We propose a submap-based technique for mapping of underwater structures with complex geometries. Our approach relies on the use probabilistic volumetric techniques to create submaps from multibeam sonar scans, as these offer increased outlier robustness. Special attention is paid problem denoising/enhancing data. Pairwise submap alignment constraints are used in factor graph framework correct navigation drift and improve map accuracy. provide experimental results obtained inspection running...

In this paper, we propose a mapping approach that constructs globally deformable virtual occupancy grid map (VOG-map) based on local submaps. Such representation allows pose graph SLAM systems to correct accumulated drift via loop closures while maintaining free space information for the purpose of path planning. We demonstrate use such implementing an underwater system in which robot actively plans paths generate accurate 3D scene reconstructions. evaluate performance simulated as well...

The problem of driving a formation from an initial to target configuration while under the effect external disturbances is studied. Additional restrictions on agent sensing as well inter-agent communication must be satisfied. We present leader-follower solution that relies simple uncertainty model trigger surfacing events. These events are then used update control signal, for which two different, provably correct, strategies proposed. Finally, we show how can characterize disturbance set....

From archaeology to the inspection of subsea structures, underwater mapping has become critical many applications. Because balanced trade-off between range and resolution, multibeam sonars are often used as primary sensor in platforms. These output an image representing intensity received acoustic echos over space, which must be classified into free occupied regions before measurements determined spatially registered. Most classifiers found literature use local thresholding techniques,...

Developing obstacle avoidance algorithms for low cost Autonomous Underwater Vehicles using single beam returned echo sounder sonar is a difficult task. In this paper, we propose an intelligent algorithm that maps the obstacle, avoids it with guarantee will not get stuck and efficiently traverses path towards destination navigation functions. We present complete system help of hybrid automata, probabilistic mapping Simulation results are presented showing validity our approach.

Typically, the reconstruction problem is addressed in three independent steps: first, sensor processing techniques are used to filter and segment data as required by front end. Second, end builds factor graph for obtain an accurate estimate of robot's full trajectory. Finally, product obtained further data, now re-projected from optimized In this paper we present approach model a way that unifies aforementioned problems under single framework particular application: sonar-based inspection...

The problem of driving a set vehicles (agents) to desired target configuration under event-based communication and measurement constraints is analyzed. We start by studying the single agent problem, where we propose waypoint based solution, along with two alternative control strategies. After deriving their properties proving some relevant results, proceed study case. generalize results obtained for this network multi network. Our strategy able position each within given distance its while...

Synthetic Aperture Sonar (SAS) is a technique to improve the spatial resolution from moving set of receivers by extending array in time, increasing effective length and aperture. This limited accuracy receiver position estimates, necessitating highly accurate, typically expensive aided-inertial navigation systems for submerged platforms. We leverage simultaneous localization mapping fuse acoustic navigational measurements obtain accurate pose estimates even without benefit absolute...

This paper presents a method for processing sparse, non-Gaussian multimodal data in simultaneous localization and mapping (SLAM) framework using factor graphs. Our approach demonstrates the feasibility of sum-product inference strategy to recover functional belief marginals from highly situations, relaxing prolific unimodal Gaussian assumption. The is more focused than conventional multi-hypothesis approaches, but still captures dominant modes via multi-modality. proposed algorithm exists...

We extend the core operation of non-parametric belief propagation (NBP), also known as multi-scale sequential Gibbs sampling, to approximate products kernel density estimated beliefs that reside on some manifold. The original algorithm, though multidimensional, implicitly assumes Euclidean ℝ <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">d</sup> space only. proposed extension generalizes any mixture Riemannian manifolds, provided primary...

One way to perform Autonomous Underwater Vehicle (AUV) localization is employ static acoustic beacons form a Long Baseline system. However, errors in the estimated positions of will negatively affect vehicle localization. To solve this problem we propose beacon position estimation technique using an Extended Kalman Filter, and also investigate how improve performance. Our solution able successfully estimate show performance by appropriately defining trajectory.

Abstract The installation of a subsea equipment such as manifold needs careful planning and coordination. Studies on the behavior dynamic responses are crucial to guarantee safety. Some important factors in these operations include current profile, waves characteristics, winches motions at topside, elastic cable (due resonance effects). Currently, most available commercial codes use simplified models for hydrodynamic forces submerged equipment. However, cases with complex geometries strong...

Stereoeletroencephalography (SEEG) has been used to define and understand the organization of epileptogenic zones brain. The fusion SEEG signals anatomy on a common referential is possible method for analysis these signals. This work describes an experiment conducted with human observers in order evaluate three different coding schemes visualize magnitude signal anatomical context.