Robot Fault Research Articles

Robot Remote Monitoring and Fault Diagnosis Based on Industrial Internet of Things

In order to realize the remote monitoring of robots, a remote monitoring platform for industrial robots is designed based on the browser/server (B/S) architecture. Through this platform, users can check the real-time running parameters and running status of robots in any place with network. The Industrial Internet of Things scheme of remote platform system is proposed to adopt three-layer structure: on-site “perception layer,” information “transmission layer,” and remote data service center. The data acquisition controller of the whole system and the core part of the sensor layer is designed. The data acquisition controller adopts the embedded platform design, which can be directly connected with the control cabinet of the industrial robot to read the running status of the robot in real time, monitor the alarm and warning data, and it is transmitted to the local server and remote service center in the first time. At the same time, the robot can receive the control command of the server for remote debugging and fault maintenance. Aiming at the data model of industrial machinery parts, a fault prediction method based on BP neural network algorithm is proposed. According to the needs of the target algorithm and the analysis of the measurement results, an attempt is made to obtain a more feasible fault diagnosis and early warning method. Through the remote monitoring system, fault early warning and corresponding troubleshooting methods are realized.

Fault Diagnosis Method for Industrial Robots Based on DBN Joint Information Fusion Technology.

Aiming at the problems of the traditional industrial robot fault diagnosis model, such as low accuracy, low efficiency, poor stability, and real-time performance in multi-fault state diagnosis, a fault diagnosis method based on DBN joint information fusion technology is proposed. By studying the information processing method and the deep learning theory, this paper takes the fault of the joint bearing of the industrial robot as the research object. It adopts the technique of combining the deep belief network (DBN) and wavelet energy entropy, and the fault diagnosis of industrial robot is studied. The wavelet transform is used to denoise, decompose, and reconstruct the vibration signal of the joint bearing of the industrial robot. The normalized eigenvector of the reconstructed energy entropy is established, and the normalized eigenvector is used as the input of the DBN. The improved D-S evidence theory is used to solve the problem of fusion of high conflict evidence to improve the fault model's recognition accuracy. Finally, the feasibility of the model is verified by collecting the fault sample data and creating the category sample label. The experiment shows that the fault diagnosis method designed can complete the fault diagnosis of industrial robot well, and the accuracy of the test set is 97.96%. Compared with the traditional fault diagnosis model, the method is improved obviously, and the stability of the model is good; the utility model has the advantages of short time and high diagnosis efficiency and is suitable for the diagnosis work under the condition of coexisting multiple faults. The reliability of this method in the fault diagnosis of the joint bearing of industrial robot is verified.

A deep transferable motion-adaptive fault detection method for industrial robots using a residual–convolutional neural network

Recently, in various industrial fields, including automated manufacturing processes, industrial robots are becoming indispensable equipment; these robots perform repetitive tasks and increase the productivity of the production line with consistent precision and accuracy. Thus, fault diagnostics of industrial robots is an essential strategy to prevent the significant economic losses that can be caused by a sudden stop of a production line due to an industrial robot fault However, previous data-driven industrial robot fault diagnostics are limited because a pre-trained model built for a specific motion may not accurately or consistently detect faults in other motions, due to motion discrepancies. To overcome this difficulty, in this paper, we propose a deep transferable motion-adaptive fault detection method that uses torque ripples for fault detection of industrial robot gearboxes. The proposed method is composed of two stages: (1) a residual–convolutional neural network is used to enhance the performance of feature extraction for simple motions, after first refining raw torque signals by filtering out the motion-dependent signals (2) a binary-supervised domain adaptation is performed to detect faults adaptively on multi-axial motions through adversarial contrastive learning. The efficacy of the proposed method was validated using experimental data from unit-axis and multi-axial welding motions collected from a real industrial robot testbed. The proposed method showed superior fault detection accuracy for the motion adaptation task, as compared to existing methods.

Joint feature enhancement mapping and reservoir computing for improving fault diagnosis performance

Complicated industrial robot structure and harsh working conditions may cause signal features collected in the condition monitoring process to be seriously disturbed. In this paper, a joint feature enhancement mapping and reservoir computing (FEM-RC) method is presented to handle the industrial robot fault diagnosis problem. Firstly, a feature enhancement mapping (FEM) method is proposed to achieve intraclass distance minimization and interclass distance equalization to obtain an enhanced feature matrix. Then, the first reservoir computing (RC) network is adopted to map the original feature matrix to the feature enhancement matrix, and the second RC network is for fault type classification. The results of the experiment carried out on a six-axial industrial robot demonstrate that compared with other peer models, the present FEM-RC has better fault diagnosis performance and robustness.

Design and Implementation of an Emergency Undocking Curriculum for Robotic Surgery.

