Improving robot accuracy is important for many industrial applications. Laser trackers and retroreflectors are often used to collect measurements for error model identification. Typically, retroreflectors are attached to the robot end effector, collecting position information. To determine orientation information, each robot pose must be measured multiple times, where the retroreflectors are located at a different position on the end effector for each measurement. Fixturing and measurement errors are introduced each time, thus affecting model accuracy and increasing measurement time. In this paper, a new retroreflector, named the SmartTRACK Sensor (STS), is used. This retroreflector can simultaneously measure both position and orientation information. Methodologies for the modeling, identification and compensation of robot kinematic errors are proposed. In the experiments conducted on a FANUC LR Mate 200 i robot, the mean residuals for the actual and modeled positions, respectively, are 4.559 to 0.076 mm, thus, a 98% reduction in geometric error was realized.

View full abstract

5-axis machine tools are widely utilized in manufacturing complex sculptured components. Typically, tool-paths for 5-axis machine tools are discontinuous and they must be smoothened to generate continuous motion. This paper proposes a novel real-time interpolation algorithm for 5-axis machine tools and industrial robots to generate continuous and rapid feed motion along discrete point-to-point tool-paths by locally blending tool position and orientation, i.e. tool-pose, within user-specified tolerances. Proposed algorithm utilizes a filtering technique to smoothen pose trajectory of the tool and continuously interpolate 6DOF motion in real-time accurately. Tool-pose errors are controlled locally at junction points, i.e. corners, of the tool-path by timing consecutive feed commands precisely. Kinematic limits of the machine axes are considered in determining maximum feedrate along consecutive moves. Simulation studies show that proposed algorithm can interpolate 5-axis machining tool-paths accurately within kinematic limits in a computationally effective and real-time suitable scheme.

View full abstract

Despite the recent rapid increase of robots in manufacturing, much of the garment industry remains largely reliant on human labor. Part of the difficulty in introducing automation to this industry lies in the complexity of modeling and manipulating flexible materials, such as fabric, especially at the high speed needed to handle the throughput expected of a modern clothing manufacturer. Previous work has demonstrated the ability to control fabric motion through a distributed actuation system under the assumption that the fabric maintains a rigid planar shape. This paper introduces a method by which a depth sensor, in conjunction with a template matching algorithm described in a companion paper [1], can identify gross fabric deformation in real time and demonstrates a control method for restoring the fabric to a rigid planar form.

View full abstract

The garment manufacturing industry is largely reliant on human labor in conjunction with hard-automation for a great deal of tasks due to robotic challenges in manipulating flexible materials (fabric). One previously proposed solution for this task is the use of a distributed actuation system, which has been demonstrated in open-loop fashion. This paper introduces a method of feedback through fast image processing by exploiting some of the features inherent to the sewing process by tracking (possibly occluded) fabric via polygon matching that allows for full closed-loop control in real time.

View full abstract

In order to reduce unnecessary power consumption by machine tools, we propose improved control methods for current idling reduction systems that shut off the power to the motor. When current idling reduction system is used, the servo lock function does not work, causing a displacement error that increases power consumption and vibration. Therefore, we devised two control methods to reduce the displacement error of such an idling reduction system. One method corrects the command movement distance when operating the idling reduction system after the feed operation. The other method delays the operation timing of the idling reduction system using a dwell command. Both methods reduce the positional shift compared with the uncorrected system.

View full abstract

A prototype tool tying a clove hitch to bind a wire harness is developed. Wire harnesses installed in airplanes are bound with belt-like strings. The binding process, tying a clove hitch first and then a double overhand knot, is still performed by human workers. To promote the efficiency of the process, simplification of the process with some tools is more practical than automation by robots. We propose a tool with a groove to form a clove hitch by getting a string through it. The bottom shape of the groove is very important to get the belt-like string through the groove without jamming. We formulated its shape based on the theory of surfaces and designed the shape by minimizing change in its curvature and torsion. We made a prototype tool and confirmed its performance with tying experiments. As a result, tying succeeded with a probability of 72%.

