I made the program with the image of shy child running away and hiding in the shade of trees and pillars when you get close. When you approach to the robot, it hides behind the back of a plate and peeks out. Also if you leave the robot alone, the robot comes close to people in reverse which makes it lovely.
I have developed a robot control system by centroid movement using a body sway meter. In the system,robot moves in all directions according to gravity center sway of a user. The system presents information acoustically and estimates movement of the center of gravity regardless of the initial centroid position of a user, which enable the user to control robot intuitively without requiring knowledge of robotics nor training. The system is also available remotely in combination with virtual vision.
I developed a system for estimating a weight of contents in a bucket of a hydraulic excavator from a camera image. A kind of contents and a volume in the bucket were estimated separately by CNN (Convolutional Neural Network) from the camera image. Then, a weight of contents in the bucket is estimated from estimated results of contents and the volume.
The system uses only 3D-URG which is mounted on the robot, to determine the human and other various objects. Point cloud data acquired with the 3D-URG are converted to image data by projecting. It discriminates an object by applying the Deep Learning in the images.
I made a remote control system by using a racing wheel for video game. Host computer and the robot share a controller and laser scanner information by TCP. The robot sends pictures to host computer by UDP. Stepping on the accelerator, the robot smoothly speeds up. The robot is searching for an obstacle by using a laser scanner. When it finds an obstacle, it is moving backward.