Current training for robotic surgery crisis management, specifically emergency robotic undocking protocol (ERUP), remains limited to anecdotal experience. A curriculum to impart the skills and knowledge necessary to recognize and complete a successful ERUP was developed using an education approach then evaluated. Baseline knowledge and confidence regarding ERUP were established for 5 robotic teams before completing 2 full-immersion simulations separated by an online self-paced learning module. In each simulation, teams operated on a perfused hydrogel model and were tasked to dissect a retroperitoneal tumor abutting a major vessel. During vascular pedicle ligation, a major vascular bleed and nonrecoverable robotic fault were remotely induced, necessitating ERUP with open conversion. After the simulation, participants completed surgery task load index (cognitive load assessment) and realism surveys. Weighted checklists scored participants' actions during each simulation. Surgical metrics including estimated blood loss, time to control bleeding, and undocking time were recorded. Curriculum retention was assessed by repeating the exercise at 6 months. Participants experienced high levels of cognitive demand and agreed that the simulation's realism and stress mimicked live surgery. Longitudinal analysis showed significant knowledge (+37.5 points, p = 0.004) and confidence (+15.3 points, p < 0.001) improvements from baseline to completion. Between simulations, checklist errors, undocking time, and estimated blood loss decreased (38⇾17, -40 seconds, and -500 mL, respectively), whereas action scores increased significantly (+27 points, p = 0.008). At 6 months, insignificant changes from curriculum completion were seen in knowledge (-4.8 points, p = 0.36) and confidence (+3.7 points, p = 0.1). This simulation-based curriculum successfully improves operative team's confidence, knowledge, and skills required to manage robotic crisis events.

Fault prognosis of industrial robots in dynamic working regimes: Find degradation in variations

Industrial robots are widely used in modern factories. Robot faults lead to the inevitable suspension of production lines. The prediction of robot failure can improve production capacity. However, it is challenging due to the variations of robots in dynamic working regimes. This paper presents a methodology of fault prognosis of industrial robots, including (1) a modeling approach of remaining useful life prediction using domain-generalization-adversarial long short-term memory to reduce the robot-to-robot variations, (2) an approach of two-stage health assessment based on principal component analysis-squared prediction error and p-chart that can reduce the disturbance of outliers in normal operations, and (3) a workflow containing feature extraction using wavelet packet decomposition, feature smoothing using exponential smoothing, feature normalization using z-score and feature selection using Pearson correlation coefficient. The case study of liquid crystal display transfer robots shows the methodology can effectively reduce the variations and improve the prediction of RUL.

Hexapod robot fault tolerant gait on slope and simulation verification

Currently, hexapod robot has been widely used in the field of outdoor detection and nuclear power disaster relief because of its adaptability to terrain, excellent movement performance and redundant fault tolerance. In this paper, fault-tolerant gait and motion planning method on the slope are researched. Firstly, the kinematics model of the robot on the slope is established, and the Jacobian matrix mapping the foot end velocity and joint velocity is derived, which is the basis of the robot movement. Then the phase sequence of the robot with one leg fault is planned to solve the problem of leg phase coordination. After then, the adjustment strategy of the centre of gravity on the slope is planned, which can effectively improve the stability of the robot. Finally, a simulation is demonstrated to verify the research.

Autonomic architecture for fault handling in mobile robots

This paper describes a generic autonomic architecture for use in developing systems for managing hardware faults in mobile robots. The method by which the generic architecture was developed is also described. Using autonomic principles, we focused on how to detect faults within a mobile robot and how specialized algorithms can be deployed to compensate for the faults discovered. We design the foundation of a generic architecture using the elements found in the MAPE-K and IMD architectures. We present case studies that show three different fault scenarios that can occur within the effectors, sensors and power units of a mobile robot. For each case study, we have developed algorithms for monitoring and analyzing data stored from previous tasks completed by the robot. We use the results from the case studies to create and refine a generic autonomic architecture that can be utilized for any general mobile robot setup for fault detection and fault compensation. We then describe a further case study which exercises the generic autonomic architecture in order to demonstrate its utility. Our proposal addresses fundamental challenges in operating remote mobile robots with little or no human intervention. If a fault does occur within the mobile robot during field operations, then having a self-automated strategy as part of its processes may result in the mobile robot continuing to function at a productive level. Our research has provided insights into the shortcomings of existing robot design which is also discussed.

Data-driven Models for Fault Classification and Prediction of Industrial Robots

Economic data acquisition and storage have been key enablers to pave the way for data-driven predictions of machine downtimes. Regarding industrial robots, such predictions can maximize the robot’s availability and effective life span. This paper focuses on the comparison of different data-driven models for robot fault prediction and classification by applying them to a data set derived from a robot test bed and illuminates the data transformation process from raw sensor data to domain knowledge motivated robot health indicators.