View full abstract

The Frobenius norm of the lifted system representation of tracking error dynamics is proposed as a metric for evaluating the tracking performance of discrete-time linear time invariant (LTI) and linear time varying (LTV) controllers. The proposed metric is introduced here in the context of feedforward tracking control of LTI single-input single-output (SISO) nonminimum phase (NMP) systems, though it is more broadly applicable. It is shown that the filtered basis functions (FBF) approach, an LTV tracking control technique studied by the authors in prior work, is the optimal solution to the rank constrained minimization of the proposed metric. Moreover, for the FBF controller, the metric is independent of plant dynamics, which is not the case for most other tracking controllers; it is also independent of the type of basis functions employed in the FBF controller. The effectiveness of the proposed metric as a tracking performance evaluation tool for both LTI and LTV tracking controllers is demonstrated analytically and numerically on LTI plants with different zero locations.

View full abstract

In recent days, high speed machining process is widely used to achieve efficient fabrication process for dies, mold and other mechanical parts. However, modern CNC controllers have problems with dynamic performance which have a great impact on accuracy and productivity. In the high-speed machining process, some factors such as jerk, acceleration, synchronizing of axis and small segment approximates the contours curve, could be a trigger of a break on feed-rate. Therefore, these problems make it difficult to predict instantaneous cutting force in actual manufacturing process.
In this paper, we proposed new feed-rate prediction model in high speed machining process. The proposed model can consider influences of CNC interpolation in the small segment approximation. The simulation results show the proposed model can predict dynamic performance of feed-rate in the high-speed machining process.

View full abstract

Aluminum alloys have been widely used for automobile and aircraft parts. In cutting of aluminum alloy, adhesion sometimes occurs to deteriorate surface finish. The study investigates the adhesion characteristics in drilling of aluminum alloy (A7075) for the cutting parameters. An evaluation approach is presented to characterize adhesion in the cutting tests.
Adhesion on the rake face of the tool is investigated in drillings of a plate for the cutting speed and the feed rate. Much adhesion appears around cutting speeds ranged from 40 to 60 m/min, which is associated with the stress and the temperature on the tool face. The adhesion area expands with the feed rate in accordance with the chip contact length on the rake face. At a low feed rate, an adhesion zone is restricted on a tool tip with the chamfer. Two adhesion zones appear on the tool tip and the chip contact area on the rake face at high feed rates.

View full abstract

In precision motion systems, suppression of high-frequency unknown disturbances is always critical and challenging. This becomes even more difficult when the system output is only available at limited sampling rates, but the frequencies of the disturbances are beyond the Nyquist frequency of the output samplers. Under such scenarios, the un-availability of the intersample outputs precludes most of traditional disturbance attenuation techniques including, for instance, loop shaping, disturbance observer (DOB), and adaptive control. To address this problem, a new adaptive disturbance suppression algorithm is proposed in this paper. Via iterative multirate extended-state estimation and parameter adaptation, the unknown beyond-Nyquist disturbances can be correctly estimated and compensated. Moreover, a special internal model structure is introduced for signals with aliasing frequencies so that the dynamics of the beyond-Nyquist disturbances can be identified more efficiently. Performance of the proposed algorithm is evaluated by simulations on a hard disk drive (HDD) benchmark problem, where the beyond-Nyquist disturbances are effectively suppressed without a priori knowledge of the disturbance models.

View full abstract

To promote the performance of harvesting perpendicularly moving base, a Y-shaped inverted cantilever piezoelectric beam is proposed, which has a tip mass at each limb. When the beam is excited randomly in the vertical direction, the dynamic instability could be activated. In this case there appears energy transferring in frequency domain. Thus the beam takes a largeamplitude oscillation and generates a large output voltage. The prototype was fabricated and the validation experiment was carried out. The results show that the occurrence of dynamic instability induced by excitation could generate large-amplitude and thus result in a large output. The Y-shape harvester is preferred in scavenging perpendicular vibration energy.

View full abstract

It is expected that the S-shape machining test which has proposed to ISO standard can be an effective method of accuracy evaluation for five-axis machining centers. However, there are few research works done for the dynamic error influence on the S-shape machining accuracy. This study investigates the influences of federate, control mode of NC controller and position loop gain of feed drive systems of a five-axis machining center. Actual S-shaped machining motions are carried out to this investigation. As the result, the effect caused by feed rate and machine control function is clarified. Combination of the control mode and position loop gain significantly influence the S-shape machining accuracy. It can be concluded that the proposed S-shape machining test could be used to evaluate the total performance including the control mode of the five-axis machining centers.

