Industrial robots are expected to share the same workspace with human workers and work in cooperation with humans to improve the productivity and maintain the quality of products. In this situation, the worker’s safety and work-time efficiency must be enhanced simultaneously. In this paper, we extend a task scheduling system proposed in the previous work by installing an online trajectory generation system. On the basis of the probabilistic prediction of the worker’s motion and the receding horizon scheme for the trajectory planning, the proposed motion planning system calculates an optimal trajectory that realizes collision avoidance and the reduction of waste time simultaneously. Moreover, the proposed system plans the robot’s trajectory adaptively based on updated predictions and its uncertainty to deal not only with the regular behavior of workers but also with their irregular behavior. We apply the proposed system to an assembly process where a two-link planar manipulator supports a worker by delivering parts and tools. After implementing the proposed system, we experimentally evaluate the effectiveness of the adaptive motion planning system.

Automation of the bin picking task with robots entails the key step of pose estimation, which identifies and locates objects so that the robot can pick and manipulate the object in an accurate and reliable way. This paper proposes a novel point pair feature-based descriptor named Boundary-to-Boundary-using-Tangent-Line (B2B-TL) to estimate the pose of industrial parts including some parts whose point clouds lack key details, for example, the point cloud of the ridges of a part. The proposed descriptor utilizes the 3D point cloud data and 2D image data of the scene simultaneously, and the 2D image data could compensate the missing key details of the point cloud. Based on the descriptor B2B-TL, Multiple Edge Appearance Models (MEAM), a method using multiple models to describe the target object, is proposed to increase the recognition rate and reduce the computation time. A novel pipeline of an online computation process is presented to take advantage of B2B-TL and MEAM. Our algorithm is evaluated against synthetic and real scenes and implemented in a bin picking system. The experimental results show that our method is sufficiently accurate for a robot to grasp industrial parts and is fast enough to be used in a real factory environment.

We propose a CNN based visual servoing scheme for precise positioning of an eye-to-hand manipulator in which the control input of a robot is calculated directly from images by a neural network. In this paper, we propose Difference of Encoded Features driven Interaction matrix Network (DEFINet), a new convolutional neural network (CNN), for eye-to-hand visual servoing. DEFINet estimates a relative pose between desired and current end-effector from desired and current images captured by an eye-to-hand camera. DEFINet includes two branches of the same CNN that share weights and encode target and current images, which is inspired by the architecture of Siamese network. Regression of the relative pose from the difference of the encoded target and current image features leads to a high positioning accuracy of visual servoing using DEFINet. The training dataset is generated from sample data collected by operating a manipulator randomly in task space. The performance of the proposed visual servoing is evaluated through numerical simulation and experiments using a six-DOF industrial manipulator in a real environment. Both simulation and experimental results show the effectiveness of the proposed method.

Cable guiding mechanisms (CGMs) directly influences the reconfigurability of cable-driven parallel robots (CDPRs). But due to the complicated kinematic model of pulley-based CGMs, the velocity and acceleration mappings from the moving platform (MP) to the cables were unknown, the continuity of cable velocity, acceleration and tension were unguaranteed, and the calibration pose-search method was based on inaccurate CGM models. In this paper, we establish an analytic and compact model for CGM. Based on this model, the velocity and acceleration mapping from the MP to the cables are derived. The continuities of cable trajectory and tension are proved. The exact identification Jacobian and its derivative are also formulated, leading to the increased fidelity of pose-search methods. The proposed method and formulations are verified by simulations and experiments on a redundant cable-driven parallel robot.

This article introduces several types of robots having physical interactions among a human/humans and a robot/robots, which we have developed so far. MR Helper (mobile robot helper) is a mobile robot with dual manipulators, which could manipulate an object in cooperation with a human. DR Helpers (distributed robot helpers) are multiple small mobile robots which transport an object in cooperation with a human/humans. MS DanceR (mobile smart dance robot) could dance the waltz as a female dance partner with a male dancer. These examples inspire possible applications of the human-robot interaction in near future

Future robots are expected to share the same workspace with humans and work in cooperation with them to improve productivity and maintain the quality of products. Considering this situation, we have developed a novel assembly task co-worker robot to support workers in their task by delivering the parts and tools to workers. Although such systems have improved work efficiency by predicting human's motion patterns, it is necessary to collect worker's data in advance and regenerate its model whenever the worker is changed. In this letter, we extend the previous system by installing an online learning algorithm and create a worker-dependent model without collecting data in advance. Trajectory prediction with high precision can be realized because of the worker-dependent model and effective utilization of the regularity of the worker's behavior. An adaptive task scheduling system based on the predicted result of the worker's behavior is proposed for improving work efficiency. Implementing the proposed algorithm, we evaluate the effectiveness of the task scheduling system by experiment.