Polishing Task Research Articles

Text Polishing with Chinese Idiom: Task, Datasets and Pre-trained Baselines

This work presents the task of text polishing, which generates a sentence that is more graceful than the input sentence while retaining its semantic meaning. Text polishing has great value in real usage and is an important component in modern writing assistance systems. However, the task is still not well studied in the literature. Further research in this important direction requires more formal task definitions, benchmark datasets, and powerful baseline models. In this work, we formulate the task as a context-dependent text generation problem and conduct a case study on the text polishing with Chinese idiom. To circumvent the difficulties of task data annotation, we propose a semi-automatic data construction pipeline based on human-machine collaboration, and establish a large-scale text polishing dataset consisting of 1.5 million instances. We propose two types of task-specific pre-training objectives for the text polishing task and implement a series of Transformer-based models pre-trained on a massive Chinese corpus as baselines. We conduct extensive experiments with the baseline models on the constructed text polishing datasets and have some major findings. The human evaluation further reveals the polishing ability of the final system.

Model predictive impedance control with Gaussian processes for human and environment interaction

Robotic tasks which involve uncertainty – due to variation in goal, environment configuration, or confidence in task model – may require human input to instruct or adapt the robot. In tasks with physical contact, several existing methods for adapting robot trajectory or impedance according to individual uncertainties have been proposed, e.g., realizing intention detection or uncertainty-aware learning from demonstration. However, isolated methods cannot address the wide range of uncertainties jointly present in many tasks.To improve generality, this paper proposes a model predictive control (MPC) framework which plans both trajectory and impedance online, can consider discrete and continuous uncertainties, includes safety constraints, and can be efficiently applied to a new task. This framework can consider uncertainty from: contact constraint variation, uncertainty in human goals, or task disturbances. An uncertainty-aware task model is learned from a few (≤3) demonstrations using Gaussian Processes. This task model is used in a nonlinear MPC problem to optimize robot trajectory and impedance according to belief in discrete human goals, human kinematics, safety constraints, contact stability, and frequency-domain disturbance rejection. This MPC formulation is introduced, analyzed with respect to convexity, and validated in co-manipulation with multiple goals, a collaborative polishing task, and a collaborative assembly task.

EMG space similarity feedback promotes learning of expert-like muscle activation patterns in a complex motor skill.

Augmented feedback provided by a coach or augmented reality system can facilitate the acquisition of a motor skill. Verbal instructions and visual aids can be effective in providing feedback about the kinematics of the desired movements. However, many skills require mastering not only kinematic, but also complex kinetic patterns, for which feedback is harder to convey. Here, we propose the electromyography (EMG) space similarity feedback, which may indirectly convey kinematic and kinetic feedback by comparing the muscle activations of the learner and an expert in the task. The EMG space similarity feedback is a score that reflects how well a set of muscle synergies extracted from the expert can reconstruct the learner's EMG when performing the task. We tested the EMG space similarity feedback in a virtual bimanual polishing task that uses a robotic system to simulate the dynamics of a real polishing operation. We measured the expert's and learner's EMG from eight muscles in each arm during the real and virtual polishing tasks, respectively. The goal of the virtual task was to smoothen the surface of a virtual object. Therefore, we defined performance in the task as the smoothness of the object at the end of a trial. We separated learners into real feedback and null feedback groups to assess the effects of the EMG space similarity feedback. The real and null feedback groups received veridic and no EMG space similarity feedback, respectively. Subjects participated in five training sessions on different days, and we evaluated their performance on each day. Subjects in both groups were able to increase smoothness throughout the training sessions, with no significant differences between groups. However, subjects in the real feedback group were able to improve in the EMG space similarity score to a significantly greater extent than the null feedback group. Additionally, subjects in the real feedback group produced muscle activations that became increasingly consistent with an important muscle synergy found in the expert. Our results indicate that the EMG space similarity feedback promotes acquiring expert-like muscle activation patterns, suggesting that it may assist in the acquisition of complex motor skills.

