The ability to represent knowledge about space and its position therein is crucial for a mobile robot. To this end, topological and semantic descriptions are gaining popularity for augmenting purely metric space representations. In this paper we present a multi-modal place classification system that allows a mobile robot to identify places and recognize semantic categories in an indoor environment. The system effectively utilizes information from different robotic sensors by fusing multiple visual cues and laser range data. This is achieved using a high-level cue integration scheme based on a Support Vector Machine (SVM) that learns how to optimally combine and weight each cue. Our multi-modal place classification approach can be used to obtain a real-time semantic space labeling system which integrates information over time and space. We perform an extensive experimental evaluation of the method for two different platforms and environments, on a realistic off-line database and in a live experiment on an autonomous robot. The results clearly demonstrate the effec-

Abstract. Global Integral Invariant Features have shown to be useful for robot localization in indoor environments. In this paper, we present a method that uses integral invariants for outdoor environments. To make the integral invariant features more distinctive for outdoor images, we split the image into a grid of subimages and calculate integral invariants for each grid cell individually. We then concatenate the results to get the feature vector for the image. Additionally, we combine this method with a particle filter to improve the localization results. We compare our approach to a Scale Invariant Feature Transform (SIFT)-based approach on images of two outdoor areas under different illumination conditions. The results show that the SIFT approach is more exact, but the integral invariant approach is faster and allows localization in significantly less than one second.

We propose an approach to building topological maps of environments based on image sequences. The central idea is to use manifold constraints to find representative feature prototypes, so that images can be related to each other, and thereby to camera poses in the environment. Our topological map is built incrementally, performing well after only a few visits to a location. We compare our method to several other approaches to representing images. During tests on novel images from the same environment, our method attains the highest accuracy in finding images depicting similar camera poses, including generalizing across considerable seasonal variations. 1.

Abstract: In this paper, we present a geometrical localization method based on a combination of global image features. Our method represents each image by two feature vectors. The first feature vector is a Weighted Gradient Orientation Histogram (WGOH). The second feature vector is a Weighted Grid Integral Invariant (WGII) feature vector based on Integral Invariants. For localization, we use a particle filter which updates the weights of the particles based on image similarities calculated from the two feature vectors. We evaluate our approach on outdoor images of two different areas and under varying illumination and compare it to a SIFT-based approach. The comparison shows that the SIFT approach is slightly more exact than our method, but our method is more than four times faster than the SIFT approach and allows a localization frequency of more than 2 Hz.

Abstract — Pose estimation of outdoor robots presents some distinct challenges due to the various uncertainties in the robot sensing and action. In particular, global positioning sensors of outdoor robots do not always work perfectly, causing large drift in the location estimate of the robot. To overcome this common problem, we propose a new approach for global localization using place recognition. First, we learn the location of some arbitrary key places using odometry measurements and GPS measurements only at the start and the end of the robot trajectory. In subsequent runs, when the robot perceives a key place, our fixed-lag smoother fuses odometry measurements with the relative location to the key place to improve its pose estimate. Outdoor mobile robot experiments show that place recognition measurements significantly improve the estimate of the smoother in the absence of GPS measurements. I.

achallengingfieldinroboticvisionresearch.Apartfromartificial environmentalsupporttechnologieslikeGPSlocalization,aselfsufficientvisualsystemisdesirable.Inthiswork,weintroduce anewheuristicapproachtooutdoorlocalizationinascenario wherenoodometryreadingsareavailable.Inanearlierwork, weemployedSIFTfeaturesandacommonparticlefiltermethod in the scenario. A modification of Particle Swarm Optimization,apopularoptimizationtechniqueespeciallyindynamically changingenvironments,isdevelopedandfittothelocalization problem,includingself-adaptivemechanisms.Thenewmethod obtainssimilarorbetterlocalizationresultsinourexperiments, whilerequiringafractionofSIFTcomparisonsofthestandard method,indicatinganall-overspeed-upby25%. I.

Abstract — This paper describes a panoramic view-based navigation in outdoor environments. We have been developing a two-phase navigation method. In the training phase, the robot acquires image sequences along the desired route and automatically learns the route visually. In the subsequent autonomous navigation phase, the robot moves by localizing itself by comparing input images with the learned route representation. To be robust to changes of weather and seasons, an object-based comparison is adopted. Our previous method applied a support vector machine (SVM) algorithm to object recognition and localization and exhibited a satisfactory performance but was sometimes sensitive to the variation of the robot’s heading. This paper thus extends the method to use panoramic images. By searching the image for the region which matches the model image the most, a new method can considerably improve the localization performance and provide the robot with globally correct directions to move. Index Terms — Outdoor mobile robot, Panoramic visionbased localization, Support vector machine.

Abstract: Research on small mobile robots is challenging due to the low computational power and limited sensing of the robots. In this paper, we present a method that enables these types of robots to localize themselves visually in indoor environments. Our approach uses a compass to cope with the restricted visual content that a low-resolution image can provide. Therefore, in the localization phase the robot orients itself towards a given direction and uses global image features to determine its position. Also, the robot’s rotation impreciseness is included in the way the mapping is done. By real-world experiments we show that our method works despite of the restricted processing capabilities and the low resolution of the images. 1.

Global image features are well-suited for the visual self-localization of mobile robots. They are fast to compute, to compare and do not require much storage space. Especially when using small mobile robots with limited processing capabilities and low-resolution cameras, global features can be preferred to local features. In this paper, we compare the accuracy and computation times of different global image features when localizing small mobile robots. We test the methods under realistic conditions, taking illumination changes and translations into account. By employing a particle filter and reducing the image resolution, we speed up the localization process considerably. 1

Abstract—In this paper, the performance of a topologicalmetric visual path following framework is investigated in different environments. The framework relies on a monocular camera as the only sensing modality. The path is represented as a series of reference images such that each neighboring pair contains a number of common landmarks. Local 3D geometries are reconstructed between the neighboring reference images in order to achieve fast feature prediction. This allows recovery from tracking failures. During navigation the robot is controlled using image-based visual servoing. The focus of the paper is on the results from a number of experiments conducted in different environments, lighting conditions and seasons. The experiments with a robot-car show that the framework is robust against moving objects and moderate illumination changes. It is also shown that the system is capable of on-line path learning. Index Terms—visual servoing, mapping, localization, visual memory, path following