A model-free learning algorithm called Predictive Sequence Learning (PSL) is presented and evaluated in a robot Learning from Demonstration (LFD) setting. PSL is inspired by several functional models of the brain. It constructs sequences of predictable sensory-motor patterns, without relying on predefined higher-level concepts. The algorithm is demonstrated on a Khepera II robot in four differe...

In this paper, we study the online path planning for khepera II mobile robot in an unknown environment. The well known heuristic A* algorithm is implemented to make the mobile robot navigate through static obstacles and find the shortest path from an initial position to a target position by avoiding the obstacles. The proposed path finding strategy is designed in a grid-map form of an unknown e...

In this paper we present a biologically inspired neural architecture for visual perception based on anticipation. The main goal of this work is to demonstrate, that anticipation is a central key to improve the perception performance of technical systems. The presented approach is able to increase the robustness of the perception process against noise or sensory dropouts. We demonstrate these pe...

A distributed mobile robot software application infrastructure is developed, improving integration and leverage between projects in a research environment. The resulting design includes a three layer CORBA based, service broker application architecture. A reference implementation and tests on B21r, LEGO Mindstorm and Khepera robots demonstrate the feasibility of the design.

This paper describes the technical implementation of the interface between the Matlab RealTime WorkShop c © and the Khepera c © robot operating system. Moreover, it shows how this tool is effective to quickly design, simulate and test controllers for autonomous robot. Finally, we discuss the feasibility of a object detection method based on acquired patterns of infrared sensors data.

This paper describes the testing tool ”ExmoR” that allows the user to experiment with different control algorithms as map building and a routine to explore the environment. The intention is to give the user the possibility of easy testing the various control algorithms for the Khepera robot without any knowledge about the robot language, the control architecture or other robot details.

We propose to study the efficiency of our mobile robot control architecture — the so-called "trend" architecture, derived from Brooks' subsumption architecture — applied to the locomotion of a group of miniature Khepera robots. In our experimental setup, each robot is equipped with a sophisticated auditory system, allowing it to communicate its position to its near neighbours by means of short ...

In this paper we describe a neural network for reactive and adaptive robot navigation. The network is based on a model of classical and operant conditioning first proposed by Grossberg [3]. The network has been successfully implemented on the real Khepera robot. This work shows the potential of applying self-organizing neural networks to the area of intelligent robotics.

We implemented variations on the Self-Organizing Distinctive State Abstraction model for robot state-discretization and task-learning in a Khepera robot. We conclude that hill-climbing as implementd in [5] is detrimental in embodied robots. Because the ultimate goal of developmental systems is intelligent, embodied robots, we believe that implementation in physical robots is an important test o...

In this paper, we compare the performance of hierarchical GP methods (Automatically Defined Functions, Module Acquisition, Adaptive Representation through Learning) with the canonical GP Implementation and with a linear genome GP system in the domain of evolving robotic controllers for a simulated Khepera miniature robot. We successfully evolve robotic controllers to accomplish obstacle avoidan...