Learning Stable Nonlinear Dynamics and Interactive Force-Aware Variable Impedance Control for Robotic Contact Tasks

This paper presents a novel approach for efficient robot programming in tasks such as polishing and grinding, which addresses the challenges of stability and adaptability in robot motion and control. The approach involves learning movement skills and variable impedance control from human demonstrations. Firstly, a non-parametric Gaussian mixture model and globally stable method are proposed for learning robot rhythmic skills with nonlinear dynamics by taking into account the characteristics of robot motion in human-robot physical interaction. Then, adaptive variable impedance control mechanisms are constructed using Gaussian mixture regression, enabling adaptive perception and online control of unknown environments. Finally, a stability-consideration variable impedance controller is designed to minimize the total control-torque while ensuring motion accuracy, which allows efficient adaptation to uncertain environments and complex tasks. The paper provides a novel framework of robot rhythmic skill learning and variable impedance control for robotic complex surface polishing tasks with nonlinear dynamics based on multiple physical human demonstrations, and verifies the proposed methods with a real Franka-robot polishing task. These implementations would be with applications in intelligent manufacturing, robot-assisted rehabilitation.

Ergonomic design and evaluation of gemstone polishing workstation

Objectives. Gemstone polishers suffer from musculoskeletal problems due to constrained working postures, substantially influenced by the poor design of conventional gemstone polishing workstations. The present study investigated the effects of three workstation adjustment parameters (illumination at workstation, polishing height, tool post position) on postural angle, muscle activity and perceived postural discomfort in order to develop recommendations for new gemstone polishing workstations. Methods. Twelve professional gemstone polishers performed the polishing task on a prototype test-rig in 27 different test conditions. Taguchi’s L27 orthogonal array was employed to estimate the optimum setting for gemstone polishing workstation design. Results. Study variables had a significant influence on postural angles. Polishing height substantially influenced muscle activity and perceived discomfort in the shoulder and lower back region. Tool post position was found to be the determinant factor for forearm muscle activity and perceived discomfort in the wrist/forearm region. Based on the findings, illumination of 1500 lux, polishing at 15 cm above the elbow height and a 20° tool position was recommended. Conclusions. Our findings present an alternative approach to using Taguchi’s design of experiment for workstation improvement, which has received very little attention in ergonomics studies. A confirmation test was conducted to validate the study recommendations.

Physiological and perceptual consequences of trust in collaborative robots: An empirical investigation of human and robot factors

Measuring trust is an important element of effective human-robot collaborations (HRCs). It has largely relied on subjective responses and thus cannot be readily used for adapting robots in shared operations, particularly in shared-space manufacturing applications. Additionally, whether trust in such HRCs differ under altered operator cognitive states or with sex remains unknown. This study examined the impacts of operator cognitive fatigue, robot reliability, and operator sex on trust symptoms in collaborative robots through both objective measures (i.e., performance, heart rate variability) and subjective measures (i.e., surveys). Male and female participants were recruited to perform a metal surface polishing task in partnership with a collaborative robot (UR10), in which they underwent reliability conditions (reliable, unreliable) and cognitive fatigue conditions (fatigued, not fatigued). As compared to the reliable conditions, unreliable robot manipulations resulted in perceived trust, an increase in both sympathetic and parasympathetic activity, and operator-induced reduction in task efficiency and accuracy but not precision. Cognitive fatigue was shown to correlate with higher fatigue scores and reduced task efficiency, more severely impacting females. The results highlight key interplays between operator states of fatigue, sex, and robot reliability on both subjective and objective responses of trust. These findings provide a strong foundation for future investigations on better understanding the relationship between human factors and trust in HRC as well as aid in developing more diagnostic and deployable measures of trust.

