Kenichi Kitahama

Image-based object recognition and dexterous hand/arm motion planning using RRTs for grasping in cluttered scene

This paper describes a new approach to realize grasping of an unknown object by a dexterous hand robot among freely placed multiple unknown objects even if they are occluded each other. This technology consists of major three functions: 1) image-based 3D reconstruction and separation of multiple objects using graph-cut theory, 2) recognition of the shape, size and orientation of each object as an object with primitive shape, i.e. either of box, cylinder or sphere, 3) multigoal and multicriteria path planning for 7 DOF arm with 13 DOF dexterous hand using enhanced RRTs (rapidly exploring random trees) algorithm to cope with multiple possible grasping strategy according to the object shape and orientation.

Vision-based scene representation for 3D interaction of service robots

In this paper, we present a new approach for scene understanding by an autonomous robot, using imaging sensors only. Our goal is to build in real-time a 3D representation of the scene which is suitable for further 3D interaction. We assume the scene is composed of a mix of known and unknown objects. We present a strategy to automatically build online a consistent 3D representation from a set computer vision algorithms results. We especially focus on 2 important aspects: the constraints and choices of each parts of the system, as well as the global layout of these parts, and the final step where all information are merged together to build a unique 3D representation. Results on real scene and system performance are finally showed to validate the proposed approach

Measurement method of normalized cornering stiffness

This paper presents a measurement method of the normalized equivalent cornering stiffness (normalized Cp), which dominates a vehicles steering responses. First, the method measures rear normalized Cp from the velocity at which the sign of vehicle slip angle is reversed in steady-state turning. The method is accurate because the velocity is independent of the errors of measuring instruments. Second, the method measures front normalized Cp from yaw velocity gain. Last, the measured normalized Cps of various vehicles are plotted on a distribution map, which indicates steering responses of each vehicle.