View full abstract

A method to predict the static strength of injection-molded products made from Glass Short-fiber Reinforced Thermoplastics (GFRTP) was examined via experiments as well as numerical simulations. Resin-flow analysis was performed to determine fiber orientations through use of the RSC model based on the observations performed using by means of X-ray computed tomography. Molten-resin viscosity was calculated by determining pressure loss in the resin when flowing into the mold. Structural analysis was performed, using the homogenization method, to calculate static mechanical properties of GFRTP products. Determination accuracy of fiber orientations was observed to have been improved upon application of experimentally obtained resin viscosity. Calculation results of structural analysis was also observed to be changed through reflection the fiber state. Further, a difference in calculated stress states was observed via examination of the influence of voids generated in the fiber and matrix.

View full abstract

A probable cause of the breaking phenomenon in the sleeve of cold-chamber die-casting machines was estimated based on observation of cracks and FE simulations. Temperature analysis of the sleeve during the casting process was performed using casting simulation software, and results obtained were mapped onto the FE model for further structural analysis. Many pit-like chips were observed to have been generated on the surface of cooling holes, and a beach mark was observed during progression of cracks on the fracture surface. A tensile stress of the order of 700MPa was measured in the vicinity of cracks, and a stress amplitude of nearly 260 MPa was observed during the die-casting operation performed as part of the FE analysis. As a major finding of this study, cooling hole corrosion and corrosion fatigue caused by stress amplitudes during die-casting were observed to be primary causes of the sleeve cracking phenomenon.

View full abstract

Incremental micro-forming is a novel forming method that can produce miniature shell parts for small batch prototyping with customized design. In lieu of utilizing rigid tools to induce plastic deformation on the sheet metal, a high-speed water jet as an alternative was successfully applied in the incremental microforming process in this study. The influence of the jet trajectory and characteristics on the final geometry of features formed on stainless-steel foil was investigated. The incremental step/pitch of the trajectory and water jet pressure was found to have a dominant effect on the surface quality and depth of the part. Several different paths were designed to create cones, pyramids, polygonal pyramids and multi-level structures. Finite Element (FE) simulations were conducted in ABAQUS to simulate the entire manufacturing process. The numerical simulations were in good agreement with the experimental results in terms of geometry. The logarithmic strain in the radial direction after forming was also measured and compared to FE simulations.

View full abstract

This study deals with the thermal stress cleavage of sapphire wafer with a thickness of 5 mm by CO₂ laser beam irradiation. The influence of the distance from the edge to the initial crack on the cleavage characteristics was investigated experimentally. As results, the waviness on the cleaved surface was quite influenced by the distance from the edge to the cleaved position. When the distance from the edge was 6 mm, the waviness of the cleaved surface was 840 µm, and its value was improved to 162 µm under the distance from the edge of 18 mm. In addition, the laser power required for the cleavage was increased as the distance from the edge was increased. On the other hand, the periodical wavy surface was obtained when the laser power was quite large. The wavy surface was improved by lowering the laser power.

View full abstract

Recent advances in machining technology have led to the creation of non-axisymmetric curved-surface turning (NACS Turning). This process can produce non-axisymmetric components more efficiently than traditional CNC grinding processes. However, NACS-Turning leaves undesired periodic cutting marks on the workpiece surface. This study describes the combination of magnetic abrasive finishing (MAF) and NACS Turning and demonstrates the feasibility of MAF to remove the described cutting marks. The MAF process was applied by adding a specialized MAF tool post to a NACS-Turning machine performing turning of hardened steel components. The result is a single setup capable of both turning and finishing.

View full abstract

Wire saws with fixed diamond abrasive are often used to cut hard and brittle materials owning to the wire saw's narrow kerf, low cutting force, and minimal material waste. Typically, the cutting force changes during the operation since the wire motion direction continuously changes due to reciprocation. Even if the process parameters (e.g., contact length between wire and part, wire tension) are fixed, large variations in the force can lead to wire saw breakage, wafer collapse, and inferior surface roughness. This study addresses this issue by regulating the normal force via automatic adjustment of the wire velocity. Experimental studies are conducted for the wire saw machining of optical glass specimens. The results show the controller can maintain a constant normal force when machining optical glass and significantly decrease wafer surface roughness as compared to the cutting process that uses a constant wire velocity.