A Virtual Element-Based Postural Optimization Method for Improved Ergonomics During Human-Robot Collaboration

Human-robot collaboration is becoming increasingly popular in the manufacturing industry, opening the door to a large range of applications by combining the complementary skills of the human worker and the robot. Collaborative robots are also a solution to decrease the operator workload and indirectly reduce the risk of occupational injuries such as musculoskeletal disorders (MSDs). The latter represents one of the major causes of absenteeism at work. Thanks to the development of human tracking devices, it is possible to monitor the operator, analyze the postures, and assess the associated MSD risk. In this paper, we present a novel ergonomics optimization framework that performs postural optimization based on the virtual element method. A feedback interface is developed whereby the user is informed about non-ergonomic postures and an improved body pose is proposed. The workpiece position controller module acts on the cobot end-effector and indirectly on the co-manipulated part in such a way that the operator’s posture is improved. The framework was validated by a user study performed on a human-robot collaboration task whereby the subject polishes a part hold by the robot. The conducted study of the user’s perception and REBA scores showed promising results. <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">Note to Practitioners</i> —This paper is motivated by the problem of non-ergonomic posture of workers in hybrid workcells. The proposed approach makes use of virtual elements (springs and dampers) to build a mechanical model of the human body posture and perform postural optimization. The obtained body joint angles are fed into two modules of the framework. First, a graphical interface displays the current pose of the user and proposes to him the improved posture. Second, a controller adapts the pose of the workpiece hold by the collaborative robot. This is realized by computing a displacement vector between the wrist current and optimized positions. The use of such a framework was demonstrated on a collaborative polishing task whereby the robot adjusts the position of the workpiece. After a user study test with 10 participants, joint data were collected and the REBA scores of different subtasks were measured and compared. The results from these preliminary experiments showed that the proposed approach improves the human body postures and offers a promising solution to enhance ergonomics by the robot assistance in case of robotic workcells. The conducted survey also shows an overall positive subject’s perception of the system.

Denim-fabric-polishing robot size optimization based on global spatial dexterity

Abstract. This paper presents a novel method to make denim-fabric-polishing robots perform their primary task flexibly and efficiently within a limited workspace. Link lengths are optimized based on an adaptive fireworks algorithm to improve the comprehensive dexterity index. A forward kinematics analysis of the denim-fabric-polishing robot is conducted via the D–H method; the workspace is analyzed according to the needs at hand to determine the range of motion of each joint. To solve the movement condition number of the Jacobian matrix, the concept of low-condition-number probability is established, and a comprehensive dexterity indicator is constructed. The influence of the robot's size on the condition number and comprehensive dexterity index is determined. Finally, the adaptive fireworks algorithm is used to establish the objective optimization function by integrating the dexterity index and other performance indicators. The optimization results show that when the comprehensive dexterity index is taken as the optimization objective, the dexterity comprehensive index and other performance indices of the robot are the lowest; that is, the robot is more flexible. Compared with the traditional genetic algorithm and particle swarm algorithm, the adaptive fireworks algorithm proposed in this paper has better solving speed and solving precision. The optimized workspace of the robot meets the requirements of the polishing task. The design also yields a sufficiently flexible, efficient, and effective robot.

Estimating Human Wrist Stiffness during a Tooling Task.

In this work, we propose a practical approach to estimate human joint stiffness during tooling tasks for the purpose of programming a robot by demonstration. More specifically, we estimate the stiffness along the wrist radial-ulnar deviation while a human operator performs flexion-extension movements during a polishing task. The joint stiffness information allows to transfer skills from expert human operators to industrial robots. A typical hand-held, abrasive tool used by humans during finishing tasks was instrumented at the handle (through which both robots and humans are attached to the tool) to assess the 3D force/torque interactions between operator and tool during finishing task, as well as the 3D kinematics of the tool itself. Building upon stochastic methods for human arm impedance estimation, the novelty of our approach is that we rely on the natural variability taking place during the multi-passes task itself to estimate (neuro-)mechanical impedance during motion. Our apparatus (hand-held, finishing tool instrumented with motion capture and multi-axis force/torque sensors) and algorithms (for filtering and impedance estimation) were first tested on an impedance-controlled industrial robot carrying out the finishing task of interest, where the impedance could be pre-programmed. We were able to accurately estimate impedance in this case. The same apparatus and algorithms were then applied to the same task performed by a human operators. The stiffness values of the human operator, at different force level, correlated positively with the muscular activity, measured during the same task.

