modification of the wavefront expansion algorithm. The first is relatively simple, but produces paths that may graze the obstacles. Instead, the second method makes use of a properly defined skeleton of the map, that lies as far as possible from such areas. The performance of the two methods is illustrated by application to both simulated and experimental maps. The proposed method is also extended to the case of nonholonomic wheeled mobile robots.

A new method for solving redundancy in robot arms by locally optimizing a dynamic criterion is presented. Optimization is performed by means of a joint decomposition technique, resulting in a particularly efficient computational scheme which is feasible for real-time control. The technique is illustrated obtaining an alternate and simpler expression in the case of torque minimization. It is known that such a local scheme may exhibit unstable behavior, due to the build-up of high joint velocities that require in turn extremely large torques. Physical intuition is exploited to propose other dynamic criteria, whose minimization generates satisfactory torque profiles. Comparative simulations on a three-link planar arm show the effectiveness of the proposed method.

We address the control problem for the Butterfly, an interesting example of 2-dof underactuated mechanical system. This robot consists of a butterfly-shaped rotational link on whose rim a ball rolls freely. The control objective is to stabilize the robot at a certain unstable equilibrium. To this end, exploiting the existence of heteroclinic trajectories, we extend a previously proposed energy-based technique. Simulation results show the effectiveness of the presented method.

The authors investigate the problem of how to achieve a cyclic joint behavior in redundant robots performing cyclic tasks, motivated by the fact that most singularity-free local resolution methods produce nonrepeatable joint motions. A controllability analysis of the inverse kinematic system makes it possible to recover the well-known repeatability conditions of T. Shamir and Y. Yomdin (1988), and to further conclude that no null space velocity can be specified if a repeatable scheme is sought, unless it is chosen as a linear term in the end-effector velocity. The problem of achieving asymptotic cyclicity for a given inversion strategy has been solved via suitable kinematic controls, which guarantee convergence to cyclic joint trajectories along the desired end-effector path. Depending on the structure of the feedforward and feedback terms in the control law, a number of different schemes are proposed, yielding exact or asymptotic end-effector tracking. The stability proofs and the satisfactory simulation results confirm the advantage of using these simple control strategies. >

We present a resolution-adaptive strategy aimed at increasing the efficiency of probabilistic path planning. Our method represents a modification of the recently proposed RRT-connect planning algorithm. The objective is to reduce the computational load by performing less collision checks in problems involving many degrees of freedom and complex environments, where the cost of such operation is very high. To this end, we introduce an adaptive rule for selecting the step size of the expansion process and delay any collision checking to the solution path verification stage. Experimental results show the benefits of the proposed technique.

An analysis of underactuated manipulators from both the dynamic and the control points of view is presented. While the unactuated joints dynamic equation is recognized to be a nonholonomic constraint in the general case, necessary and sufficient conditions are given to identify special cases in which such a constraint is integrable. In contrast to most examples in the literature, the unactuated joints dynamics is an instance of a second-order nonholonomic constraint. it is shown that smooth feedback stabilization to a single equilibrium point is not possible, and a feedback scheme providing stabilization to a manifold of equilibrium positions is proposed. >

Robotic manipulation by rolling contacts is an appealing method for achieving dexterity with relatively simple hardware. While there exist techniques for planning motions of rigid bodies in rolling contact under nominal conditions, an inescapable challenge is the design of robust controllers of provable performance in the presence of model perturbations. As a preliminary step in this direction, we present in this paper an iterative robust planner of arbitrary accuracy for the plate-ball manipulation system subject to perturbations on the sphere radius. The basic tool is an exact geometric planner for the nominal system, whose repeated application guarantees the desired robustness property on the basis of the iterative steering paradigm. Simulation results under perturbed conditions show the effectiveness of the method.

This paper addresses the problem of generating whole-body motions for a humanoid robot that must execute a certain task in an environment containing obstacles. The assigned task trajectory is deformable, and the planner may exploit this feature for finding a solution. Our framework consists of two main components: a constrained motion planner and a deformation mechanism. The basic idea is that the constrained motion planner attempts to solve the problem for the original task. If this proves to be too difficult, the deformation mechanism modifies the task using appropriate heuristic functions. Then, the constrained motion planner is invoked again on the deformed task. If needed, this procedure is iterated. The proposed algorithm has been successfully implemented for the NAO humanoid in V-REP.