Fault analysis of industrial robots based on self‐organised critical theory

Industrial robots, as a structurally sophisticated mechatronics system, have a high cost of routine maintenance and repair. Repairs after fault require the corresponding manpower and material resources, and have hysteresis. If the fault can be predicted in a timely and accurate manner, the maintenance process can be carried out in advance, and the hidden dangers can be eliminated to fundamentally solve the fault problem. Based on the self-organised critical theory (SOC theory), this article draws lessons from its self-organisation evolution model and uses the self-organised criticality of industrial robot fault to establish an autoregressive moving average model (ARMA model) for industrial robots. According to the analysis of residual value and the explanation for the faults of industrial robots, find ways and means to prevent and reduce faults.

Maximizing the Size of Self-Motion Manifolds to Improve Robot Fault Tolerance

One measure of a robot's fault tolerance is the size of its self-motion manifold. This letter presents a new methodology for finding the largest self-motion manifold(s) of kinematically redundant robots, that consists of two algorithms. Because large self-motion manifolds occur near singular configurations, the first algorithm is designed to identify singularities of all ranks, including high-rank singularities. One unique feature of this algorithm is its ability to deal with the ill conditioned nature of singular vectors when there are multiple nearly equal singular values. The second algorithm is constructed to compute the self-motion manifolds that contain these singular configurations by iteratively moving along the null space of the robot's Jacobian. An important aspect of this computation is to deal with singularities along the manifold, where the null space is high dimensional. These two algorithms are applied to the well-known Mitsubishi PA-10 to illustrate their effectiveness at identifying singularities and computing the largest self-motion manifold(s).

Research on Fault Diagnosis Method of Industrial Robots Based on Case-Based Reasoning

The application enterprises of industrial robots generally do not have the ability to independently diagnose and repair the faults of industrial robots. When robots fail and stop operation, they often need to notify external service providers to diagnose and repair the fault sites of enterprises, which seriously affects the production tempo of enterprises. The loss of the enterprise is aggravated. Therefore, according to the dynamic evolution of the industrial robot fault diagnosis method and the characteristics of the case reasoning method, this paper applies the Case-Based Reasoning method to the industrial robot fault diagnosis. By using the concept of Inverse Document Frequency (IDF) index in information theory, the case organization, case retrieval and case retention technology in CBR technology are designed, which provides an effective method for the diagnosis of industrial robot faults.

An Augmented Reality System to Support Fault Visualization in Industrial Robotic Tasks

The digitalization is transforming the very nature of factories, from automated systems to intelligent ones. In this process, industrial robots play a key role. Even if repeatability, precision and velocity of the industrial manipulators enable reaching considerable production levels, factories are required to face an increasingly competitive market, which requires being able to dynamically adapt to different situations and conditions. Hence, facilities are moving toward systems that rely on the collaboration between humans and machines. Human workers should understand the behavior of the robots, placing trust in them to properly collaborate. If a fault occurs on a manipulator, its movements are suddenly stopped for security reasons, thus workers may not be able to understand what happened to the robot. Therefore, the operators' stress and anxiety may increase, compromising the human-robot collaborative scenario. This work fits in this context and it proposes an adaptive Augmented Reality system to display industrial robot faults by means of the Microsoft HoloLens device. Starting from the methodology employed to identify which virtual metaphors best evoke robot faults, an adaptive modality is presented to dynamically display the metaphors in positions close to the fault location, always visible from the user and not occluded by the manipulator. A comparison with a non adaptive modality is proposed to assess the effectiveness of the adaptive solution. Results show that the adaptive modality allows users to recognize faults faster and with fewer movements than the non adaptive one, thus overcoming the limitation of the narrow field-of-view of the HoloLens device.

Robot fault detection and remaining life estimation for predictive maintenance

In this work some possible solutions to implement a Robotics-oriented predictive maintenance approach are discussed. The data-driven methodology is described from the data collection to the design of an appropriate dataset and finally to the use of some of the most promising algorithms in the field of machine learning. The whole process is composed by several building blocks that can be combined to realize a data analysis on industrial robots. Some of the most promising techniques in Predictive Maintenance for Industrial machines were included in the proposed methodology, together with a Survival Analysis study, and then evaluated with proper performance metrics. Experimenting this methodology on a real use-case with Comau industrial robots showed the validity of the approach and opened to the inclusion of such a process in a service-oriented solution.