View full abstract

Near-field acoustic levitation is a method to levitate an object in a close range by utilizing vibration of the object or the ground surface at a high frequency. In this paper, the frequency dependence of levitation force is numerically and experimentally investigated, which is important for the fundamental understanding of near-field acoustic levitation and the design optimization for non-contact actuators and bearings based on the phenomenon. Computational fluid dynamics simulation is used to study a simplified two-dimensional model. It is found that the levitation force reaches local maximal values at multiple critical frequencies, which is associated with the pressure distribution pattern within the air film. The nonlinear relationship between the levitation force and frequency is verified through the experimental study. The first critical frequency corresponding to the maximum levitation force has been successfully identified in the experiment.

View full abstract

We propose an optical displacement measurement sensor on a machine tool to measure the deflection of a metal-workpiece, which is as precise as a conventional contact-type dial gauge. Our sensor allows automatic and fast measurement by non-contact measurement. We have proposed a multi-directional triangulation system which uses a lens array consisting of four lenses, and scanning exposure method which scans the projection beam on a workpiece during the exposure time of an image sensor. Our techniques improved the accuracy by detecting the reflected light from four angles and scanned positions. In this paper, we show that our proposed method achieves on-machine displacement measurement of metal-workpieces with same accuracy as a dial gauge. In addition, we evaluate the measurement error in the shape of machining traces and show that our sensor measures the electrical discharge machining traces whose reflected direction characteristic is uniformity more precisely than the cutting machining traces.

View full abstract

Commonly used methods for gearbox fault diagnosis involve feature extraction from measured signals to capture its state variation, followed by a fault identification process. These methods are regarded as feature-based process and the extracted features, such as RMS value and kurtosis, are used as input for fault diagnosis. However, fault-related transient impulses, which are embedded in the signals, are lost in feature extraction, leading to reduced diagnosis accuracy. To overcome this shortcoming, the deep residual network with hybrid dilated convolution (ResNet-HDC) is constructed for gearbox fault diagnosis in this paper, which possesses two advantages: 1) deep residual network for deep feature extraction, and 2) hybrid dilated convolution for blurred signal handling. Experimental study performed on a gearbox test rig has shown that the ResNet-HDC is effective for gearbox fault diagnosis.

View full abstract

Specific energy is one of the key parameters for evaluating the performance of grinding processes. This paper presents an approach to estimate specific energy in grinding processes by fusing data from multiple low-cost sensors. Six sensors were adopted to measure total machine current and voltage, grinding motor current and voltage, vibration and acoustic emission. Time and frequency domain analysis are conducted to extract features from raw sensorial data which partially represent the grinding specific energy variation. An artificial neural network (ANN) model was established to fuse the extracted features and determine the mathematical correlation to the specific energy. The approach was experimentally tested on a commercial surface grinding machine. The results of experimental validation indicate that the specific energy for the grinding process under study can be accurately estimated with the developed approach with an average accuracy of 72.94%.

View full abstract

Smart Manufacturing (SM) systems often use the internet for communication. Sensors, platforms, and services that communicate in SM need to securely connect to internet and communicate with one another in a standard language. An increasingly popular language for communication in IoT in general and more specifically for hardware in manufacturing is MTConnect. The goal of this work is to demonstrate an approach for development of a platform compatible with MTConnect that performs high frequency data acquisition from machines, processes collected data, and securely interacts with Internet and web Application Programming Interfaces (APIs) for cloud computing and web interactions. The proposed platform consists of two separate sections, a Local Area Network (LAN) and an Internet Area Network (IAN). The LAN communicates with machines via MTConnect, performs fog computations, and transfers data to the IAN. The IAN receives high frequency data from LAN. The platform facilitates several tasks such as inserting data to a database, fetching data to apps, sending messages and notifications, and communicating with web services such as Amazon Web Services and Particle Clouds.

View full abstract

Fault pattern recognition in complex mechanical systems such as gearbox has always been a great challenge. The performance of a classic fault pattern recognition approach heavily depends on domain expertise and the classifier applied. This paper proposes a deep convolutional neural network-based transfer learning approach that not only entertains adaptive feature extractions, but also requires only a small set of training data. The proposed transfer learning architecture essentially consists of two sequentially connected pieces; first is of a pretrained deep neural network that serves to extract features automatically, the second piece is a neural network aimed for classification which is to be trained using data collected from gearbox experiment. The proposed approach performs gear fault pattern recognition using raw accelerometer data. The achieved accuracy indicates that the proposed approach is not only sensitive and robust in performance, but also has the potential to be applied to other pattern recognition practices.

