149544 research outputs found
Sort by
Generation of Annotated Data Using Pre-Trained Diffusion Models: Alleviating the data pain in the context of architectural facade segmentation
No full textThe success of modern machine learning methods such as
deep learning has broadened its influence on almost all research fields, including architecture and the built world. A common requirement for these data-centric technologies is the availability of data for training.
Dataset creation and curation is a time-consuming and error-prone
process. Depending on the complexity of annotations needed, this can
become a significant bottleneck. Particularly in tasks such as in
semantic segmentation, where labelling a single urban image may
require several minutes. Segmented data is essential for analysing and developing methodologies for objects in the built world. Yet existing façade segmentation datasets often suffer from inconsistencies, limited size and varying annotation standards. This work proposes an approach to bridge the data gap, using generative models. A pretrained stable diffusion model is finetuned on a custom dataset created with the purpose of providing a wide range of architectural styles and image types. In a subsequent stage, the pretrained model is used to train an ensemble of pixel classifiers to predict class labels for generated images. A data generation pipeline is proposed and prototypically tested, focusing on the generation of segmented data for facade labels
Experimental Validation and Calibration of GNSS-Based Phase Synchronization for Bistatic and Multistatic SAR Missions
No full textThis article addresses the critical issue of phase synchronization in multistatic synthetic aperture radar (SAR). We present the experimental validation of a global navigation satellite system (GNSS)-based synchronization technique planned for use in ESA’s upcoming Earth Explorer mission, Harmony. In this technique, the radar payload and GNSS receiver utilize the same main oscillator, and radar synchronization is achieved through the postprocessing of carrier phase data from the GNSS receiver and precise baseline determination (PBD) outputs. This article presents an experimental procedure that serves as a general proof of concept of the technique, a method for assessing the achievable synchronization accuracy for a given GNSS receiver, and a method to estimate the covariance matrix to optimize the weighting between the various carrier phase observables. We present point-to-point estimation and smoothing approaches. The technique achieved in a laboratory environment relative synchronization errors below 515 fs ( 1σ ), or 1° for a 5.4-GHz radar system, in a zero-baseline scenario, and below 1.5° at 5.4 GHz in a short-baseline scenario, in which the systems are physically separated
Collaborative Code-Based Differential GNSS for Multi-User Positioning
No full textThis paper introduces a collaborative differential GNSS (C-DGNSS) method designed to enhance positioning accuracy in scenarios with limited satellite visibility, relying solely on GNSS pseudorange measurements and requiring no inter-ranging information. C-DGNSS exploits the correlation among single difference observations from multiple users to a common base station, which are not leveraged by conventional DGNSS. The estimation problem is formulated as an iterative weighted least squares (WLS) estimator, known to be optimal under the Gaussian assumption, and its variance lower bound is derived from the Cramér-Rao bound (CRB). C-DGNSS is compared against conventional DGNSS and the ideal DGNSS bound, the latter assuming noiseless pseudorange measurements for the base station. We evaluate the impact of the number aiding users, their geometry, and measurement precision on positioning error, as well as the effect of favorable and poor geometric dilution of precision (GDOP) for the aided users. Simulation results demonstrate that the proposed scheme outperforms standard DGNSS even with a single aiding user, approaching the ideal bound as the number of aiding users grows and their geometry improves. Accuracy gains of up to 28.5\% are achieved
OpenDRIVE road network dataset of Schwarzer Berg in Brunswick (Version 1.0.0)
No full textThis lane-detailed HD map as OpenDRIVE dataset models road network parts of the "Schwarzer Berg" region in the city of Brunswick, Lower Saxony, Germany. The main application scopes of this data are in the domain of automated driving: simulation, verification and validation
Experiments and Modeling of a Single Multicopter Rotor Response to Sinusoidal Vertical Gusts
No full textMulticopters operate in environments subject to strongly gusting winds and need good aeromechanical models to improve the aircraft. A common, convenient, assumption is that the gusting inflow is quasi-static at each instant, but this assumption has never been tested. This paper shows that there is a solid physical basis for the simplified aerodynamic models of a single multicopter response to gusts. Experiments and computations show that using the static relationship between thrust or power and aerodynamic angle of attack for a single multicopter rotor (the quasi-static assumption) in sinusoidally pitching sideflow can be used to predict the thrust or power for unsteady variation of the angle of attack if the instantaneous flow
angle of the freestream is known. Vertical (angle) gusts up to 1885◦/s (k = 2.2 based on the diameter) and with a wavelength longer than the rotor diameter were shown to be covered by this assumption
Snapshot Estimator for Magnetic Field-based Train Localization with Uncalibrated Magnetometers
No full textMagnetic field-based train localization uses distortions of the Earth magnetic field caused by magnetic material in the vicinity of the tracks. From the distortions, a position estimate can be obtained by comparing magnetometer measurements to a map of the magnetic field. The comparison to the map requires the calibration of the train-mounted magnetometer. However, calibration is difficult because common calibration techniques require a rotation of the magnetometer and the platform it is mounted on. Thus, in this paper we propose a maximum likelihood (ML) snapshot estimator that simultaneously estimates the calibration parameters and the train position. The snapshot estimator reduces the ML estimation problem to a onedimensional optimization problem that can be solved efficiently. To show the feasibility of the approach, an evaluation is carried out based on measurements recorded with a train driving through Berlin
Physics-informed robotic airflow exploration and mapping with a swarm of mobile robots
No full textAirflow is the key transport mechanism for airborne substances like gas or particulate matter. It is of great interest in many applications ranging from evacuation planning to analyzing indoor ventilation systems. However, accurately determining a spatial map of the airflow is difficult and time-consuming since environmental parameters and boundary conditions are often unknown. This work introduces a novel adaptive sampling strategy for mobile robots. The strategy allows multiple mobile robots with anemometers to autonomously collect airflow measurements and generate a two-dimensional spatial map of the airflow field. Using a Domain-knowledge Assisted Exploration approach, the robots respond in real-time to the measurements already taken and determine the most informative locations online for further measurements. We incorporate the Navier-Stokes Partial Differential Equations to fuse the collected data with model assumptions. By casting the airflow model into a probabilistic framework, we can quantify uncertainties in the airflow field and develop an intelligent, uncertainty-driven exploration strategy inspired by optimal experimental design principles. This strategy combines an estimated uncertainty map with a rapidly exploring random tree path planner. Additionally, using the Navier-Stokes equations allows us to interpolate spatially between measurements in a physics-informed way, enabling us to construct a more accurate airflow map. We implemented and evaluated the proposed concept in simulations and experiments in a laboratory environment, where five mobile robots explore artificially generated airflow fields. The results indicate that our approach can correctly estimate the airflow and show that the proposed adaptive exploration strategy gathers information more efficiently than a predefined sampling pattern