Design of a polishing tool for collaborative robotics using minimum viable product approach

ABSTRACTA collaborative tool for robotic polishing is developed in this work in order to allow the simultaneous operation of the robot system and human operator to cooperatively carry out the polishing task. For this purpose, the collaborative environment is detailed and the polishing application is designed. Moreover, the polishing tool is developed and its implementation using the minimum viable product approach is obtained. Furthermore, a robust hybrid position-force control is proposed to use the developed tool attached to a robot system and some experiments are given to show its performance.

Augmented reality based approach for on-line quality assessment of polished surfaces

Augmented reality is considered one of the enabling technologies of the fourth industrial revolution, within the Industry 4.0 program and beyond. Indeed, augmented reality solutions can increase the working quality and the productivity and allow a better use of the human resources. This technology can help the operator in the industrial applications during the crucial phases of the processes. Since the quality assessment of the surfaces is recognized to be a key phase in the polishing process, in this paper we propose a novel method that exploits augmented reality to support the operators during this phase. The metrology data measured by a surface measurement system are directly projected on the polished component through an augmented reality headset worn by the operators and used to assess the quality of the worked surfaces. Rather than imagine how a certain parameter change can affect the result achieved, the information is directly there on the component’s surface. Users can see from the data where refinements are required and make better and faster decisions, which is compelling for its potential beyond industrial polishing. The proposed method is implemented and validated on an industrial cell, where the robot automatically perform the polishing task and move the head of the surface measurement system along the surface to measure the metrology parameters. Thanks to the proposed approach, the end-user and the operator can directly see on the component if the quality reached satisfies the specifications or if some parts of the surface require further refinements through additional polishing steps.

Learning Force‐Relevant Skills from Human Demonstration

Many human manipulation skills are force relevant, such as opening a bottle cap and assembling furniture. However, it is still a difficult task to endow a robot with these skills, which largely is due to the complexity of the representation and planning of these skills. This paper presents a learning‐based approach of transferring force‐relevant skills from human demonstration to a robot. First, the force‐relevant skill is encapsulated as a statistical model where the key parameters are learned from the demonstrated data (motion, force). Second, based on the learned skill model, a task planner is devised which specifies the motion and/or the force profile for a given manipulation task. Finally, the learned skill model is further integrated with an adaptive controller that offers task‐consistent force adaptation during online executions. The effectiveness of the proposed approach is validated with two experiments, i.e., an object polishing task and a peg‐in‐hole assembly.

A model-based residual approach for human-robot collaboration during manual polishing operations

A fully robotized polishing of metallic surfaces may be insufficient in case of parts with complex geometric shapes, where a manual intervention is still preferable. Within the EU SYMPLEXITY project, we are considering tasks where manual polishing operations are performed in strict physical Human-Robot Collaboration (HRC) between a robot holding the part and a human operator equipped with an abrasive tool. During the polishing task, the robot should firmly keep the workpiece in a prescribed sequence of poses, by monitoring and resisting to the external forces applied by the operator. However, the user may also wish to change the orientation of the part mounted on the robot, simply by pushing or pulling the robot body and changing thus its configuration. We propose a control algorithm that is able to distinguish the external torques acting at the robot joints in two components, one due to the polishing forces being applied at the end-effector level, the other due to the intentional physical interaction engaged by the human. The latter component is used to reconfigure the manipulator arm and, accordingly, its end-effector orientation. The workpiece position is kept instead fixed, by exploiting the intrinsic redundancy of this subtask. The controller uses a F/T sensor mounted at the robot wrist, together with our recently developed model-based technique (the residual method) that is able to estimate online the joint torques due to contact forces/torques applied at any place along the robot structure. In order to obtain a reliable residual, which is necessary to implement the control algorithm, an accurate robot dynamic model (including also friction effects at the joints and drive gains) needs to be identified first. The complete dynamic identification and the proposed control method for the human-robot collaborative polishing task are illustrated on a 6R UR10 lightweight manipulator mounting an ATI 6D sensor.