View full abstract

In this report, we have described the experimental result of the quantity of using the acceleration sensor movement and precision inspection of the posture angle calculation of the XY axis movement. A simple and compact self-localization system could be constructed, and confirmed the accuracy of position estimation by using 3-axis MEMS acceleration sensors and small imaging sensors with CMOS camera.
With the aim of constructing 3D-maps with position information added to images in an infrastructure, we experimentally confirmed that position estimation accuracy could be improved by reducing generated vibration. We also confirmed the usefulness of an image synthesis method that does not need feature points. By adding distance information obtained from self-position estimation to created composite images, we constructed 3D-maps that reproduce actual in-vivo conditions.

View full abstract

This paper presents laser-based people tracking using two networked 64-layer laser scanners (LSs) fixed in an environment. The tilt angles, heights, and relative poses of the two LSs are calibrated based on the laser-scanned data. After the position data of people are extracted from the laser-scanned data using a background subtraction method, poses and sudden behaviours of people, such as dashing and falling down, are estimated based on the interacting-multimodel (IMM) estimator. Experimental results validate the effectiveness of the proposed method.

View full abstract

This paper presents the novel consensus algorithm that drives a group of agents to achieve same behaviors. The algorithm features as flexible and robust design that includes cooperative law and adaptive law. The proposed approach can adapt to different network topology as long as the network has a spanning tree. The cooperative law utilizes local information from available neighbors and drives agents to achieve global consensus. The adaptive law rejects the local disturbances so that the uncertainties do not affect network. Thus, the multi-agent systems arrive locally and globally convergence. The simulation results for different network topology verify proposed flexible consensus algorithm.

View full abstract

Grinding is a machining process commonly used in industry to achieve a desired level of product quality (e.g., surface finish). The condition of the grinding wheel is one of the features of the grinding process that has the largest effect on the quality of the finished product. This research proposes a methodology to predict grinding wheel condition based on surface roughness (Ra). In this methodology, two different regression models based on the multilayer perceptron artificial neural network (MLPANN) and random forest (RndF) machine learning algorithms were first compared based on their R ² test values. Then, a study on the best performing models was conducted to identify a sensor (or combination of sensors) that best predicts tool condition. The results showed that the RndF-based regression model performs marginally better than the MLP-ANN based regression model. It was also found that the tangential component of force plays a significant role in determining Ra.

View full abstract

We have proposed a frame-wise Direction of Arrival (DOA) estimation method with mixed signals observed at a pair of microphones. The proposed method has been confirmed can successfully track the moving sound source in a real time manner under the conditions that the distance between the microphone and the target sound source is within 0.5m and the angular velocity of the target is 110 º/sec or less by simulations.
In this paper, we investigate whether our approach is valid in a real environment. From several experiment results, it has been confirmed that our approach can estimate DOA of the moving sound source in a real environment where the distance of the microphone and the target sound source is within 0.5m and true DOA is from -30º to 30º.

View full abstract

Condition monitoring and fault diagnosis is of significance to improve the safety and reliability of motor system. This paper presents a novel approach for motor fault diagnosis based on multimodal deep learning. Particularly, the case that the motor fault is coupled with the sensor fault has been taken into consideration. One-dimensional convolutional neural network and multi-layer perceptron are employed in the proposed model to extract features from vibration signals and current signals respectively, and a fully connected layer gets the shared representation of various modal. The proposed model has been applied to analyze experimental signals collected from different motor conditions. Results show that the proposed model is effective in diagnosing motor faults and sensor faults for improved induction motor status assessing and maintenance.

View full abstract

The position measurement for heavy-duty machine tools by sequential multilateral method is modeled as a nonlinear redundant (overdetermined) problem, which is generally solved by ordinary least square (OLSQ) method. Due to the influences of repeatability error and distance detection error of laser tracer, the positions retrieved via OLSQ are associated with low stability due to noise and error in the laser tracers data. To reduce the influence of data uncertainties and find a stable solution, Tikhonov regularization is adopted in this paper to mathematically minimize the fluctuation in the calculation results. Simulation is conducted based on Monte Carlo method to prove the concept. The results show that the solution of the proposed method could be more accurate with lower average error and standard deviation compared with that from OLSQ method.

View full abstract

A cross-docking terminal is a transshipment facility in supply chains, where products transported by inbound trucks are unloaded at inbound doors, sorted, and reloaded on outbound trucks at outbound doors. In this study we address the truck-to-door scheduling problem in a multi-door cross-docking terminal where temporary storage is considered. We propose two types of timeindexed formulation for the problem to assign trucks to dock doors and determine their arrival and departure times so that tardiness and earliness as well as unsatisfied demand are minimized. We examine the effectiveness of the proposed formulations by numerical experiment.

