We propose a data-driven scene flow estimation algorithm exploiting the observation that many 3D scenes can be explained by a collection of agents moving as rigid bodies. At the core of our method lies a deep architecture able to reason at the object-level by considering 3D scene flow in conjunction with other 3D tasks. This object level abstraction enables us to relax the requirement for dense scene flow supervision with simpler binary background segmentation mask and ego-motion annotations. Our mild supervision requirements make our method well suited for recently released massive data collections for autonomous driving, which do not contain dense scene flow annotations. As output, our model provides low-level cues like pointwise flow and higher-level cues such as holistic scene understanding at the level of rigid objects. We further propose a test-time optimization refining the predicted rigid scene flow. We showcase the effectiveness and generalization capacity of our method on four different autonomous driving datasets. We release our source code and pre-trained models under github.com/zgojcic/Rigid3DSceneFlow.

We introduce PREDATOR, a model for pairwise point-cloud registration with deep attention to the overlap region. Different from previous work, our model is specifically designed to handle (also) point-cloud pairs with low overlap. Its key novelty is an overlap-attention block for early information exchange between the latent encodings of the two point clouds. In this way the subsequent decoding of the latent representations into per-point features is conditioned on the respective other point cloud, and thus can predict which points are not only salient, but also lie in the overlap region between the two point clouds. The ability to focus on points that are relevant for matching greatly improves performance: PREDATOR raises the rate of successful registrations by more than 20% in the low-overlap scenario, and also sets a new state of the art for the 3DMatch benchmark with 89% registration recall. [Code release]

We propose 3DSmoothNet, a full workflow to match 3D point clouds with a siamese deep learning architecture and fully convolutional layers using a voxelized smoothed density value (SDV) representation. The latter is computed per interest point and aligned to the local reference frame (LRF) to achieve rotation invariance. Our compact, learned, rotation invariant 3D point cloud descriptor achieves 94.9% average recall on the 3DMatch benchmark data set, outperforming the state-of-the-art by more than 20 percent points with only 32 output dimensions. This very low output dimension allows for near realtime correspondence search with 0.1 ms per feature point on a standard PC. Our approach is sensor- and scene-agnostic because of SDV, LRF and learning highly descriptive features with fully convolutional layers. We show that 3DSmoothNet trained only on RGB-D indoor scenes of buildings achieves 79.0% average recall on laser scans of outdoor vegetation, more than double the performance of our closest, learning-based competitors. Code, data and pre-trained models are available online at https://github.com/zgojcic/3DSmoothNet.

Ultra-wideband (UWB) time of flight (ToF) ranging is nowadays one of the most attractive technologies to implement indoor localization solutions with reliable dm-level accuracy. UWB systems are generally resistant to multipath interference. However, non-line-of-sight (NLOS) components with small relative delays may introduce errors in the estimated distances well exceeding the typically achievable accuracy. We present an empirical accuracy assessment of UWB ranging using a commercial UWB system. A particular focus is on the magnitude and spatial patterns of multipath errors. A large dataset comprising distances between 0.2 and 100 m was collected in a geodetic metrology lab and outdoors for this purpose. We derived ground truth of the distances with superior accuracy using a laser tracker and a total station. The results show that multipath errors are highly repeatable if the geometrical configuration does not change, but can degrade the accuracy of the distance measurements locally by almost an order of magnitude. While typical biases were on the order of a few cm, peak values exceeding 60 cm were found, and the standard deviation, typically about 3 cm or less, increased up to almost 15 cm in certain locations. We have used a part of the large dataset to train an artificial neural network (ANN) for error prediction. The ANN uses the measured ranges and additional signal features, also provided by the ranging nodes, as inputs. For error mitigation, the ANN output was subtracted from the original range measurements. This improved the ranging accuracy over the whole test data by practically eliminating the bias on average and reducing the overall standard deviation by a factor of 3. The results indicate that the relation between the output of the UWB node and the range errors can be modeled to a practically useful degree even without in-depth knowledge of the specific UWB node's measurement process and characteristics. The proposed approach can help to significantly improve the UWB ranging performance in presence of multipath and thus also the accuracy of a corresponding localization system.

