Set-membership approach to the kidnapped robot problem

Abstract

Abstract-This article depicts an algorithm which matches the output of a Lidar with an initial terrain model to estimate the absolute pose of a robot. Initial models do not perfectly fit the reality and the acquired data set can contain an unknown, and potentially large, proportion of outliers. We present an interval based algorithm that copes with such conditions, by matching the Lidar data with the terrain model in a robust manner. Experimental validations using different terrain model are reported to illustrate the performance of the method.

Indoor Localization for Augmented Reality Devices Using BIM, Point Clouds, and Template Matching

“…While Boniardi et al [23] achieve a higher accuracy and reliability, the approach requires knowledge about the initial position of the device. The UAV localization presented by Desrochers et al [14] has a similar accuracy to our algorithm but requires a significantly longer computation time on similar-sized environments, making it not applicable for AR localization. In our algorithm, aligning the recorded point cloud through normal analysis allows for a significant reduction in computation time and applying template matching on floor plans ensures a high accuracy with good performance.…”

“…Magnusson et al [13] showed that using point clouds for real-time localization is feasible by using normal distribution transforms for loop closure. Using the geometry of the surroundings has the advantage of being independent of lighting conditions, as shown by Desrochers et al [14]. The authors achieve a localization accuracy of 20 cm and 20 • for localizing Unmanned Aerial Vehicles (UAVs) positions in known environments, but execution time for large environments can be up to 60 s. Using point-cloud matching in AR localization offers a significant simplification to the domain: AR localization may apply domain-specific heuristics to achieve a registration of its surrounding environment, while general point-cloud algorithms are built to work on arbitrary geometry.…”

Hitchhiking Based Symbiotic Multi-Robot Navigation in Sensor Networks

“…Hence, if the driver is lost, then it will be difficult for the hitchhiker to localize itself in the map as it is completely unaware of its current position in the map. This problem is similar to the famous 'kidnapped robot problem' for which solutions are available in literature [24][25][26]. However, we propose to recover from this problem in the first place by transferring the current estimated pose (x δ , y δ , θ δ ) and associated uncertainty (Σ δ ) information to the hitchhiker intermittently in intervals of δ s. This is graphically shown in Figure 2 where a driver is shown transferring information intermittently.…”

Robust data processing in the presence of uncertainty and outliers: Case of localization problems

“…For example, when solving a localization problem -whether it is a problem of locating a robot (see, e.g., [3]) or of locating a satellite (see, e.g., [4]) -we are interested in the coordinates p 1 , . .…”

“…This idea has been actively used, as a heuristic idea, to deal with data processing under outliers, see, e.g., [3], [7], [10]. Several practical applications of this heuristic idea are described, e.g., in [3].…”