View full abstract

Transition from low-mix high volume manufacturing to high-mix low-volume manufacturing has increased importance of job-shop scheduling. Because of non-deterministic polynomial time hardness of job-shop scheduling problem (JSP), approximate optimization based on meta-heuristics have been actively discussed, and methods using simulated annealing (SA) have been proposed. Because SA has a disadvantage that good solutions cannot be obtained efficiently if the initial solution is not given appropriately, methods for solving this problem have been proposed for JSPs aiming at minimizing makespan. In high-mix low-volume manufacturing, it is important to minimize lead times in addition to makespan. This research deals with development of an efficient method using SA for JSPs aiming at minimizing average lead time (ALT). Methods of neighborhood limitation in SA for reducing ALT were developed by focusing on waiting time of operations. It was proven that using the proposed methods in SA with appropriate probability is effective by numerical examples.

View full abstract

This paper proposes an efficient scheduling method for job shop with considering the constraint of the number of operators. An operation of a job can be processed by a predetermined number of operators for its operation. The total number of operators is also given. There are two types of operations: manual operations and automatic ones. Automatic operations do not need human operations except for setup and removal operations. Time horizon is comprised of multiple shifts and overtime work is allowed between shifts. The primary objective of scheduling is to meet due dates of jobs and the secondary objective is to minimize overtime. A genetic algorithm incorporating priority rule is applied for this problem. The effectiveness of the proposed method is examined through numerical experiments. The problem is also formulated by an integer programing model. Some small scale examples are solved optimally and used for comparison with the proposed method.

View full abstract

In recent years, it is required to reduce the amount of waste since there is a growing interest in environmental issues. Hence, it is important to reuse and recycle the used products efficiently. In this paper, we propose a scheduling method by using an evolutionary algorithm in order to find an optimal disassembly schedule including a lot of components of the used products. The relationship between two components in assembly structure can be represented by a tree structure. A whole disassembly process of a product is represented by a combination of tree structures. Each tree structure is exchanged by genetic operations such as crossover and mutation. We carry out a lot of computational experiments though the developed scheduling system. The experimental results demonstrate that the new method is superior to the previous method from the viewpoint of the minimization of the whole disassembly processing times.

View full abstract

To realize the criteria of the sustainable manufacturing industry and the flexibility corresponding to the diverse needs of users, we first put forward the concept of a manufacturing facility that consists of a core platform and add-on modules (P-A manufacturing facility). In a prior study, we established a circulation market model that takes P-A manufacturing facilities as objective products and constructed a multi-agent simulation system from the perspective of long-term use and life cycle assessment theory. Here, more behavioral agents that may lead to unknown effects on circulation are added to our model. Different societal scenarios are edited for simulation, and differences generated from different scenarios are compared and evaluated. This proposed approach using multiagent simulation system is expected to help to select some optimal societal scenarios and to provide feedback relevant to application in real societal situations.

View full abstract

Key performance indicators (KPI) play an important role in the manufacturing system. In practices, it is often needed to control certain KPI value to be at a specified level. However, the relationship between system inputs and KPIs and among KPIs are very complex. In this paper, we focus on the KPI control for optimizing the system inputs to reach desired outputs values based on an approximated predictive model for KPIs. The model can achieve high flexibility for analysis, thus providing accurate KPI control. We demonstrate the effectiveness of the KPI control through simulation studies.

View full abstract

With the advancement of Information and Communication Technology (ICT), smart manufacturing system has been improved extremely in the world, and the information acquired from smart devices in manufacturing equipment is used for improving manufacturing efficiency and integrity. In our group, a value co-creative smart factory platform has been constructed, and also the production scheduling algorithm in consideration of users’ preferences and machines’ condition has been developed. In our smart factory platform, the condition of the simulated products is acquired using RF-ID system, and then the schedule of each production machine is adjusted using a multi-agent system with machine conditions, users’ preference, desired due date and prices using the developed interactive system. In the production schedules, a scheduling method introducing an auction mechanism has been developed. In this proceeding, the current development of smart factory platform with interactive system and the production scheduling method involving the auction mechanisms are discussed.

