The acquisition and self-improvement of novel motor skills is among the most important problems in robotics. Motor primitives offer one of the most promising frameworks for the application of machine learning techniques in this context. Employing an improved form of the dynamic systems motor primitives originally introduced by Ijspeert et al. [2], we show how both discrete and rhythmic tasks can be learned using a concerted approach of both imitation and reinforcement learning. For doing so, we present both learning algorithms and representations targeted for the practical application in robotics. Furthermore, we show that it is possible to include a start-up phase in rhythmic primitives. We show that two new motor skills, i.e., Ball-in-a-Cup and Ball-Paddling, can be learned on a real Barrett WAM robot arm at a pace similar to human learning while achieving a significantly more reliable final performance.

Learning Motor Primitives for Robotics

Robotic grinding is considered as an alternative towards the efficient and intelligent machining of complex components by virtue of its flexibility, intelligence and cost efficiency, particularly in comparison with the current mainstream manufacturing modes. The advances in robotic grinding during the past one to two decades present two extremes: one aims to solve the problem of precision machining of small-scale complex surfaces, the other emphasizes on the efficient machining of large-scale complex structures. To achieve efficient and intelligent grinding of these two different types of complex components, researchers have attempted to conquer key technologies and develop relevant machining system. The aim of this paper is to present a systematic, critical, and comprehensively review of all aspects of robotic grinding of complex components, especially focusing on three research objectives. For the first research objective, the problems and challenges arising out of robotic grinding of complex components are identified from three aspects of accuracy control, compliance control and cooperative control, and their impact on the machined workpiece geometrical accuracy, surface integrity and machining efficiency are also identified. For the second aim of this review, the relevant research work in the field of robotic grinding till the date are organized, and the various strategies and alternative solutions to overcome the challenges are provided. The research perspectives are concentrated primarily on the high-precision online measurement, grinding allowance control, constant contact force control, and surface integrity from robotic grinding, thereby potentially constructing the integration of “measurement – manipulation – machining” for the robotic grinding system. For the third objective, typical applications of this research work to implement successful robotic grinding of turbine blades and large-scale complex structures are discussed. Some research interests for future work to promote robotic grinding of complex components towards more intelligent and efficient in practical applications are also suggested.

Robotic grinding of complex components: A step towards efficient and intelligent machining – challenges, solutions, and applications

With the continuous improvement of automation, industrial robots have become an indispensable part of automated production lines. They widely used in a number of industrial production activities, such as spraying, welding, handling, etc., and have a great role in these sectors. Recently, the robotic technology is developing towards high precision, high intelligence. Robot calibration technology has a great significance to improve the accuracy of robot. However, it has much work to be done in the identification of robot parameters. The parameter identification work of existing serial and parallel robots is introduced. On the one hand, it summarizes the methods for parameter calibration and discusses their advantages and disadvantages. On the other hand, the application of parameter identification is introduced. This overview has a great reference value for robot manufacturers to choose proper identification method, points further research areas for researchers. Finally, this paper analyzes the existing problems in robot calibration, which may be worth researching in the future.

An overview of calibration technology of industrial robots

Industry 4.0 is taking human-robot collaboration at the center of the production environment. Collaborative robots enhance productivity and flexibility while reducing human’s fatigue and the risk of injuries, exploiting advanced control methodologies. However, there is a lack of real-time model-based controllers accounting for the complex human-robot interaction dynamics. With this aim, this paper proposes a Model-Based Reinforcement Learning (MBRL) variable impedance controller to assist human operators in collaborative tasks. More in details, an ensemble of Artificial Neural Networks (ANNs) is used to learn a human-robot interaction dynamic model, capturing uncertainties. Such a learned model is kept updated during collaborative tasks execution. In addition, the learned model is used by a Model Predictive Controller (MPC) with Cross-Entropy Method (CEM). The aim of the MPC+CEM is to online optimize the stiffness and damping impedance control parameters minimizing the human effort (i.e, minimizing the human-robot interaction forces). The proposed approach has been validated through an experimental procedure. A lifting task has been considered as the reference validation application (weight of the manipulated part: 10 kg unknown to the robot controller). A KUKA LBR iiwa 14 R820 has been used as a test platform. Qualitative performance (i.e, questionnaire on perceived collaboration) have been evaluated. Achieved results have been compared with previous developed offline model-free optimized controllers and with the robot manual guidance controller. The proposed MBRL variable impedance controller shows improved human-robot collaboration. The proposed controller is capable to actively assist the human in the target task, compensating for the unknown part weight. The human-robot interaction dynamic model has been trained with a few initial experiments (30 initial experiments). In addition, the possibility to keep the learning of the human-robot interaction dynamics active allows accounting for the adaptation of human motor system.

