Search

In the trash-collection challenge of the Nakanoshima Robot Challenge, an autonomous robot must collect trash (bottles, cans, and bentos) scattered in a defined area within a time limit. A method for collecting the trash is to use machine learning to recognize the objects, move to the target location, and grasp the objects. An autonomous robot can achieve the target position and posture by rotating on the spot at the starting point, moving in a straight line, and rotating on the spot at the destination, but the rotation requires stopping and starting. To achieve faster movement, we implemented a smooth movement approach without sequential stops using a spline curve. When using the training data previously generated by the authors in their laboratory for object recognition, the robot could not correctly recognize objects in the environment of the robot competition, where strong sunlight shines through glass, because of the varying brightness and darkness. To solve this problem, we added our newly generated training data to YOLO, an image-recognition algorithm based on deep learning, and performed machine learning to achieve object recognition under various conditions. [ABSTRACT FROM AUTHOR]

This paper proposes a method for the recognition of the opened/closed states of automatic sliding glass doors to allow for automatic robot-controlled movement from outdoors to indoors and vice versa by a robot. The proposed method uses an RGB-D camera as a sensor for extraction of the automatic sliding glass doors region and image recognition to determine whether the door is opened or closed. The RGB-D camera measures the distance between the opened or moving door frames, thereby facilitating outdoor to indoor movement and vice versa. Several automatic sliding glass doors under different experimental conditions are experimentally investigated to demonstrate the effectiveness of the proposed method. [ABSTRACT FROM AUTHOR]

The development of a chopstick-like two-fingered micromanipulator based on a hybrid mechanism is presented. The microhand consists of two 3-prismatic--revolute--spherical (PRS) parallel modules connected serially in a mirror image style. Each module has a long glass pipette as an end effector. The development process consists of three phases. In the first phase, analysis and mathematical modeling, a novel solution of the inverse kinematics problem (IKP) of a 3-revolute--prismatic--spherical (RPS) parallel module, is derived and applied with proper modification to the case of 3-PRS of the proposed mechanism. The solution is extended to the two-fingered hybrid mechanism of the microhand. In the optimization and design phase, the optimization of the chosen design parameters of a theoretical 3-PRS parallel module is carried out using two approaches: discretization method and genetic algorithms. Based on the optimal design parameters, a CAD model of the 3-PRS finger module is built, and a complementary optimization step using the ANSYS Workbench program is carried out to determine suitable characteristics of the pin flexure hinge. Finally, the total CAD model of the two-fingered hand is built. In the realization and implementation phase, the description of the hardware system of the two-fingered microhand prototype is presented. The program description, calibration method, practical Jacobian matrices, practical workspace, and error analysis of the prototype are discussed. Index Terms--Genetic and evolutionary algorithms (GEAs), hybrid-structure micromanipulators, inverse kinematics problem (IKP) analysis, 3-revolute--prismatic--spherical (RPS) and 3-prismatic--revolute--spherical (PRS) parallel mechanisms.

A method to detect minute flaws on metal parts is proposed to remove defective parts before assembling in a factory. The input grayscale images of metal parts are directly used to recognize flaws without any image conversion to shorten recognition time. The recognition problem to find flaws and detect their position on the metal parts is converted to another problem that searches for the maximum peak and the variables producing the peak. Then, the recognition problem can be treated as an optimization problem, and this conversion allows us to utilize high genetic algorithm performances in optimization. The effectiveness and problems of the proposed method are studied on the standing points of recognition speed and quantitative recognition ability. Based on the analysis, we furthermore improved our system to increase the flaw detection rate; the lighting direction was changed to find the best lighting condition that can emphasize the contrast between the metal surface and the flaw by using the reflection character of the hairline on the metal, which is resulted by a polishing process. Index Terms--Genetic algorithm (GA), metal surface inspection, model-based matching.