View full abstract

Recently, the demand of LNG is increasing and KHI is one of the world’s leading manufactures. We must satisfy the customer’s request: high quality, fast delivery, and low cost. In order to maximize efficiency of the production line, it is important to standardize work and create “standard work combination tables” for each worker. However, in the case of one-of-a-kind products such as an LNG storage tank, it is difficult to prepare the optimal standard work combination table through iteration of trial-and-error. A discrete event simulator is a useful tool by which the performance of the production line can be virtually checked. The authors developed a method to create a standard work combination table for a flow shop line including multiple workers using discrete event simulation. As a result, we created an efficient standard work combination table in a short time without actual factory operation. Thus, we drew out the capacity of the production line to the maximum extent from the start of production.

View full abstract

Supply chain resilience has been receiving an attention for the risk management. In this paper, we propose a rescheduling strategy in order to recover the productivity of supply chain after facing unexpected disruptions in manufacturing companies. When suppliers suspend manufacturing processes due to disruptions and products cannot satisfy a given due date, the suppliers reduce the delays and penalty charges for the delays by contracting out the delayed products to external suppliers and restore production capacity to the state before the disruption occurs. This research proposes priority rules using dispatching rules in order to select suitable products for outsourcing from the delayed ones. We carried out computational experiments through the developed simulation system of supply chains. The experimental results demonstrate that outsourcing is effective for the recovery of productivity in supply chains.

View full abstract

Nowadays, manufacturing enterprises need to plan their strategies considering environment as well as economy. Therefore, closed-loop supply chains have been proposed. Closed-loop supply chains consist of forward supply chains and reverse supply chains in which remanufacturers reuse and recycle the used products. Previous studies proposed a method for remanufacturers to collect many products by estimating suitable product value at the collection time.
This study proposes a method to collect appropriate rechargeable batteries by estimating remaining value at the collection time. The variation of internal resistance of batteries is calculated by using a function of the current internal resistance and the period of use. When we measure internal resistance values, they change depending on charging or discharging. Thus we propose a method to estimate internal resistance by using the least square method. We carry out the computational simulation of closed-loop supply chains to demonstrate the effectiveness of the proposed method.

View full abstract

One of the key issues to realize green manufacturing will be to build a practical method of predicting power/energy consumption at each individual manufacturing equipment. We have proposed a regression-based model to predict an in-process power consumption in end milling on a A5052 alloy. This paper extends our previous work to a face milling process of some materials. The capability of our conventional model in face milling is demonstrated by introducing a time decomposition approach to precisely estimate an in-process power consumption. In the face milling process, a contact area between the tool and workpiece tends to rapidly vary. Thus, we consider to divide a time period during cutting into three characteristic domains in order to precisely predict electric power consumption over the entire cutting process. Some experimental results demonstrate the performance of our proposed model to predict power consumption in face milling.

View full abstract

In order to achieve the concept of delight design, a specific technology for mass customization producing must be developed. Biological structures have evolved to fit their purpose, and their adaptability and robustness are attractive to delight design and mass customization. Machine design which reflects the feature of biological structures is likely to lead the innovative realization of ultra-lightweight and high-strength structure. However, the structure that merely imitate biological form is not mechanically effective. In order to manufacture products which are inspired by biological form and also have all required properties, we need a new design theory and manufacturing system. We have developed a new design and production technology called ‘Bio Innovative Design Technology', which consists of; 1) Design database with forms in biology, 2) Design optimization techniques, 3) Development of braiding technology.
This article reports an outline of Bio Innovative Design Technology, and then report a fundamental manufacturing method of solid geometric preforms by braiding technology which is suitable for carbon fiber reinforced plastic.

View full abstract

On the tubular fabric produced by the circular knitting machine, the following defects sometimes appear such as vertical lines, horizontal lines, contamination, and holes. These defects are caused by malfunctions of the knitting machine or the poor quality of yarn used. The proposed detecting systems for these defects by other researchers have not been implemented in the industry because they appear to require excessive operating time. Therefore, the aim of this study is to create a simple analyzing system for vertical line defects because they cause massive wastes of fabric. By using prototype software, vertical line defects could be recognized by evaluating moving averages of averaged brightness of vertical successive pixels obtained from the captured fabric image, and the functions of de-skew and determination of number of pixels per loop width in the software will be able to be adapted to the future system on circular knitting machines.

