The International Journal of Robotics Research

High-fidelity simulation is a key enabling technology for the widespread deployment of large unmanned ground vehicles (UGVs). However, current approaches for lidar simulation leave much to be desired, particularly for scenes with vegetation. We introduce a novel 3D mapping technique that learns high-fidelity models for geo-specific lidar simulation directly from pose tagged lidar data. We introduce a novel stochastic, volumetric model that captures and can reproduce the statistical interactions of lidar with terrain. We show how to automatically learn the model directly from 3D mapping data collected by a UGV in the target environment. We extend our approach using terrain-classification techniques to develop a hybrid surface–volumetric model that combines the efficiency of surface modeling for areas that are well approximated by large surfaces (e.g. roads, bare earth) with our volumetric approach for more complex areas (e.g. bushes, trees) without sacrificing overall fidelity. We quantitatively compare the performance of our approach against more conventional methods on large outdoor datasets from urban and off-road environments. Our results show significant performance gains using our volumetric and hybrid approaches over the state-of-the-art, laying the ground work for truly high-fidelity simulation engines for UGVs.

Registration of range sensor measurements is an important task in mobile robotics and has received a lot of attention. Several iterative optimization schemes have been proposed in order to align three-dimensional (3D) point scans. With the more widespread use of high-frame-rate 3D sensors and increasingly more challenging application scenarios for mobile robots, there is a need for fast and accurate registration methods that current state-of-the-art algorithms cannot always meet. This work proposes a novel algorithm that achieves accurate point cloud registration an order of a magnitude faster than the current state of the art. The speedup is achieved through the use of a compact spatial representation: the Three-Dimensional Normal Distributions Transform (3D-NDT). In addition, a fast, global-descriptor based on the 3D-NDT is defined and used to achieve reliable initial poses for the iterative algorithm. Finally, a closed-form expression for the covariance of the proposed method is also derived. The proposed algorithms are evaluated on two standard point cloud data sets, resulting in stable performance on a par with or better than the state of the art. The implementation is available as an open-source package for the Robot Operating System (ROS).

In this paper we propose a vision-based online mapping of large-scale environments. Our approach uses a hybrid representation of a fully metric Euclidean environment map and a topological map. This novel hybrid representation facilitates our scalable online hierarchical bundle adjustment approach. The proposed method achieves scalability by solving the local registration through embedding neighboring keyframes and landmarks into a Euclidean space. The global adjustment is performed on a segmentation of the keyframes and posed as the iterative optimization of the arrangement of keyframes in each segment and the arrangement of rigidly moving segments. The iterative global adjustment is performed concurrently with the local registration of the keyframes in a local map. Thus, the map is always locally metric around the current location, and likely to be globally consistent. Loop closures are handled very efficiently benefiting from the topological nature of the map and overcoming the loss of the metric map properties of previous approaches. The effectiveness of the proposed method is demonstrated in real-time on various challenging video sequences.

We present an efficient and featureless approach to bathymetric simultaneous localization and mapping (SLAM) that utilizes a Rao–Blackwellized particle filter (RBPF) and Gaussian process (GP) regression to provide loop closures in areas with little to no overlap with previously explored terrain. To significantly reduce the memory requirements (thereby allowing for the processing of large datasets) a novel map representation is also introduced that, instead of directly storing estimates of seabed depth, records the trajectory of each particle and synchronizes them to a common log of bathymetric observations. Upon detecting a loop closure each particle is weighted by matching new observations to the current predictions generated from a local reconstruction of their map using GP regression. Here the spatial correlation in the environment is fully exploited, allowing predictions of seabed depth in areas that may not have been directly observed previously. The results demonstrate how observations of seafloor structure with partial overlap can be used by bathymetric SLAM to improve map self consistency when compared with dead reckoning fused with long-baseline (LBL) observations. In addition we show how mapping corrections can still be achieved even when no map overlap is present.

In this paper, we propose a stochastic differential equation-based exploration algorithm to enable exploration in three-dimensional indoor environments with a payload-constrained micro-aerial vehicle (MAV). We are able to address computation, memory, and sensor limitations by using a map representation which is dense for the known occupied space but sparse for the free space. We determine regions for further exploration based on the evolution of a stochastic differential equation that simulates the expansion of a system of particles with Newtonian dynamics. The regions of most significant particle expansion correlate to unexplored space. After identifying and processing these regions, the autonomous MAV navigates to these locations to enable fully autonomous exploration. The performance of the approach is demonstrated through numerical simulations and experimental results in single- and multi-floor indoor experiments.