Conformal polishing approach: Tool footprint analysis

Polishing process is one of the most critical manufacturing processes during a metal part production because it determines the final quality of the product. Free-form surface polishing is a handmade process with lots of rejected parts, scrap generation and time and energy consumption. Two different research lines are being developed: prediction models of the final surface quality parameters and an analysis of the amount of material removed depending on the polishing parameters to predict the tool footprint during the polishing task. This research lays the foundations for a future automatic conformal polishing system. It is based on rotational and translational tool with dry abrasive in the front mounted at the end of a robot. A tool to part concept is used, useful for large or heavy workpieces. Results are applied on different curved parts typically used in tooling industry, aeronautics or automotive. A mathematical model has been developed to predict the amount of material removed in function of polishing parameters. Model has been fitted for different abrasives and raw materials. Results have shown deviations under 20% that implies a reliable and controllable process. Smaller amount of material can be removed in controlled areas of a three-dimensional workpiece.

Recovery of ibuprofen from pharmaceutical wastes using ionic liquids

This work aims at developing a process to valorise pharmaceutical wastes through the recovery of pharmaceutically active compounds. The ibuprofen extraction and isolation from solid pharmaceutical wastes is used here as a case study and an integrated approach comprising the ibuprofen solid–liquid extraction, the removal of the insoluble excipients present in the pills, the target drug recovery and the recycling of the aqueous solutions is proposed. The present work is centred on the optimization of the first (solid–liquid extraction) and third (drug recovery) steps mentioned above. For the solid–liquid extraction step, various ionic liquid aqueous solutions were tested, tetrabutylammonium chloride ([N4444]Cl) being adopted to further optimize the process. A solution composed of 45 wt% of [N4444]Cl + 5 wt% of citrate buffer + 50 wt% of H2O led to the highest ibuprofen extraction efficiency (EEIBU = 97.92 ± 2.65%) while in the absence of citrate the extraction efficiency was somewhat lower (EEIBU = 93.53 ± 0.62%). The polishing task was affected by the type of aqueous solution utilized during the solid–liquid extraction step: in the presence of citrate buffer water was not prone to induce significant ibuprofen precipitation (maximum REIBU of 34.71 ± 4.00%) the aqueous KCl solution being the best option (maximum REIBU of 87.97 ± 1.00%); when no citrate buffer is used water can be used as an anti-solvent with a maximum REIBU of 91.60 ± 0.19% while aqueous KCl solutions lead to an REIBU up to 97.07 ± 0.14%. Based on these results an integrated process is proposed for the ibuprofen recovery and isolation aimed at adding value to pharmaceutical wastes.

Acoustic emission-based in-process monitoring of surface generation in robot-assisted polishing

The applicability of acoustic emission (AE) measurements for in-process monitoring of surface generation in the robot-assisted polishing (RAP) was investigated. Surface roughness measurements require interruption of the process, proper surface cleaning and measurements that sometimes necessitate removal of the part from the machine tool. In this study, stabilisation of surface roughness during polishing rotational symmetric surfaces by the RAP process was monitored by AE measurements. An AE sensor was placed on a polishing arm in direct contact with a bonded abrasive polishing tool, and a cylindrical workpiece in Vanadis 4E steel was polished in 40 polishing passes from an initial turned surface roughness Ra = 3.1 µm down to Ra = 0.07 µm. The polishing task was performed in five intervals and after 4, 8, 20, 30 and 40 passes, the resulting surface roughness was measured. The results show a decreasing trend in measured AE signal power and RMS, which is well qualitatively correlated with the development of surface roughness during polishing. The trend allows the identification of an asymptote representing the process completion (stabilisation of surface roughness), reliable for correct in-process determination of the process endpoint. This makes it possible to reliably determine the right time for changing the polishing media to finer abrasive when applying a given set of parameters is no longer effective to create a smoother surface, thus improving the efficiency of the process. The findings enabling automatic detection of optimal process endpoint allow intelligent process control, creating fundamental elements in development of robust fully automated RAP process for its widespread industrial application.

