Vacuum-based end effectors are widely used in industry and are often preferred over parallel-jaw and multifinger grippers due to their ability to lift objects with a single point of contact. Suction grasp planners often target planar surfaces on point clouds near the estimated centroid of an object. In this paper, we propose a compliant suction contact model that computes the quality of the seal between the suction cup and local target surface and a measure of the ability of the suction grasp to resist an external gravity wrench. To characterize grasps, we estimate robustness to perturbations in end-effector and object pose, material properties, and external wrenches. We analyze grasps across 1,500 3D object models to generate Dex-Net 3.0, a dataset of 2.8 million point clouds, suction grasps, and grasp robustness labels. We use Dex-Net 3.0 to train a Grasp Quality Convolutional Neural Network (GQ-CNN) to classify robust suction targets in point clouds containing a single object. We evaluate the resulting system in 350 physical trials on an ABB YuMi fitted with a pneumatic suction gripper. When evaluated on novel objects that we categorize as Basic (prismatic or cylindrical), Typical (more complex geometry), and Adversarial (with few available suction-grasp points) Dex-Net 3.0 achieves success rates of 98%, 82%, and 58% respectively, improving to 81% in the latter case when the training set includes only adversarial objects. Code, datasets, and supplemental material can be found at http://berkeleyautomation.github.io/dex-net.

Dex-Net 3.0: Computing Robust Vacuum Suction Grasp Targets in Point Clouds Using a New Analytic Model and Deep Learning

In order to overcome the problem of power generation in distributed energy, microgrid(MG) emerges as an alternative scheme. Compared with the ac microgrids, the dc microgrids have the advantages of high system efficiency, good power quality, low cost, and simple control. However, due to the complexity of the distributed generation system, the conventional droop control shows the drawbacks of low current sharing accuracy. Therefore, the improved primary control methods to enhance current sharing accuracy are systematically reviewed, such as particle swarm optimization programming, probabilistic algorithm and voltage correction factor scheme. However, it is difficult to achieve stable and coordinated operation of the dc microgrids by relying on the primary control. Hence, the various secondary control approaches, such as dynamic current sharing scheme, muti-agent system (MAS) control and virtual voltage control methods have been summarized for voltage regulation. Furthermore, the energy management system (EMS), modular-based energy router (MBER) and other coordinated control methods are reviewed to achieve power management. Besides, various control methods to compensate the effect of communication delay are summarized. Moreover, linear matrix inequality (LMI), Lyapunov-Krasovskii functional stability and Takagi-Sugeno model prediction scheme can be adopted to eliminate the influence of communication delay. In addition, due to the constant power loads (CPL) exhibit negative impedance characteristics, which may result in the output oscillation of filter. Thus, various control approaches have been reviewed to match the impedance, such as the nonlinear disturbance observer (NDO) feedforward compensation method, linear programming algorithm, hybrid potential theory and linear system analysis of polyhedral uncertainty. The merits and drawbacks of those control strategies are compared in this paper. Finally, the future research trends of hierarchical control and stability in dc microgrids and dc microgrid clusters are also presented.

/pdf/review-of-power-sharing-voltage-restoration-and-1kru0dubd1.pdf

Review of Power Sharing, Voltage Restoration and Stabilization Techniques in Hierarchical Controlled DC Microgrids

Vacuum-based end effectors are widely used in industry and are often preferred over parallel-jaw and multifinger grippers due to their ability to lift objects with a single point of contact. Suction grasp planners often target planar surfaces on point clouds near the estimated centroid of an object. In this paper, we propose a compliant suction contact model that computes the quality of the seal between the suction cup and local target surface and a measure of the ability of the suction grasp to resist an external gravity wrench. To characterize grasps, we estimate robustness to perturbations in end-effector and object pose, material properties, and external wrenches. We analyze grasps across 1,500 3D object models to generate Dex-Net 3.0, a dataset of 2.8 million point clouds, suction grasps, and grasp robustness labels. We use Dex-Net 3.0 to train a Grasp Quality Convolutional Neural Network (GQ-CNN) to classify robust suction targets in point clouds containing a single object. We evaluate the resulting system in 350 physical trials on an ABB YuMi fitted with a pneumatic suction gripper. When evaluated on novel objects that we categorize as Basic (prismatic or cylindrical), Typical (more complex geometry), and Adversarial (with few available suction-grasp points) Dex-Net 3.0 achieves success rates of 98$\%$, 82$\%$, and 58$\%$ respectively, improving to 81$\%$ in the latter case when the training set includes only adversarial objects. Code, datasets, and supplemental material can be found at this http URL .

Dex-Net 3.0: Computing Robust Robot Suction Grasp Targets in Point Clouds using a New Analytic Model and Deep Learning

Multi-converter electronic systems are becoming widely used in many industrial applications; therefore, the stability of the whole system is a big concern to the real-world power supplies applications. A multi-converter system comprised of cascaded converters has a basic configuration that consists of two or more converters in series connection, where the first is a source converter that maintains a regulated dc voltage on the intermediate bus while remaining are load converters that convert the intermediate bus voltage to the tightly regulated outputs for the next system stage or load. Instability in cascaded systems may occur due to the constant power load (CPL), which is a behavior of the tightly regulated converters. CPLs exhibit incremental negative resistance behavior causing a high risk of instability in interconnected converters. In addition, there are other problems apart from the CPL, e.g., non-linearities due to the inductive element and uncertainties due to the imprecision of a mathematical model of dc-dc converters. Aiming to effectively mitigate oscillations effects in the output of source converter loaded with a CPL, in this paper, an interval robust controller, by linear programming based on Kharitonov rectangle, is proposed to regulate the output of source converter. Several tests were developed by using an experimental plant and simulation models when the multi-converter buck-buck system is subjected to a variation of power reference. Both simulation and experimental results show the effectiveness of the proposed controller. Furthermore, the performance indices computed from the experimental data show that the proposed controller outperforms a classical control technique.