In this pilot study, a system that allows disabled persons to improve or recover their mobility and communication within the surrounding environment was implemented and validated. The system is based on a software controller that offers to the user a communication interface that is matched with the individual's residual motor abilities. Fourteen patients with severe motor disabilities due to progressive neurodegenerative disorders were trained to use the system prototype under a rehabilitation program. All users utilized regular assistive control options (e.g., microswitches or head trackers) while four patients learned to operate the system by means of a non-invasive EEG-based Brain-Computer Interface, based on the subjects' voluntary modulations of EEG sensorimotor rhythms recorded on the scalp.

We present a method for performing evasive motions with a humanoid robot. In the considered scenario, the robot is standing in a workspace, when a moving obstacle (e.g., a human, or another robot) enters its safety area and heads towards it; the humanoid must plan and execute in real-time a maneuver that avoids the collision. The proposed method goes through several conceptual steps. Once the entrance of the moving obstacle in the safety area is detected, its approach direction relative to the robot is determined. On the basis of this information, a suitable evasion maneuver represented by footsteps is generated. From these, an appropriate trajectory is computed for the Center of Mass of the humanoid. Finally, joint motion commands are generated so as to track such trajectory. All computations make use of closed-form expressions and are therefore suitable for real-time implementation. The proposed approach is validated via simulations and experiments on a NAO humanoid. The possibility of adapting the basic method so as to be used in a replanning framework is also investigated.

We consider the problem of stabilizing a 2R robot which moves in the horizontal plane by using a single actuator at the base. This system is representative of the class of underactuated mechanical systems that are not controllable in the first approximation. The presence of a drift term in the dynamic equations makes the application of most existing control techniques impossible. The proposed stabilization method makes use of three basic tools, namely (i) partial feedback linearization of the dynamic equations, (ii) computation of a nilpotent approximation of the system, and (iii) iterative application of an open-loop control designed on the nilpotent system. Although the procedure is presented for the 2R robot case, it provides guidelines for devising a method of general applicability.

Kinematic redundancy endows a robotic manipulator with the possibility of executing self-motions, that is changing its configuration without moving the end-effector This ability may be used to assume the most convenient posture for a given task, to avoid singularities or workspace obstacles, as well as to obtain closed joint motion on cyclic end-effector paths. We investigate the self-motion stabilization problem: the objective is to design a feedback scheme which drives the robot joint variables to a reference value keeping the end-effector still. Two main approaches are presented. The first is based on the idea of projecting the error term in the null space of the Jacobian matrix. Pure proportional feedback yields simple stability in general, and asymptotic stability in special cases. More interestingly, by using time-varying feedback joint convergence to the desired configuration is achieved. The second approach exploits the concept of reduction in the space of redundant degrees of freedom to design stabilization schemes which take advantage of the mechanical structure of the manipulator. Both approaches are illustrated by application to planar robot arms. >

We present a simple approach for planning the motion of nonholonomic robots among obstacles. Existing methods lead to open-loop solutions which are either obtained in two stages, approximating a previously built holonomic path, or computationally intensive, being based on configuration space discretization. Our nonholonomic planner employs a direct projection strategy to modify online the output of a holonomic incremental planner, and generates velocity control inputs that realize the desired motion in a least-squares sense. As a result, a feedback scheme is obtained which can use only local sensor information. The proposed approach is applied to unicycle kinematics, with artificial potential fields or vortex fields as local holonomic planners. >

The authors present an iterative learning controller that applies to nonholonomic mobile robots, as well as other systems that can be put in chained form. The learning algorithm exploits the fact that chained-form. The learning algorithm exploits the fact that chained-form systems are linear under piecewise-constant inputs. The proposed control scheme requires the execution of a small number of experiments to drive the system to the desired state in finite time, with nice convergence and robustness properties with respect to modeling inaccuracies as well as disturbances. To avoid the necessity of exactly reinitializing the system at each iteration, the basic method is modified so as to obtain a cyclic controller, by which the system is cyclically steered through an arbitrary sequence of states. As a case study, a carlike mobile robot is considered. Both simulation and experimental results are reported to show the performance of the method.

