Narongdech Keeratipranon

Manifold learning for robot navigation.

In this paper we introduce methods to build a SOM that can be used as an isometric map for mobile robots. That is, given a dataset of sensor readings collected at points uniformly distributed with respect to the ground, we wish to build a SOM whose neurons (prototype vectors in sensor space) correspond to points uniformly distributed on the ground. Manifold learning techniques have already been used for dimensionality reduction of sensor space in navigation systems. Our focus is on the isometric property of the SOM. For reliable path-planning and information sharing between several robots, it is desirable that the robots build an internal representation of the sensor manifold, a map, that is isometric with the environment. We show experimentally that standard Non-Linear Dimensionality Reduction (NLDR) algorithms do not provide isometric maps for range data and bearing data. However, the auxiliary low dimensional manifolds created can be used to improve the distribution of the neurons of a SOM (that is, make the neurons more evenly distributed with respect to the ground). We also describe a method to create an isometric map from a sensor readings collected along a polygonal line random walk.

A Direct Localization Method Using only the Bearings Extracted from Two Panoramic Views Along a Linear Trajectory

To operate successfully in any environment, mobile robots must be able to localize themselves accurately. In this paper, we describe a direct method (in the sense it does not use an iterative search) based on vision for localizing a mobile robot in an environment with only two observations along a linear trajectory. We only assume that the robot can visually identify landmarks and measure their bearings. Contrary to other existing approaches to landmark based navigation, we do not require any other sensors (like range sensors or wheel encoders) or the prior knowledge of relative distances between the landmarks. Given its low cost, the range of potential applications of our localization system is very wide. In particular, this system is ideally suited for domestic robots such as autonomous lawn-mowers and vacuum cleaners.

Bearing-only Simultaneous Localization and Mapping for Vision-Based Mobile Robots

To navigate successfully, a mobile robot must be able to estimate the spatial relationships of the objects of interest accurately. A SLAM (Simultaneous Localization and Mapping) system employs its sensor data to build incrementally a map of an unknown environment and to localize itself in the map simultaneously. Thanks to recent advances in computer vision and cheaper cameras, vision sensors have become popular to solve SLAM problems. The proposed bearing-only SLAM system requires only a single camera which is simple and affordable for the navigation of domestic robots such as autonomous lawn-mowers and vacuum cleaners. Existing approaches to bearing-only SLAM require the readings from an odometer to estimate the robot locations prior to landmark initialization. This chapter presents a new 2-dimensional bearing-only SLAM system that relies only on the bearing measurements from a single camera. Our proposed system does not require any other sensors like range sensors or wheel encoders. The trade-off is that it requires the robot to be able to move in a straight line for a short while to initialize the landmarks. Once the landmark positions are estimated, localization becomes easy. The induced map created by our method is only determined up to a scale factor as only bearing information is used (no range or odometry information). All the object coordinates in the map multiplied by a scale factor would not change the bearing values.

Bearing similarity measures for self-organizing feature maps

The neural representation of space in rats has inspired many navigation systems for robots. In particular, Self-Organizing (Feature) Maps (SOM) are often used to give a sense of location to robots by mapping sensor information to a low-dimensional grid. For example, a robot equipped with a panoramic camera can build a 2D SOM from vectors of landmark bearings. If there are four landmarks in the robot’s environment, then the 2D SOM is embedded in a 2D manifold lying in a 4D space. In general, the set of observable sensor vectors form a low-dimensional Riemannian manifold in a high-dimensional space. In a landmark bearing sensor space, the manifold can have a large curvature in some regions (when the robot is near a landmark for example), making the Eulidian distance a very poor approximation of the Riemannian metric. In this paper, we present and compare three methods for measuring the similarity between vectors of landmark bearings. We also discuss a method to equip SOM with a good approximation of the Riemannian metric. Although we illustrate the techniques with a landmark bearing problem, our approach is applicable to other types of data sets.

An improved probability density function for representing landmark positions in bearing-only SLAM systems

To navigate successfully, a mobile robot must be able to estimate the spatial relationships of the objects of interest in its environment accurately. The main advantage of a bearing-only Simultaneous Localization and Mapping (SLAM) system is that it requires only a cheap vision sensor to enable a mobile robot to gain knowledge of its environment and navigate. In this paper, we focus on the representation of the spatial uncertainty of landmarks caused by sensor noise. We follow a principled approach for computing the Probability Density Functions (PDFs) of landmark positions when an initial observation is made. We characterize the PDF p(r, α) of a landmark position expressed in polar coordinates when r and α are independent, and the marginal probability p(r) of the PDF is constrained to be uniform.

Robot Soccer KheperaSot League: Challenges and Future Directions

Robot soccer fosters AI and intelligent robotics research by providing a standard problem where a wide range of technologies can be integrated and examined. In order for a robot team to actually perform in a soccer game, various technologies must be incorporated including: design principles of autonomous agents, multi-agent collaboration, strategy acquisition, real-time reasoning, robotics, and sensor-fusion. In this paper, we discuss the speci c features of the KheperaSot league, describe the winner of the last two World Cups, and discuss the future directions of the KheperaSot league.