Axel Rottmann

Learning maps in 3D using attitude and noisy vision sensors

In this paper, we address the problem of learning 3D maps of the environment using a cheap sensor setup which consists of two standard web cams and a low cost inertial measurement unit. This setup is designed for lightweight or flying robots. Our technique uses visual features extracted from the web cams and estimates the 3D location of the landmarks via stereo vision. Feature correspondences are estimated using a variant of the PROSAC algorithm. Our mapping technique constructs a graph of spatial constraints and applies an efficient gradient descent-based optimization approach to estimate the most likely map of the environment. Our approach has been evaluated in comparably large outdoor and indoor environments. We furthermore present experiments in which our technique is applied to build a map with a blimp.

A probabilistic sonar sensor model for robust localization of a small-size blimp in indoor environments using a particle filter

In recent years, autonomous miniature airships have gained increased interest in the robotics community. This is due to their ability to move safely and hover for extended periods of time. The major constraints of miniature airships come from their limited payload which introduces substantial constraints on their perceptional capabilities. In this paper, we consider the problem of localizing a miniature blimp with lightweight ultrasound sensors. Since the opening angle of the sound cone emitted by a sonar sensor depends on the diameter of the membrane, small-size sonar devices introduce the problem of high uncertainty about which object has been perceived. We present a novel sensor model for ultrasound sensors with large opening angles that allows an autonomous blimp to robustly localize itself in a known environment using Monte Carlo localization. As we demonstrate in experiments with a real blimp, our novel sensor model outperforms a popular sensor model that has in the past been shown to work reliably on wheeled platforms.

Using AdaBoost for Place Labeling and Topological Map Building

Indoor environments can typically be divided into places with different functionalities like corridors, kitchens, offices, or seminar rooms. We believe that the ability to learn such semantic categories from sensor data or in maps enables a mobile robot to more efficiently accomplish a variety of tasks such as human-robot interaction, path-planning, exploration, or localization. In this work, we first propose an approach based on supervised learning to classify the pose of a mobile robot into semantic classes. Our method uses AdaBoost to boost simple features extracted from vision and laser range data into a strong classifier. We furthermore present two main applications of this approach. Firstly, we show how our approach can be utilized by a moving robot for robust online classification of the poses traversed along its path using a hidden Markov model. Secondly, we introduce a new approach to learn topological maps from geometric maps by applying our semantic classification procedure in combination with probabilistic labeling. Experimental results obtained in simulation and with real robots demonstrate the effectiveness of our approach in various environments.

Learning non-stationary system dynamics online using Gaussian processes

Gaussian processes are a powerful non-parametric framework for solving various regression problems. In this paper, we address the task of learning a Gaussian process model of non-stationary system dynamics in an online fashion. We propose an extension to previous models that can appropriately handle outdated training samples by decreasing their influence onto the predictive distribution. The resulting model estimates for each sample of the training set an individual noise level and thereby produces a mean shift towards more reliable observations. As a result, our model improves the prediction accuracy in the context of non-stationary function approximation and can furthermore detect outliers based on the resulting noise level. Our approach is easy to implement and is based upon standard Gaussian process techniques. In a real-world application where the task is to learn the system dynamics of a miniature blimp, we demonstrate that our algorithm benefits from individual noise levels and outperforms standard methods.

Adaptive autonomous control using online value iteration with gaussian processes

In this paper, we present a novel approach to controlling a robotic system online from scratch based on the reinforcement learning principle. In contrast to other approaches, our method learns the system dynamics and the value function separately, which permits to identify the individual characteristics and is, therefore, easily adaptable to changing conditions. The major problem in the context of learning control policies lies in high-dimensional state and action spaces, that needs to be explored in order to identify the optimal policy. In this paper, we propose an approach that learns the system dynamics and the value function in an alternating fashion based on Gaussian process models. Additionally, to reduce computation time and to make the system applicable to online learning, we present an efficient sparsification method. In experiments carried out with a real miniature blimp we demonstrate that our approach can learn height control online. Further results obtained with an inverted pendulum show that our method requires less data to achieve the same performance as an off-line learning approach.

How to Extract the Sheet Resistance and Hall Mobility From Arbitrarily Shaped Planar Four-Terminal Devices With Extended Contacts

Van der Pauws method enables the sheet resistance Rsq and the Hall mobility μH to be extracted from arbitrarily shaped simply connected planar samples with four peripheral pointlike contacts. This paper generalizes the method for devices with extended contacts. It is found that Rsq and μH can be extracted using only six resistance measurements in the absence of a magnetic field and a single magnetic sensitivity measurement. Conversely, if the μH of a simply connected planar conducting device with peripheral contacts is known, the magnetic sensitivity of the device can be predicted based on six resistance measurements in the absence of a magnetic field, without any further knowledge of the device geometry. The new method is applied to a variety of differently shaped diffused silicon n-wells with peripheral contacts. The extracted sheet resistance and Hall mobility values show excellent consistency and are in agreement with the fabrication specifications.

Gaussian process based state estimation for a gyroscope-free IMU

A gyroscope-free inertial measurement unit (GF-IMU) utilizes only accelerometers to determine the relative movement of a body. We consider the problem of merging the individual accelerometer measurements using an Unscented Kalman filter (UKF) to estimate the body motion. Conventionally, this is realized by a parametric observation model, which gives the expected sensor measurements. In this paper, we replace this model by a Gaussian process (GP). GPs are a state-of-the-art non-parametric Bayesian regression framework. Thereby, the measurements are determined based on a sampled set of training data. No physical principles of the system must be described. In addition, we apply sparse GPs using pseudo-inputs to reduce computation time, while the estimation accuracy remains nearly constant. As a result, the filter cycle time decreases by a factor of 1.92. We present accuracy measurements obtained on a 3D rotation table and compare the results to estimates generated with a parametric model.