We present a local approach for planning the motion of a car-like robot navigating among obstacles, suitable for sensor-based implementation. The nonholonomic nature of the robot kinematics is explicitly taken into account. The strategy is to modify the output of a generic local holonomic planner, so as to provide commands that realize the desired motion in a least-squares sense. A feedback action tends to align the vehicle with the local force field. In order to avoid the motion stops away from the desired goal, various force fields are considered and compared by simulation.

We consider the feedback control problem for a wheeled mobile robot with the kinematics of a unicycle, a typical example of nonholonomic robotic system. It is shown that dynamic feedback linearization can be used to design a simple control law which is valid for trajectory tracking as well as point stabilization tasks. In particular, for both cases exponential convergence with linear transients in the cartesian space is obtained. Experimental results for a laboratory prototype prove the effectiveness of the proposed method.

The problem of sensor-based robot motion planning in unknown environments is addressed. The proposed solution approach prescribes the repeated sequence of two fundamental processes: perception and navigation. In the former, the robot collects data from its sensors, builds local maps and integrates them with the global maps so far reconstructed, using fuzzy logic operators. During the navigation process, a planner based on the A* algorithm proposes a path from the current position to the goal. The robot moves along this path until one of two termination conditions is verified namely (i) an unexpected obstructing obstacle is detected, or (ii) the robot is leaving the area in which reliable information has been gathered. Experimental results are presented for a Nomad 200 mobile robot.

Existing methods for nonholonomic motion planning can be applied to exactly nilpotentizable or flat systems. However, some nonholonomic systems do not fall into this class. For these, it is expected that the availability of globally defined nilpotent approximations will be important for the synthesis of efficient steering and stabilization strategies. In this paper, we consider as a case study the vehicle consisting of a car pulling two off-hooked trailers, and we show how to compute a nilpotent approximation that is valid in the neighborhood of both regular and singular points. Simulation results to test the accuracy of the proposed procedure are given.

This paper presents a control scheme for localizing and encircling a target using a multi-robot system. The task is achieved in a distributed way, in that each robot only uses local information gathered by on-board relative-position sensors assumed to be noisy, anisotropic, and unable to detect the identity of the measured object. Communication between the robots is provided by limited-range transceivers. Experimental results with stationary and moving targets support the theoretical analysis.

We present a method for trajectory planning and control of planar robots with a passive rotational last joint. These underactuated mechanical systems, which are subject to nonholonomic second-order constraints, are shown to be fully linearized and input-output decoupled by means of a nonlinear dynamic feedback. This objective is achieved in a unified framework, both in the presence or absence of gravity. The linearizing output is the position of the center of percussion of the last link. Based on this result, one can plan smooth trajectories joining in finite time any initial and desired final state of the robot; in particular, transfers between inverted equilibria and swing-up maneuvers under gravity are easily obtained. We also address the problem of avoiding the singularity induced by the dynamic linearization procedure through a careful choice of output trajectories. A byproduct of the proposed method is the straightforward design of exponentially stable tracking controllers for the generated trajectories. Simulation results are reported for a 3R robot moving in a horizontal and vertical plane. Possible extensions of the approach and its relationships with the differential flatness technique are briefly discussed.

We discuss some recent control techniques for underactuated manipulators - a special instance of mechanical systems having fewer input commands than degrees of freedom. This class includes robots with passive joints, elastic joints, or flexible links. Structural system properties are investigated showing that robots with passive joints are the most difficult to control. With reference to these, solutions are proposed for the typical problems of trajectory planning and tracking, and of set-point regulation. The relevance of nonlinear control techniques such as dynamic feedback linearization and iterative state steering is clarified through illustrative examples.

Algorithmic methods for solving robot redundancy are considered, providing new strategies for optimizing general kinematic criteria at the acceleration level. Acceleration resolution offers two potential advantages: it allows to take into account dynamic performance and can be directly used with control techniques for accurate tracking. Local first and second-order optimal resolution schemes are described in a convenient unified framework and classified according to their iterative or exact nature. Instances where an equivalence exists between velocity and acceleration methods are pointed out. Existing second-order methods do not apply to the case of mixed criteria, namely objective functions depending both on joint configuration and velocity. A discrete-time formulation is used to design algorithms solving this problem. Simulations with a planar redundant arm show the benefits achieved with the new optimization scheme.