Model-Based Reinforcement Learning Variable Impedance Control for Human-Robot Collaboration

Application of novel force control strategies to enhance robotic abrasive belt grinding quality of aero-engine blades

Learning variable impedance control is a powerful method to improve the performance of force control. However, current methods typically require too many interactions to achieve good performance. Data-inefficiency has limited these methods to learn force-sensitive tasks in real systems. In order to improve the sampling efficiency and decrease the required interactions during the learning process, this paper develops a data-efficient learning variable impedance control method that enables the industrial robots automatically learn to control the contact force in the unstructured environment. To this end, a Gaussian process model is learned as a faithful proxy of the system, which is then used to predict long-term state evolution for internal simulation, allowing for efficient strategy updates. The effects of model bias are reduced effectively by incorporating model uncertainty into long-term planning. Then the impedance profiles are regulated online according to the learned humanlike impedance strategy. In this way, the flexibility and adaptivity of the system could be enhanced. Both simulated and experimental tests have been performed on an industrial manipulator to verify the performance of the proposed method.

Efficient Force Control Learning System for Industrial Robots Based on Variable Impedance Control.

Kinematic parameters’ calibration is a powerful method to improve the accuracy of the robot. This paper proposes an effective kinematic self-calibration method for dual-manipulators based on virtual constraints to estimate the actual kinematic parameters of the robots. This method only needs a camera mounted on one robot end-effector (EE) and a calibration target attached to another robot EE. First, a new calibration error model based on the straight line constraint is established to formulate the positions’ misalignment error with the kinematic parameters’ error. Then, the particle swarm optimization algorithm is developed to generate the optimal calibration poses of the robots under the constraints, which are used to ensure the poses feasible and the measurement errors acceptable. Finally, the kinematic parameter errors are identified with the Levenberg–Marquardt algorithm. The experiments of the kinematic parameters’ calibration with the dual-manipulators system are designed. The experimental results showed that the high positional accuracy of both robots can be achieved.

/pdf/kinematic-self-calibration-method-for-dual-manipulators-1g8be03bku.pdf

Kinematic Self-Calibration Method for Dual-Manipulators Based on Optical Axis Constraint

The invention belongs to the field of control over a mechanical arm, and relates to an industrial mechanical arm force/position mixed control method based on a Kalman filter. The industrial mechanical arm force/position mixed control method based on the Kalman filter includes the steps that a six-dimensional force/moment sensor is installed at a wrist of a universal industrial mechanical arm; a mathematical model of actual contact force is built according to parameters of the sensor and a tail end tool of the mechanical arm; the actual contact force is calculated through the Kalman filter according to the measurement value of the force sensor; and according to the position information of six joints of the mechanical arm, the current position of the tail end of the mechanical arm and the like are calculated through forward kinematics. According to the industrial mechanical arm force /position mixed control method based on the Kalman filter, the environment rigidity is estimated on line in real time, interference in the force sensor is effectively inhibited, the stability of flexible control is increased, the universal industrial arm is made to have flexibility, and various complex tasks such as outline tracking, grinding, deburring and assembling can be completed.

Industrial mechanical arm force/position mixed control method based on Kalman filter

This paper proposes a hybrid force/position controller for industrial robotic manipulator based on Kalman filter. Firstly, the mathematical model of real contact force is built to estimate the actual contact force by applying Kalman filter using a force derivative to achieve the system state description. The estimated actual contact force is used to control the end-effector force as well as estimating the stiffness of environment. To refine the environment stiffness estimation the Recursive Least Square (RLS) technique has been employed. Owing to the fact that the general industrial manipulator only provides the position control mode, the position-based hybrid force/position control architecture is designed and realized by using the position tracking mode of the motion control card. The main advantages of the implemented controller is simplicity, computational efficiency and robustness to unknown environment, it is convenience for the general industrial manipulators. Besides, it lends itself for industrial manipulators in order to achieve compliant behavior and perform complex tasks. The proposed control structure is successfully validated by practical experiments. The results show that the controller has a satisfactory performance in term of force control and trajectory tracking and robustness to force/torque sensor measurement interferences.

Hybrid force/position control of industrial robotic manipulator based on Kalman filter

The invention belongs to the technical field of mechanical arm control, and particularly relates to a method of writing standard Chinese characters by using a six-degree-of-freedom mechanical arm under the control of force. The method comprises the steps of extracting the contour information of Chinese characters, planning the path of the tail end of the mechanical arm, designing an impedance control method, finding the position of a contact surface and controlling the force of writing processes and particularly comprises the following steps of: extracting the position information of contour points of Chinese characters based on the TTF character library of Windows, and falling contour curves into straight lines and spline curves; and converting the extracted contour position coordinates of Chinese characters into the coordinates of Cartesian space of an end effector of the mechanical arm. The method has the beneficial effects that the six-degree-of-freedom mechanical arm is used for completing Chinese character writing tasks under the control of force; the method is universal for industrial mechanical arms, can be widely applied to practical teaching and science and technology exhibitions and can also be used for completing tasks such as spraying and engraving in industrial fields; Chinese characters can be written on brittle objects by using the method; and the accuracy and safety of writing are ensured.

Chao Li

Papers

Efficient Force Control Learning System for Industrial Robots Based on Variable Impedance Control.

Kinematic Self-Calibration Method for Dual-Manipulators Based on Optical Axis Constraint

Industrial mechanical arm force/position mixed control method based on Kalman filter

Hybrid force/position control of industrial robotic manipulator based on Kalman filter

Method of writing standard Chinese characters by using six-degree-of-freedom mechanical arm under control of force