View full abstract

A method to design the two-dimensional shapes of patterns of two piece brassiere cup is proposed when its target three dimensional shape is given. A brassiere cup consists of several patterns and their shapes are designed by repeatedly making a paper cup model and checking its three-dimensional shape. For improvement of design efficiency of brassieres, such trial and error must be reduced. As a cup model for check is made of paper not cloth, it is assumed that the surface of the model is composed of several developable surfaces. When the two-dimensional shape of an edge of a paper pattern and the target three dimensional shape of the edge are given, the surface of the pattern is uniquely determined. Then, the pattern with two edges can be designed by optimizing both shapes of edges so that determined developable surfaces from individual target edge shapes coincide with each other. It was experimentally verified that the edges of a designed pattern using the proposed method formed the target shapes.

View full abstract

This paper studies the modeling and control of a class of fully actuated, thrust vectoring unmanned aerial vehicles (UAVs) for aerial locomotion and manipulation. In this paper, we develop a novel fully actuated vector thrust UAV with all propellers able to tilt about two perpendicular axes, so that the thrust force generated by each propeller is a fully controllable vector in 3D space. The dynamic model of a UAV with arbitrary number of such vectorized propellers is constructed, and the control algorithm for position/orientation tracking is developed to generate the desired 6-dimensional body force wrench. To resolve the redundancy in actuation, a thrust force optimization problem minimizing power consumption while achieving the desired body force wrench is formulated. The optimization problem has the thrust vectors as variables, and is shown to be convex with linear equality constraints, so that the global minima can be easily derived. Various simulation results are presented to verify the proposed solution.

View full abstract

Traffic congestion in the city is one of urgent problems to solve in order to maintain stable civic life and keep citizens ’daily behavior safely. Any technologies discussed and mathematically formulated are necessary for the realization of Intelligent Transportation Systems (ITS) in any modern urban area. Platooning, the linking of cadre vehicles, is important to control traffic flow in city crowed road, with which the method of traffic control without congestion is considered in this paper.
A series of works such as harvesting and transporting in a farm is one of such works with so care as not to damage the harvest in order to maintain the value of harvests. We are developing an autonomous cart for gathering a harvest with the bed to be controlled to keep in horizontally level, in order to avoid harvests gathered in particular area of the bed and avoid to be damaged in harvests. It is proposed a method of control of maintaining the horizontal level of the bed of the cart with air cylinder suspension systems.

View full abstract

Along with the growth of elderly population, robot walking helpers play an important role in assisting their activities in daily lives. The passive type of robot walking helper moves only by user-applied forces with controlled brakes used to steer the walker. It can safely assist the elderly in walking. Since the environment may be with moving obstacles, it is important for the passive walking helper to move away from these moving obstacles. In this paper, we propose using the optimization method for trajectory planning and control of the passive robot walking helper. The collision-free trajectory and the brake torque can be efficiently derived for safely guiding the user to the desired position. Simulations are performed to demonstrate the efficiency of the proposed approach.

View full abstract

Working robots are expected to be applied to work in various environments with complicated 3D structure such as plants and bridges. Such complicated 3D structures include multiple planar surfaces such as floor, wall, and ceiling. Robots are required to move on such environment composed of planar structures. A locomotion working robot ASTERISK with 6 limbs has been developed for moving and working on complicated environments. We developed a new type of ASTERISK whose limb has an electromagnet at the tip. Then it has locomotion ability in anti-gravity situation. We designed a transfer motion between planar surfaces for the ASTERISK. The transfer motion is verified by computer simulation. Furthermore, the motion is implemented to actual ASTERISK. Experimental results show the feasibility of the transfer motion of ASTERISK.

View full abstract

Super smart society is one of the recent concepts, which are gaining a lot of attention in Japan. It includes around 12 different systems, which are connected between each other, and at the same time each system consists of several subsystems. Modeling this kind of systems can be tricky task. It happens because of the intersystem connections. Also at the end of each system there is society or citizens, which makes modeling more complicated. Having this in mind main objective of this research is to develop multiscale modeling approach for social systems. Multiscale modeling was chosen because of its ability to break down case into several levels of abstraction, which might be very handy when we are trying to model concepts such as Society 5.0. Our previous papers, for example [1], were concentrated on model of energy sector and on triple scale model. In this paper main concern will be providing an introduction about theory of multiscale modeling for social systems. Computational experiments demonstrated that application of multiscale modeling provides necessary attributes for handling complex systems such as Society 5.0.

View full abstract