Robotics has revolutionized several industries across the globe through technologies never seen before. However, society still expects more than what today’s robots are really capable of. To deal with such expectations, innovative algorithms must be translated into viable solutions so that robots can continue to improve labor in workplaces and ordinary people’s lives. Among the most important are motion generation algorithms, whose advancements can solve some of the biggest barriers towards more flexible and safe human-robot interactions and operations in dynamic environments. The developments in optimization algorithms and computer processor technology have lednonlinear model predictive control(NMPC) to gain popularity in robotics as a motion generation technique. The main reason for that is its ability to minimize a cost function while respecting a set of constraints that typically represent the system’s physical and operational limitations. Ultimately, these are the elements that grant NMPC an improved performance compared to classic approaches. Despite being a promising technique, the non-negligible computational burden associated with the online solution of the underlying optimal control problems has decisively limited its roll-out to robotic systems subject to short sampling times and resource-constrained hardware. This thesis proposes multiple tailored algorithms for real-time motion generation in robotic systems based on high-performance implementations of NMPC. The primary computational bottlenecks concern the numerical simulation of the continuous-time nonlinear dynamic models and the online solution of the stemming large yet well-structured nonlinear program. The thesis shows that it is possible to achieve significant speed-ups in solution times while preserving nonlinearities through efficient software implementations that cover standard building blocks from nonlinear programming, tailored quadratic programming solvers, and fast approximate schemes for NMPC. Additionally, a discussion is provided in terms of dynamic modeling. It encompasses the design decisions required to create a model that ex- poses the system limitations so that high-quality motions can be attained due to a more accurate representation. Among the ever-growing plethora of robotic systems, the thesis focuses on the motion generation of a double inverted pendulum and a quadrotor. In particular, two numerical simulations addressing human-robot interaction and operation in dynamic environments and two real-world applications dealing with position control demonstrate a significant improvement in control performance, with solution times in the range of micro- and milliseconds.

One of the key problems in planning and control of redundant robots is the fast gen- eration of controls when multiple tasks and constraints need to be satisfied. In the literature, this problem is classically solved by multi-task prioritized approaches, where the priority of each task is determined by a weight function, or with trajectory optimiza- tion techniques. In this thesis we propose a framework that, through the combination of machine learning and control theory approach, can be efficiently applied both for multi task prioritization and trajectory optimization to automatically find optimal behaviour. First we learn the temporal profiles of the task priorities, represented as parametrized weight functions: we automatically determine their parameters through a constrained stochastic optimization procedure. Then we extend the proposed method for trajectory optimization scenario where we learn the task trajectories for whole-body balancing tasks. In both cases we ensure that the optimized movements are safe and never violate any of the robot and problem constraints. For this purpose we compare three constrained variants of CMA-ES on several benchmarks, among which two are new robotics bench- marks of our design using the KUKA LWR. We retain (1+1)-CMA-ES with covariance constrained adaptation [30] as the best candidate to solve our problems. In order to tackle the limitations of the algorithm described in [30], in this thesis we propose an extension of (1+1)-CMA-ES with Constrained Covariance Adaptation (CCA) that ad- dresses all the issue that affects the learning module of our framework. We show the effectiveness of the proposed framework for the task priority learning on a simulated 7 DOF Kuka LWR and both a simulated and a real Kinova Jaco arm. We compare the performance of our approach to a state-of-the-art method based on soft task prioriti- zation, where the task weights are typically hand-tuned. Then we apply our method on two whole-body experiments with the iCub humanoid robot to show its scalability property. Finally we test our learning framework to the prioritized whole-body torque controller of iCub, to optimize the robot’s trajectory for standing up from a chair.; One of the key problems in planning and control of redundant robots is the fast gen- eration of controls when multiple tasks and constraints need to be satisfied. In the literature, this problem is classically solved by multi-task prioritized approaches, where the priority of each task is determined by a weight function, or with trajectory optimiza- tion techniques. In this thesis we propose a framework that, through the combination of machine learning and control theory approach, can be efficiently applied both for multi task prioritization and trajectory optimization to automatically find optimal behaviour. First we learn the temporal profiles of the task priorities, represented as parametrized weight functions: we automatically determine their parameters through a constrained stochastic optimization procedure. Then we extend the proposed method for trajectory optimization scenario where we learn the task trajectories for whole-body balancing tasks. In both cases we ensure that the optimized movements are safe and never violate any of the robot and problem constraints. For this purpose we compare three constrained variants of CMA-ES on several benchmarks, among which two are new robotics bench- marks of our design using the KUKA LWR. We retain (1+1)-CMA-ES with covariance constrained adaptation [30] as the best candidate to solve our problems. In order to tackle the limitations of the algorithm described in [30], in this thesis we propose an extension of (1+1)-CMA-ES with Constrained Covariance Adaptation (CCA) that ad- dresses all the issue that affects the learning module of our framework. We show the effectiveness of the proposed framework for the task priority learning on a simulated 7 DOF Kuka LWR and both a simulated and a real Kinova Jaco arm. We compare the performance of our approach to a state-of-the-art method based on soft task prioriti- zation, where the task weights are typically hand-tuned. Then we apply our method on two whole-body experiments with the iCub humanoid robot to show its scalability property. Finally we test our learning framework to the prioritized whole-body torque controller of iCub, to optimize the robot’s trajectory for standing up from a chair.