A fault-tolerant approach to the control of a battery assembly system

The paper concerns fault-tolerant control of a real battery assembly system which is under a pilot implementation at RAFI GmbH Company (one of the leading electronic manufacturing service providers in Germany). The proposed framework is based on an interval analysis approach, which along with max-plus algebra, allows describing uncertain discrete event system such as the production one being considered in this paper. Having a mathematical system description, a model predictive control-based fault tolerant strategy is developed which can cope with both processing, transportation and mobile robot faults. In particular, it enables tolerating (up to some degree) the influence of these faults on the overall system performance. As a result, a novel robust predictive fault-tolerant strategy is developed that is applied to the advanced battery assembly system. The final part of the paper shows the implementation and experimental validation of the proposed strategy. The proposed approach is tested against single as well as simultaneous faults concerning processing, transportation and mobile robots.

Thruster fault diagnosis method based on Gaussian particle filter for autonomous underwater vehicles

Autonomous Underwater Vehicles (AUVs) generally work in complex marine environments. Any fault in AUVs may cause significant losses. Thus, system reliability and automatic fault diagnosis are important. To address the actuator failure of AUVs, a fault diagnosis method based on the Gaussian particle filter is proposed in this study. Six free-space motion equation mathematical models are established in accordance with the actuator configuration of AUVs. The value of the control (moment) loss parameter is adopted on the basis of these models to represent underwater vehicle malfunction, and an actuator failure model is established. An improved Gaussian particle filtering algorithm is proposed and is used to estimate the AUV failure model and motion state. Bayes algorithm is employed to perform robot fault detection. The sliding window method is adopted for fault magnitude estimation. The feasibility and validity of the proposed method are verified through simulation experiments and experimental data.

Practical relevance of faults, diagnosis methods, and tolerance measures in elastically actuated robots

Elastically actuated robots promise safe human–robot interaction and energy-efficient motions. Yet, increased complexity and critical operation states might increase the practical fault risk. This paper explores faults in such robots using expert data from the field and identifies components that show increased fault occurrence: the highest fault sensitivity occurs in kinematics, electronics, sensors, and software. Since elastic actuators are an active field of research, few cases of industrial application exist and thus most experts in this study have academic background. Beyond assessing fault sensitivity, countermeasures such as redundant design are compiled. A brief literature review discusses fault diagnosis and fault-tolerant design with respect to these insights. Despite the availability of a few promising methods in robotics, neither diagnosis nor tolerance do receive sufficient recognition leaving potential for practical application.

Multiple Leader Candidate and Competitive Position Allocation for Robust Formation against Member Robot Faults.

This paper proposes a Multiple Leader Candidate (MLC) structure and a Competitive Position Allocation (CPA) algorithm which can be applicable for various applications including environmental sensing. Unlike previous formation structures such as virtual-leader and actual-leader structures with position allocation including a rigid allocation and an optimization based allocation, the formation employing the proposed MLC structure and CPA algorithm is robust against the fault (or disappearance) of the member robots and reduces the entire cost. In the MLC structure, a leader of the entire system is chosen among leader candidate robots. The CPA algorithm is the decentralized position allocation algorithm that assigns the robots to the vertex of the formation via the competition of the adjacent robots. The numerical simulations and experimental results are included to show the feasibility and the performance of the multiple robot system employing the proposed MLC structure and the CPA algorithm.

Insect Pheromone Strategy for the Robots Coordination

Insect colony inspires scientists for years to create similar behavior in the robotic application. The main goal of our work was to develop simple and powerful algorithm which will accept dynamically changes in the size of a robot swarm due the mission. This algorithm is suitable for situations where unpredictable conditions may lead to robot fault in multi-robotics system and mission completion is endangered. In this article we would like to investigate properties of a simple pheromone based algorithm. The algorithm operates as cellular automata and partially uses an insect pheromone strategy for the robots coordination. Our abstract model is a decentralized adaptive system with a shared memory which represents the environment.

Case-based reasoning combined with statistics for diagnostics and prognosis

Many approaches used for diagnostics today are based on a precise model. This excludes diagnostics of many complex types of machinery that cannot be modelled and simulated easily or without great effort. Our aim is to show that by including human experience it is possible to diagnose complex machinery when there is no or limited models or simulations available. This also enables diagnostics in a dynamic application where conditions change and new cases are often added. In fact every new solved case increases the diagnostic power of the system. We present a number of successful projects where we have used feature extraction together with case-based reasoning to diagnose faults in industrial robots, welding, cutting machinery and we also present our latest project for diagnosing transmissions by combining Case-Based Reasoning (CBR) with statistics. We view the fault diagnosis process as three consecutive steps. In the first step, sensor fault signals from machines and/or input from human operators are collected. Then, the second step consists of extracting relevant fault features. In the final diagnosis/prognosis step, status and faults are identified and classified. We view prognosis as a special case of diagnosis where the prognosis module predicts a stream of future features.