This paper is concerned with a control design method for multi-fingered robot hands, whose degrees of freedom (DOF) are not enough to control all of the components of the object motion and the internal force. In the control design of the multi-fingered robot hands, controllers are usually designed for the variables with respect to the object motion and the internal force, under the assumption that the systems have enough DOF to control all the components of them. However, if we consider the systems with fewer DOF, although the control variables directly linked to the components of the object motion and the internal force will be preferable for the grasping/manipulation tasks, control design methodology to meet such requirements has not been established yet. In this paper, we consider the systems whose DOF are not enough to control all of the components of the object motion and the internal force, and propose a control design method whose control variables are directly linked to the components of the object motion and the internal force. The control variables can be specified as functions of the object motion and the internal force directly, and a linearizing compensator for the variables is derived. A numerical example of a robot hand with interlocking joints is shown to prove effectiveness of the proposed method.

View full abstract

Sensors of robots that act in unstructured environment sometimes do not provide complete observation, due to occlusion or limitation of sensing range. This paper presents a motion generation method for robot with multiple sensors with limited sensing ranges. The proposed method introduce extension of the action-observation mapping to outside of the sensing range of a sensor, based on the diffusion-based learning of Jacobian matrices between control input and observation variable. Multiple observation spaces can be integrated by finding correspondence between the virtual observation spaces. When a target observation variable is given to the robot, it can generate a motion from an observation space toward the target with another observation space using the extended observation space. The proposed framework is verified by two robot tasks, reaching motion toward the floor with a manipulator and navigation of mobile robot around the wall. In both cases, observation space by camera with limited view was extended and appropriate motion trajectories were obtained.

View full abstract

Consider a special district (group) which is composed of multiple companies (agents), and where each agent responds to an energy demand and has a CO₂ emission allowance imposed. A distributed energy management system (DEMS) optimizes energy consumption of a group through energy trading in the group. In this paper, we extended the energy distribution decision and optimal planning problem in DEMSs from a single period problem to a multiple periods one. The extension enabled us to consider more realistic constraints such as demand patterns, the start-up cost, and minimum running/outage times of equipment. At first, we extended the market-oriented programming (MOP) method for deciding energy distribution to the multiple periods problem. The bidding strategy of each agent is formulated by a 0-1 mixed non-linear programming problem. Secondly, we proposed decomposing the problem into a set of single period problems in order to solve it faster. In order to decompose the problem, we proposed a CO₂ emission allowance distribution method, called an EP method. We confirmed that the proposed method was able to produce solutions whose group costs were close to lower-bound group costs by computational experiments. In addition, we verified that reduction in computational time was achieved without losing the quality of solutions by using the EP method.

View full abstract

A new layered image processing system is proposed for robots under motion and is evaluated for its performance by implementing it onto a small-size humanoid robot. The proposed system is a layered system in both resolution and processing method. As such, the system can employ appropriate processing methods for layers with different resolutions and, as a result, can avoid unnecessary processing of images corrupted by robot motions. It can improve the image processing speed of floor images because of this and because the image processing is performed simultaneously with robot motions. The proposed method and the results of verification experiments are discussed here.

View full abstract

In the field of signal processing, AD/DA converter takes an important role for data compressing and circuits design. This paper proposes new design method of AD/DA system with pre-filter and post-filters for achieving small quantization noise effect. Both filters are designed to minimize noise and distortion toward original signal. The effectiveness is illustrated by numerical example.

View full abstract