In this thesis, we focus on the use of robot vision sensors to solve the path following problem, which in recent years, has been a popular target for engaged researchers around the world. Initially, we shall present the working assumptions that we used, along with the relevant kinematic and sensor models of interest, and with the main characteristics of visual servoing. Then, we shall attempt to survey the research carried out in the field of vision-based path following. Following this, we shall present two schemes that we designed for reaching and following paths. These respectively involve nonholonomic robots and legged robots. Both schemes require only some visible path features, along with a coarse camera model, and, under certain conditions, may guarantee convergence even when the initial error is large. The first control scheme is designed to enable nonholonomic mobile robots with a fi pinhole camera to reach and follow a continuous path on the ground. Two visual servoing controllers (position-based and image-based) have been designed. For both controllers, a Lyapunov-based stability analysis has been carried out. The performance of the two controllers is validated and compared by simulations and experiments on a car-like robot. The second control scheme is more application-oriented than the first and has been used in ASPICE, an assistive robotics project, to enable the autonomous navigation of a legged robot equipped with an actuated pinhole camera. The robot uses a position-based visual servoing controller to follow artificial paths on the ground, and travel to some required destinations. Apart from being a useful testing-ground for this path following scheme, the ASPICE project has also permitted the development of many other aspects in the field of assistive robotics. These shall be outlined at the end of the thesis.; Resume en anglais uniquement.

In this thesis, we focus on the use of robot vision sensors to solve the path following problem, which in recent years, has been a popular target for engaged researchers around the world. Initially, we shall present the working assumptions that we used, along with the relevant kinematic and sensor models of interest, and with the main characteristics of visual servoing. Then, we shall attempt to survey the research carried out in the field of vision-based path following. Following this, we shall present two schemes that we designed for reaching and following paths. These respectively involve nonholonomic robots and legged robots. Both schemes require only some visible path features, along with a coarse camera model, and, under certain conditions, may guarantee convergence even when the initial error is large. The first control scheme is designed to enable nonholonomic mobile robots with a fi pinhole camera to reach and follow a continuous path on the ground. Two visual servoing controllers (position-based and image-based) have been designed. For both controllers, a Lyapunov-based stability analysis has been carried out. The performance of the two controllers is validated and compared by simulations and experiments on a car-like robot. The second control scheme is more application-oriented than the first and has been used in ASPICE, an assistive robotics project, to enable the autonomous navigation of a legged robot equipped with an actuated pinhole camera. The robot uses a position-based visual servoing controller to follow artificial paths on the ground, and travel to some required destinations. Apart from being a useful testing-ground for this path following scheme, the ASPICE project has also permitted the development of many other aspects in the field of assistive robotics. These shall be outlined at the end of the thesis.; Resume en anglais uniquement.