Robotica

A generally accepted helicopter model used for control includes the effect of Small Body Forces (SBF) which couple the vehicle's rotational subsystem inputs to its translational dynamics. SBF result from tail rotor thrust and lateral forces due to main rotor flapping. It is well-known that SBF lead to a theoretically challenging stabilization problem for the tracking error dynamics. Hence, much of the existing work has neglected SBF in order to simplify control design. We design a controller that directly compensates the influence of the tail rotor component of the SBF. The design is validated in simulation and flight tests.

Higher-order derivatives of kinematic mappings give insight into the motion characteristics of complex mechanisms. Screw theory and its associated Lie group theory have been used to find these derivatives of loop closure equations up to an arbitrary order. In this paper, this is extended to the higher-order derivatives of the solution to these loop closure equations to provide an approximation of the finite motion of serial and parallel mechanisms. This recursive algorithm, consisting solely of matrix operations, relies on a simplified representation of the higher-order derivatives of open chains. The method is applied to a serial, a multi-DOF parallel, and an overconstrained mechanism. In all cases, adequate approximation is obtained over a large portion of the workspace.

We provide necessary and sufficient conditions for all projective transformations of the projectivized dual quaternion model of rigid body displacements that are induced by coordinate changes in moving and/or fixed frame. These transformations fix the quadrics of a pencil and preserve the two families of rulings of an exceptional three-dimensional quadric. Moreover, we fully characterize the constraint varieties of dyads with revolute and prismatic joints in the dual quaternion model. The constraint variety of a dyad with two revolute joints is a regular ruled quadric in a three-space that contains a “null quadrilateral.” If a revolute joint is replaced by a prismatic joint, this quadrilateral collapses into a pair of conjugate complex null lines and a real line but these properties are not sufficient to characterize such dyads. We provide a complete characterization by introducing a new invariant, the “Study fibre projectivity,” and we present examples that demonstrate its potential to explain hitherto not sufficiently well-understood phenomena.

A novel reconfigurable tri-prism mobile robot with eight modes is proposed. The robot is composed of two feet connected by three U-R-U (universal-revolute-universal) limbs. The robot incorporates the kinematic properties of sphere robots, squirming robots, tracked robots, wheeled robots and biped robots. In addition, the somersaulting and turning modes are also explored. After the description of the robot, the DOF (degree-of-freedom) is calculated based on screw theory. The 3D model and simulations indicate that the robot can cross several typical obstacles and can also be folded via two approaches. Finally, the prototype experiments are presented to verify the feasibility of the proposed mobile robot in different motion mode.

A three-dimensional (3D) model of a novel 5-DoF type parallel manipulator with a couple-constrained wrench is constructed and its couple-constrained wrench is analyzed. First, the formulas are derived for solving the displacement, velocity, acceleration of the moving platform and moving links, and a workspace is constructed. Second, the formulas are derived for solving the inertial wrenches of the moving links. Third, a dynamics equation is established by considering the inertial wrenches and friction, and the formulas are derived for solving the dynamically active forces and the dynamically couple-constrained wrench. Finally, a numerical example is given to demonstrate the analytic solution of the kinematics and the dynamics, and the analytical solutions are verified by utilizing a simulation mechanism.

A robot should be able to estimate an accurate and dense 3D model of its environment (a map), along with its pose relative to it, all of it in real time, in order to be able to navigate autonomously without collisions.

As the robot moves from its starting position and the estimated map grows, the computational and memory footprint of a dense 3D map increases and might exceed the robot capabilities in a short time. However, a global map is still needed to maintain its consistency and plan for distant goals, possibly out of the robot field of view.

In this work, we address such problem by proposing a real-time stereo mapping pipeline, feasible for standard CPUs, which is locally dense and globally sparse and accurate. Our algorithm is based on a graph relating poses and salient visual points, in order to maintain a long-term accuracy with a small cost. Within such framework, we propose an efficient dense fusion of several stereo depths in the locality of the current robot pose.

We evaluate the performance and the accuracy of our algorithm in the public datasets of Tsukuba and KITTI, and demonstrate that it outperforms single-view stereo depth. We release the code as open-source, in order to facilitate the system use and comparisons.

Safe and accurate navigation for autonomous trajectory tracking of quadrotors using monocular vision is addressed in this paper. A second order Sliding Mode (2-SM) control algorithm is used to track desired trajectories, providing robustness against model uncertainties and external perturbations. The time-scale separation of the translational and rotational dynamics allows to design position controllers by giving a desired reference in roll and pitch angles, which is suitable for practical validation in quad-rotors equipped with an internal attitude controller. A Lyapunov based analysis proved the closed-loop stability of the system despite the presence of unknown external perturbations. Monocular vision fused with inertial measurements are used to estimate the vehicle's pose with respect to unstructured scenes. In addition, the distance to potential collisions is detected and computed using the sparse depth map coming also from the vision algorithm. The proposed strategy is successfully tested in real-time experiments, using a low-cost commercial quadrotor.

