Kohki Kikuchi

Navigation strategies referring to insect homing in flying robots

Sensor-based navigation is one of the most important researches for mobile robots domain and then there are many research works for sensor-based navigation. We propose the insects inspired landmark navigation to pinpoint the target position in this paper. There exist several research works such as snapshot model and average landmark model which request the absolute bearing information (compass) and visual one. Whereas our method called relative angle model (RAM) need only visual information in order to extract relative positions of 3 landmarks. By simulation and experiment using 3D blimp-type flying robot, we find that the pinpoint 3D navigation works out well by RAM.

Development on face robot for real facial expressions

The goal of this study is to investigate the intelligence necessary for a machine such as a human-friendly robot working in an environment where humans and robots coexist. It seems to be required to manufacture a human-friendly robot that can communicate with human beings as a test bed or platform. Since face and facial expressions are crucial factors for communication, we have been developing a face robot which has a human-like face and can express facial expressions similar to human being. The important factors of the face robot for coexistence and communication with humans are: 1. human size and compactness; and 2. ease of control. While we have used air cylinders with pressurized air for the first-generation of face robot, to accomplish these two factors, we decided to use a SMA (shaped memory alloy) actuator driven by electricity and built the second-generation of face robot. Although we succeeded in making a human size face robot, it came to clear that the SMA does not have enough power and speed. To overcome these issues, we developed an air-tube-actuator driven by pressurized air and manufactured the third-generation of face robot. In the paper, we show how to build the third-generation of face robot and its basic ability.

A study on functional characteristics of robotic system with morphology and intelligence

Robotic behavior emerges from not only intelligence but also the relationship between morphology, intelligence and the environment. We investigate and analyze the relationship between morphology and intelligence acquired from different tasks, and the characteristic differences of functions emerges from this relationship. Robotic morphology and intelligence are encoded using genetic programming: GP operations, and robot dynamics are handled using a two-dimensional dynamics simulation. As the robotic morphology, we utilize position-changeable visual sensors through evolution by GP operations, and as the robotic intelligence, we use a decision tree which is able to evolve by GP operations. A task of the robot is to keep a certain distance from an object. The simulation results clarify the importance of harmony and/or balance between morphology and intelligence, and show that a new function emerges as the result of harmony and/or balance between morphology, intelligence and a task.

A methodology to investigate robotic intelligence

Abstract -To pursue intelligence is our research purpose. By considering knowledge from related research fields such as psychology, biology, cognitive science, physiology, etc., we have managed to take into account three factors, i.e. (i) to have a physical body to be able to interact with the environment, (ii) to implement a hierarchical structure for intelligence and (iii) not to prepare modules expressing functional similarities a priori. By applying harmony theory - a mathematical model produced in the field of cognitive science - we succeeded in implementing these three factors and creating a hierarchical structure for robot intelligence. The greates asset of this research is that we do not have to prepare functionally isolated modules a priori. The Harmony Theory Network (HTN) automatically organizes the robot behavior in a hierarchical manner as robot intelligence in relation to the environment. That is to say, we find that (i) differences in the environment influence the structure of the HTN, (ii) the structure of the HTN acknowledges differences in sensor tasks automatically, (iii) the HTN organizes the connections for sensory and motor processing in its lower layer, and coordinates the well-organized robot behavior in its higher layer by combining the sensory-motor relation obtained by the lower layer, and (iv) it is possible to interpret the meaning of the HTN structure afterwards in relation to the robot behavior.

Neural network application to capture control in aqua-robot manipulator

This paper deals with implementation of a neural network (NN) algorithm to an aqua-robot manipulator control to capture a floating object in water, and an evaluation of the NN control performance in comparison with that of the conventional PD feedback control. The nonlinear dynamics of a floating object in water, caused by approaching motion of the aqua-robot manipulator to the object, is learned in a three-layered neural network and the neural network is implemented in the control system of the aqua-robot manipulator in parallel to a conventional PD feedback controller. The performance of the neural network capture control is experimentally evaluated and found essentially high in terms of the capture time needed and capture-success ratio.