Inspection of ship hulls and marine structures using autonomous underwater vehicles has emerged as a unique and challenging application of robotics. The problem poses rich questions in physical design and operation, perception and navigation, and planning, driven by difficulties arising from the acoustic environment, poor water quality and the highly complex structures to be inspected. In this paper, we develop and apply algorithms for the central navigation and planning problems on ship hulls. These divide into two classes, suitable for the open, forward parts of a typical monohull, and for the complex areas around the shafting, propellers and rudders. On the open hull, we have integrated acoustic and visual mapping processes to achieve closed-loop control relative to features such as weld-lines and biofouling. In the complex area, we implemented new large-scale planning routines so as to achieve full imaging coverage of all the structures, at a high resolution. We demonstrate our approaches in recent operations on naval ships.

This paper describes a solution to robot navigation on curved 3D surfaces. The navigation system is composed of three successive subparts: a perception and representation, a path planning, and a control subsystem. The environment structure is modeled from noisy lidar point clouds using a tool known as tensor voting. Tensor voting propagates structural information from points within a point cloud in order to estimate the saliency and orientation of surfaces or curves found in the environment. A specialized graph-based planner establishes connectivities between robot states iteratively, while considering robot kinematics as well as structural constraints inferred by tensor voting. The resulting sparse graph structure eliminates the need to generate an explicit surface mesh, yet allows for efficient planning of paths along the surface, while remaining feasible and safe for the robot to traverse. The control scheme eventually transforms the path from 3D space into 2D space by projecting movements into local surface planes, allowing for 2D trajectory tracking. All three subparts of our navigation system are evaluated on simulated as well as real data. The methods are further implemented on the MagneBike climbing robot, and validated in several physical experiments related to the scenario of industrial inspection for power plants.

This article proposes two methods based on cooperation between climbing and ground robots in order to address the mapping problem for autonomous inspection of three-dimensional (3D) structures. A pole climbing robot was developed to autonomously inspect a 3D human-made structure. The robot is able to climb over 3D human-made structures with bends and T-junctions. In the previous version of the system, the robot operator had to provide a set of data, resembling the map of the 3D structure, to the path planning algorithm of the climbing robot. However, the necessity of a priori knowledge of the structure’s geometry reduces the autonomy of the system. In this article we propose two solutions in which the structure in front of the robot is mapped autonomously. In the first proposed method, ground robots act as a mobile observer with a wide coverage for a climbing robot, to detect and estimate the size of the structure being climbed. We will present a case study in which multiple terrain robots provide the model of a structure which should be explored by a pole climbing robot. Each terrain robot is equipped with a low-cost wide angle Visual Graphics Array (VGA) camera, and some markers are fixed on the climbing robot. At each navigation step, the climbing robot and terrain robots cooperate to model a part of the structure which should be explored by the climbing robot. We also present a second approach in which a depth image from a Microsoft Kinect (a motion sensing input device by Microsoft) is fused with the information from the camera in order to eliminate the dependency of the system on color and light conditions.

In this paper we address the problem of controlling the motion of a group of unmanned aerial vehicles (UAVs) bound to keep a formation defined in terms of only relative angles (i.e. a bearing formation). This problem can naturally arise within the context of several multi-robot applications such as, e.g. exploration, coverage, and surveillance. First, we introduce and thoroughly analyze the concept and properties of bearing formations, and provide a class of minimally linear sets of bearings sufficient to uniquely define such formations. We then propose a bearing-only formation controller requiring only bearing measurements, converging almost globally, and maintaining bounded inter-agent distances despite the lack of direct metric information.

The controller still leaves the possibility of imposing group motions tangent to the current bearing formation. These can be either autonomously chosen by the robots because of any additional task (e.g. exploration), or exploited by an assisting human co-operator. For this latter ‘human-in-the-loop’ case, we propose a multi-master/multi-slave bilateral shared control system providing the co-operator with some suitable force cues informative of the UAV performance. The proposed theoretical framework is extensively validated by means of simulations and experiments with quadrotor UAVs equipped with onboard cameras. Practical limitations, e.g. limited field-of-view, are also considered.