This paper reports the design, analysis, and control of a miniature vibration-driven planar locomotion robot called Shell. A vibration-driven system is able to achieve locomotion based on internal oscillations and anisotropic friction forces. In this robot design, two parallel oscillators are employed to provide propelling forces, and a blade-like support is designed to generate anisotropic frictional contact with the ground. If the two parallel oscillators are of different frequencies and amplitudes, two-dimensional locomotion of the robot can be achieved. To predict the robot's planar locomotion, a dynamic model is developed. Controlling the robot's locomotion and especially, the locomotion modes can be achieved by adjusting the vibration frequencies of the two internal oscillators. Experimental results show that Shell can be controlled to move rectilinearly and along circles with certain curvatures. In addition, by combining these basic trajectories, Shell can move freely on a horizontal plane.

This work proposes a strategy for position control and obstacle avoidance in a quadcopter based on constrained generalized predictive control and geometric attitude control. The approach allows real-time trajectory tracking using optimal control actions and avoids collisions with static obstacles whose position is known. An experimental validation of the proposed controller is presented.

In this paper, the kinematic accuracy problem caused by geometric errors of a 2(3HUS+S) parallel manipulator is described. The kinematic equation of the manipulator is obtained by establishing a D–H (Denavit–Hartenberg) coordinate system. A D–H transformation matrix is used as the error-modeling tool, and the kinematic error model of the manipulator integrating manufacturing and assembly errors is established based on the perturbation theory. The iterative Levenberg–Marquardt algorithm is used to identify the geometric errors in the error model. According to the experimentally measured attitudes, the kinematic calibration process is simulated using MATLAB software. The simulation and experiment results show that the attitude errors of the moving platforms after calibration are reduced compared with before the calibration, and the kinematic accuracy of the manipulator is significantly improved. The correctness and effectiveness of the error model and the kinematic calibration method of the 2(3HUS+S) parallel manipulator for simulation of hip joint motion are verified.

This paper presents a simple and highly visual approach for the type synthesis of a family of overconstrained parallel mechanisms that have one translational and two rotational movement capabilities. It considers, especially, mechanisms offering the accuracy and dynamic response needed for machining applications. This family features a spatial limb plus a member of a class of planar symmetrical linkages, the latter connected by a revolute joint either to the machine frame at its base link or to the platform at its output link. Criteria for selecting suitable structures from among numerous candidates are proposed by considering the realistic practical requirements for reconfigurability, movement capability, rational component design and so on. It concludes that a few can simultaneously fulfil the proposed criteria, even though a variety of structures have been presented in the literature. Exploitation of the proposed structures and evaluation criteria then leads to a novel five degrees of freedom hybrid module named TriMule. A significant potential advantage of the TriMule over the Tricept arises because all the joints connecting the base link and the machine frame can be integrated into one single, compact part, leading to a lightweight, cost effective and flexible design particularly suitable for configuring various robotized manufacturing cells.

This paper proposes a detailed investigation on the new neural-based feed-forward PID direct force control (FNN-PID-DF) approach applied to a highly nonlinear 2-axes pneumatic artificial muscle (PAM) manipulator in order to ameliorate its force output performance. Founded on the novel inverse neural NARX model dynamically identified to learn well all nonlinear characteristics of the contact force dynamics of the 2-axes PAM-based manipulator, the novel proposed neural FNN-PID-DF force controller is innovatively implemented in order to directly force control the 2-axes PAM robot system used as a rehabilitation device subjected to internal systematic interactions and external contact force deviations. The performance of the experimental tests has proven the advantages and merits of the new force control method compared to the classical PID force control method. The new neural FNN-PID-DF force controller guides the wrist/hand of subject/patient to successfully generate the predefined desired force values.

The ability to reliably estimate free space is an essential requirement for efficient and safe robot navigation. This paper presents a novel system, built upon a stochastic framework, which estimates free space quickly from stereo data, using self-supervised learning. The system relies on geometric data in the close range of the robot to train a second-stage appearance-based classifier for long range areas in a scene. Experiments are conducted on board an unmanned ground vehicle, and the results demonstrate the advantages of the proposed technique over other self-supervised systems.

This paper suggests an integrated navigation system for an unmanned ground vehicle operating in an unknown cluttered environment. The navigator supports time-critical mobility making it possible for a mobile robot to reach a target from the first attempt without the need for a dedicated exploration and mapping stage. The robot only uses necessary and sufficient egocentric local sensory data collected on its way to the target. The construction of the navigation structure revolves around key properties of the harmonic potential field approach to motion planning. The robot's trajectory is well-behaved and direct-to-the-goal. It contains only the minimum number of detours necessary to accommodate the sensory data and maintain the robot in a safe, goal-oriented state. The navigation structure is developed and its theoretical basis is explained. Extensive experimental validation of its properties and performance is carried-out using the X80 robotic platform.