/pdf/interval-robust-controller-to-minimize-oscillations-effects-43nr2st8iy.pdf

Interval Robust Controller to Minimize Oscillations Effects Caused by Constant Power Load in a DC Multi-Converter Buck-Buck System

In recent years, dc microgrid (MG) is increasing rapidly in electric power grids and other isolated systems, integrating more efficiency and suite better some of the renewable energy sources, storage units, and dc loads. However, dc MG stability analysis becomes a challenge when constant power loads (CPLs) are applied to dc bus, which introduces destabilizing effects in the system due to its negative impedance characteristics. This paper presents a novel robust controller, based on linear programming based on the Chebyshev theorem as a robust control technique considering the Kharitonov’s theorem that ensures the minimization of the total deviation from the desired performance in a closed-loop system, specified by a family of characteristic polynomials. The purpose of the proposed controller is to tightly regulate the dc bus voltage, ensuring MG stability due to the effects of power variation on CPLs. The simulation and experimental tests are performed by using a MATLAB/Simulink simulator and a developed prototype of the DC MG system, respectively, to ratify the robustness and effectiveness of the proposed method of robust controller design.

/pdf/novel-robust-methodology-for-controller-design-aiming-to-208rl610x5.pdf

Novel Robust Methodology for Controller Design Aiming to Ensure DC Microgrid Stability Under CPL Power Variation

A fractional-order model predictive control with extended prediction self-adaptive control (FOMPC-EPSAC) strategy is proposed for the AR.Drone quadrotor system. The objective is to achieve an optimal trajectory tracking control for an AR.Drone quadrotor by using a fractional order integral cost function in the conventional MPC-EPSAC algorithm. In addition, a particle swarm optimization (PSO) algorithm is applied to find the optimal weighting matrices, which depend on the terms ($\alpha $, $\beta $) of the fractional order cost function. Some simulation results show the superiority of FOMPC-EPSAC over conventional MPC-EPSAC with respect to trajectory tracking and robustness under wind disturbances.

A fractional order predictive control for trajectory tracking of the AR.Drone quadrotor

Las redes moviles ad hoc (MANET) se caracterizan por la movilidad de los dispositivos participantes y la comunicacion salto a salto entre ellos. La creacion de estas redes ocurre de una manera no planificada y respondiendo a la distribucion de los dispositivos en un momento determinado. Esta investigacion se enfoca en el desarrollo de una aplicacion movil Android capaz de crear un entorno ad hoc para el intercambio de datos, que puede proporcionar al usuario, y a sus aplicaciones, beneficios en los entornos constituidos por dispositivos que forman redes moviles ad hoc. Estas redes son utilizadas en situaciones en las que no existe posibilidad de desplegar una infraestructura de comunicacion, para lo cual se analizaron las tecnicas de encaminamiento de redes sin infraestructura y el comportamiento de los nodos, ademas del trafico de la red. Los resultados muestran que la solucion propuesta por medio de broadcast, debido a la tecnica de inundacion de paquetes, permite solventar comunicaciones a una cantidad no mayor a 50 nodos activos, por lo que se espera que esta investigacion permita a los usuarios encontrar los servicios que satisfagan sus necesidades y, cuando no haya esta posibilidad, combinar otras funcionalidades disponibles en forma de servicios.

/pdf/implementacion-de-una-aplicacion-movil-en-un-entorno-ad-hoc-3vvxtmqurv.pdf

Implementación de una aplicación móvil en un entorno AD HOC para el intercambio de información

Proportional-Integral-Derivative controller has been applied in a wide range of industrial applications since its simpleness and non-hardware limitations. However, only depending on the tuning technique, its efficiency can be highly improved in order to reach process requirements. This paper presents a cascade scheme control system for the temperature in a steam distillation for essential oils extraction process. Temperature is one of the most crucial parameters that can define the quality of the essential oil in a distillation process. Although, as many real systems, there are some disturbing signals that need to be treated in a controller design due to robustness, fast response and closed - loop stability approach. The tuning technique, for both, inner and outer controller, is based on Internal Model Control (IMC) rules, where pressure has been considered as the disturbing signal. Moreover, a contrast is made between the performance of the IMC rules and other tuning methods: Ziegler Nichols and Skogestad - IMC. The simulation results demonstrate that the IMC controllers under the cascade scheme gives a better performance compared to the other methods when set point tracking and disturbance rejection are measured.

Douglas Antonio Plaza Guingla

Papers

A fractional order predictive control for trajectory tracking of the AR.Drone quadrotor

Implementación de una aplicación móvil en un entorno AD HOC para el intercambio de información

Improvements on Disturbance Rejection in Steam Distillation through IMC-based PID Design