Dynamic Modeling and Force Control of a Redundant Robot for Polishing Applications

AbstractFor many robotic applications with tasks such as cutting, assembly or polishing, it is necessary to get in contact with the surrounding. In this paper a redundant robot with seven degrees of freedom in a metal polishing task is considered. For simulation as well as for the controller design a dynamic model of the robot and a contact model are required. The equations of motion of the robot are calculated with the Projection Equation in subsystem representation and the contact model contains linear tool elasticities and work piece elasticities. In the case of a polishing task, a constant contact force during the process is required even if the robot moves along a trajectory. Thus some degrees of freedom of the robot tool center point have to be position controlled while the other ones have to be force controlled. The redundant robot offers the possibility to avoid singular positions or to maximize the available end‐effector forces within the inverse kinematics and is therefore best suited for polishing large objects. The actual process forces are measured with a six axis force‐torque‐sensor mounted at the tool center point. These forces are used in a parallel force/position control law to achieve the desired behavior. Results from measurements of a test arrangement are presented. (© 2014 Wiley‐VCH Verlag GmbH &amp; Co. KGaA, Weinheim)

Vision Sensor Application for Intelligent Polishing Robotic Cell

Robotics especially on polishing process has not been exploited successfully in manufacturing, despite tremendous research efforts and potential needs in industries. There is still much concern on how robots are used in industries and stumbling blocks in the deployment of technology. In this study, the techniques to alleviate difficulties in using robots and deploying advanced technologies in the real manufacturing environment will be developed. However, the mirror polishing process still depends on human expert, compared to CNC and CAD/CAM technology. On top of that, undesirable working condition exists due to dust and noise. Next, it is quite difficult to find skilled technicians where existing industrial robot (normally with repetitive accuracy of 0.1mm) not suitable for mirror polishing task requirement (10nm). This study addresses idea to develop and automate the polishing process by using robots. It applies specifically on flat steel surfaces and the image of the workpiece is captured by vision sensor. The image captured is extracted to develop ANFIS (Adaptive Neuro Fuzzy Inference System) to determine the roughness so that correct polishing parameter will be applied on the surface of the workpiece when polishing is done. ANFIS is well suited to the management of uncertainty, is introduced to address the uncertainty problem associated with surface roughness.

603 ロボットによる力制御のための能動作業台 : 小型部品のバフ研磨作業への応用(OS9-1,オーガナイズドセッション:9 ロボット技術とメカトロ制御技術)

603 ロボットによる力制御のための能動作業台 : 小型部品のバフ研磨作業への応用(OS9-1,オーガナイズドセッション:9 ロボット技術とメカトロ制御技術)

Development of an Automatic Mold Polishing System

In this study, an automatic mold polishing system (AMPS) has been successfully developed. The AMPS integrates with mold geometry process kernel, path planner, process planner, and force control robot into a system. The polishing path is generated and drawn on the three-dimensional mold surface for verification. The AMPS can also automatic process the task planning according to the polishing requirements of the mold. The task list includes all the procedures and optimal polishing parameters. The polishing tasks were then executed by a force controllable robot. The experiment shows that the resultant of surface roughness satisfies the expectation in the process planning. Note to Practitioners -In the manufacturing of molds, the most time-consuming process is on the final stage the polishing process. The polishing processes in the mold industry are almost conducted manually. The worker needs exceptional skill that can only be trained by experienced master. For mass production of molds, automatic polishing of the mold is a necessary. From the long-term research effort, we have constructed an automatic mold polishing system (AMPS). The AMPS includes a dedicated 5-axes robot to execute the polishing task; a force control mechanism to hold the tool and apply constant pressure on the mold surface; and a software system that can read the geometry of the mold surface and automatically generate the detail polishing process and the polishing path for the robot. The AMPS now can handle polishing process for planar surfaces with obstacle inside and surfaces of revolution. Research effort is now concentrated on dealing with the free-from surfaces. Besides, reading the Initial Graphics Exchange System (IGES) format from various computer-aided design systems is another important issue for the system to be more useful. In all, the AMPS is just begin to accumulate any further researches and developments that may sustain various requirements of polishing applications.