This paper reviews the state-of-the-art in robotic tails intended for inertial adjustment applications on-board mobile robots. Inspired by biological tails observed in nature, robotic tails provide a separate means to enhance stabilization, and maneuverability from the mobile robot's main form of locomotion, such as legs or wheels. Research over the past decade has primarily focused on implementing single-body rigid pendulum-like tail mechanisms to demonstrate inertial adjustment capabilities on-board walking, jumping and wheeled mobile robots. Recently, there have been increased efforts aimed at leveraging the benefits of both articulated and continuum tail mechanism designs to enhance inertial adjustment capabilities and further emulate the structure and functionalities of tail usage found in nature. This paper discusses relevant research in design, modeling, analysis and implementation of robotic tails onto mobile robots, and highlight how this work is being used to build robotic systems with enhanced performance capabilities. The goal of this article is to outline progress and identify key challenges that lay ahead.

We present singularity-free workspace optimization of a class of two-degree-of-freedom (2-DoF) parallel wrists with large rotation range capability. The wrists in consideration are kinematically equivalent to two families of 2-DoF homokinetic couplings. The first family comprises fully parallel wrists with N (N ≥ 3) double-universal ( $\mathcal{UU}$ ) legs. The second family comprises spherical N- $\mathcal{UU}$ parallel wrists with interconnecting revolute ( $\mathcal{R}$ ) joints. Both families belong to the more general class of zero-torsion parallel manipulators, and are, therefore, collectively referred to as zero-torsion wrists (ZTWs). We carry out a unified singularity-free workspace optimization by utilizing geometric properties of zero-torsion motion manifolds. Our work may serve as a conceptual guide to the design of ZTWs for large tilt-angle applications.

This paper presents visual and 3D structure inspection for steel structures and bridges using a developed climbing robot. The robot can move freely on a steel surface, carry sensors, collect data and then send to the ground station in real-time for monitoring as well as further processing. Steel surface image stitching and 3D map building are conducted to provide a current condition of the structure. Also, a computer vision-based method is implemented to detect surface defects on stitched images. The effectiveness of the climbing robot's inspection is tested in multiple circumstances to ensure strong steel adhesion and successful data collection. The detection method was also successfully evaluated on various test images, where steel cracks could be automatically identified, without the requirement of some heuristic reasoning.

Due to physically demanding tasks in construction, workers are exposed to significant safety and health risks. Measuring and evaluating body kinematics while performing tasks helps to identify the fundamental causes of excessive physical demands, enabling practitioners to implement appropriate interventions to reduce them. Recently, non-invasive or minimally invasive motion capture approaches such as vision-based motion capture systems and angular measurement sensors have emerged, which can be used for in-field kinematics measurements, minimally interfering with on-going work. Given that these approaches have pros and cons for kinematic measurement due to adopted sensors and algorithms, an in-depth understanding of the performance of each approach will support better decisions for their adoption in construction. With this background, the authors evaluate the performance of vision-based (RGB-D sensor-, stereovision camera-, and multiple camera-based) and an angular measurement sensor-based (i.e., an optical encoder) approach to measure body angles through experimental testing. Specifically, measured body angles from these approaches were compared with the ones obtained from a marker-based motion capture system that has less than 0.1 mm of errors. The results showed that vision-based approaches have about 5–10 degrees of error in body angles, while an angular measurement sensor-based approach measured body angles with about 3 degrees of error during diverse tasks. The results indicate that, in general, these approaches can be applicable for diverse ergonomic methods to identify potential safety and health risks, such as rough postural assessment, time and motion study or trajectory analysis where some errors in motion data would not significantly sacrifice their reliability. Combined with relatively accurate angular measurement sensors, vision-based motion capture approaches also have great potential to enable us to perform in-depth physical demand analysis such as biomechanical analysis that requires full-body motion data, even though further improvement of accuracy is necessary. Additionally, understanding of body kinematics of workers would enable ergonomic mechanical design for automated machines and assistive robots that helps to reduce physical demands while supporting workers' capabilities.

In this paper, a new model for joint dynamic friction of industrial robot manipulators is presented. In particular, the effects of the temperature in the joints are considered. A polynomial-based model is proposed and the parameter estimation is performed without the need of a joint temperature sensor. The use of an observer is then proposed to compensate for the uncertainty in the initial estimation of the temperature value. A large experimental campaign show that the model, in spite of the simplifying assumptions made, is effective in estimating the joint temperature and therefore the friction torque during the robot operations, even for values of velocities that have not been previously employed.