A non-vector-based dissimilarity measure is proposed by combining vector-based distance metrics and set operations. This proposed compound dissimilarity measure (CDM) is applicable to quantify the similarity of collections of attribute/feature pairs where not all attributes are present in all collections. This is a typical challenge in the context of e.g., fingerprinting-based positioning (FbP). Compared to vector-based distance metrics (e.g., Minkowski), the merits of the proposed CDM are i) the data do not need to be converted to vectors of equal dimension, ii) shared and unshared attributes can be weighted differently within the assessment, and iii) additional degrees of freedom within the measure allow to adapt its properties to application needs in a data-driven way. We indicate the validity of the proposed CDM by demonstrating the improvements of the positioning performance of fingerprinting-based WLAN indoor positioning using four different datasets, three of them publicly available. When processing these datasets using CDM instead of conventional distance metrics the accuracy of identifying buildings and floors improves by about 5% on average. The 2d positioning errors in terms of root mean squared error (RMSE) are reduced by a factor of two, and the percentage of position solutions with less than 2m error improves by over 10%.

Foot-mounted inertial positioning (FMIP) can face problems of inertial drifts and unknown initial states in real applications, which renders the estimated trajectories inaccurate and not obtained in a well defined coordinate system for matching trajectories of different users. In this paper, an approach adopting received signal strength (RSS) measurements for Wifi access points (APs) are proposed to align and calibrate the trajectories estimated from foot mounted inertial measurement units (IMUs). A crowdsourced radio map (RM) can be built subsequently and can be used for fingerprinting based Wifi indoor positioning (FWIP). The foundation of the proposed approach is graph-based simultaneously localization and mapping (SLAM). The nodes in the graph denote users' poses and the edges denote the pairwise constrains between the nodes. The constrains are derived from: (1) inertial estimated trajectories; (2) vicinity in the RSS space. With these constrains, an error functions is defined. By minimizing the error function, the graph is optimized and the aligned/calibrated trajectories along with the RM are acquired. The experimental results have corroborated the effectiveness of the approach for trajectory alignment, calibration as well as RM construction.

A multi-objective evolutionary algorithm for optimal anchor placement of an infrared indoor positioning system is proposed. As a novel contribution the multipath error, a severe cause of inaccuracies (up to some tens of cm), is taken into account as an objective together with precision in a genetic algorithm. Multipath error and precision of the estimated position are assessed as scalar functions of the client position within an entire region of interest using models of the phase-based measurements and their propagation into estimated position. The precision is assessed using the trace of the covariance matrix of the estimated coordinates assuming position-dependent variances. A specular model is applied to calculate the respective multipath errors, and the appropriate number of multipath components to be included is analyzed. The method is applied to a geometrically simple scenario using realistic system parameters and environmental parameters to demonstrate the feasibility and highlight the impact of multipath on optimization of anchor placement. The results are sets of quasi-optimal configurations within a non-restricted anchor-location space.

A new measuring architecture for phase-based infrared ranging applied to indoor positioning is presented. The motivation to develop this proposal was to reduce the critical effect of multipath interferences on the differential distances estimated by the ranging system. The multipath mitigation feature of the proposed system is based on applying a spread spectrum modulation on the emitted signal so that the line-of-sight component reaching the receivers can be discriminated from the non-line-of-sight components by applying a selective coherent demodulation. An overview of the system architecture is provided, together with the first approach to the model of the demodulation process and its validation. The conclusions obtained from the study of the proposal are used to provide preliminary results of the mitigation capabilities of the system, and to analyze the key issues involved in its implementation. The results show that the effectiveness of the proposal is strongly related with the infrared link bandwidth. Mitigation levels above 50% for multipath 4.5 meters longer than the direct path are reached if the infrared link bandwidth is higher than 200 MHz.