1167 research outputs found
Sort by
AR-CP: Uncertainty-Aware Perception in Adverse Conditions with Conformal Prediction and Augmented Reality For Assisted Driving
Get PDFDeep learning models play a crucial role in improving driver assistance systems and environmental perception. However, their tendency toward overconfident predictions poses risks to driver safety, particularly in adverse conditions. To address this, we propose AR-CP, an uncertainty-aware framework integrating conformal prediction and augmented reality (AR). AR-CP starts with a conformal prediction step, generating an uncertainty-aware prediction set. Then, AR simplifies and clarifies the visualization of the closest common parent class, reducing misinformation. We present a rigorous formulation and theoretical analysis, evaluating AR-CP on the ROAD dataset. Results demonstrate superior performance compared to existing methods, ensuring safer driving experiences with reduced mental load and heightened situation awareness, as validated by an immersive user study involving 15 participants.0.
Data and code: Motility-induced coexistence of a hot liquid and a cold gas
No full textIf two phases exist at the same time, such as a gas and a liquid, they have the same temperature. This fundamental law of equilibrium physics is known to apply even to many non-equilibrium systems. However, recently, there has been much attention in the finding that inertial self-propelled particles like Janus colloids in a plasma or microflyers could self-organize into a hot gas-like phase that coexists with a colder liquid-like phase. With the data and code provided here, we show that a kinetic temperature difference across coexisting phases can occur even in equilibrium systems when adding generic (overdamped) self-propelled particles. In particular, we consider mixtures of overdamped active and inertial passive Brownian particles and show that when they phase separate into a dense and a dilute phase, both phases have different kinetic temperatures. Surprisingly, we find that the dense phase (liquid) cannot only be colder but also hotter than the dilute phase (gas). This effect hinges on correlated motions where active particles collectively push and heat up passive ones primarily within the dense phase. Our results answer the fundamental question if a non-equilibrium gas can be colder than a coexisting liquid and create a route to equip matter with self-organized domains of different kinetic temperatures
Dataset for Event Detection in Gait Analysis from 3D-Printed Piezoelectric PLA-Based Insole on an Instrumented Treadmill
No full textIn the experiment, one test person wearing a ferroelectret insole in the right shoe walks on an instrumented treadmill. The dataset contains data from four ferroelectret sensors in the insole and vertical ground reaction forces (GRF) from an instrumented treadmill across five different walking speeds. The ferroelectret insole data are filtered and given in Volt. The GRF treadmill data from the right side are filtered and given in Newton. The data are segmented into steps from one heel strike to the same foot's following heel strike and normalized into 1000 data points per step. Each row contains one step with 1000 columns. The amount of steps/rows depends on the walking speed: 74 for slowest v050, 124 for fastest v150.
Please find the full description in the corresponding publication, full instructions for usage in the README.md file. Copyright of thumbnail image 2024, IEEE
Compressor Atlas Copco GA22_VSD
No full textData set of the compressor Atlas Copco GA22_VSD in the ETA Research Factory supply system
5Pils dataset
No full textThe 5Pils dataset accompanies the paper "'Image, tell me your story!' Predicting the original meta-context of visual misinformation". The dataset contains the meta-context annotations of 1,676 images fact-checked by the organizations Factly, Pesachcek, and 211Check.
The dataset does not contain the images. To download the images, please refer to our GitHub code (https://github.com/UKPLab/5pils/tree/main).
5Pils is made available under a CC-BY-SA-4.0 license.
Please cite our paper if you find 5Pils useful to your work.1.
Source Code for Concerted Control: Simulating Robust Bipedal Gaits at Various Speeds in MuJoCo
No full textThis repository contains source code associated with the paper "Concerted Control: Modulating Joint Stiffness Using GRF for Gait Generation at Different Speeds" by Shunsuke Koseki, Omid Mohseni, Dai Owaki, Mitsuhiro Hayashibe, Andre Seyfarth, and Maziar A. Sharbafi.
The code simulates a bipedal model using the MuJoCo physics engine, representing a human with a height of 180cm and a weight of 80kg. The model's movement is constrained to the sagittal plane and includes seven degrees of freedom: one torso joint (between the pelvis and the torso), two hip joints, two knee joints, and two ankle joints.
The controller implemented in the model is a bioinspired, simple, and easy-to-implement walking controller, termed Concerted Control. This controller leverages a shared common signal to coordinate movements across multiple joints without relying on predefined trajectories. It builds on our previously developed Force Modulated Compliance (FMC) control concept, which modulates joint stiffness based on ground reaction forces (GRF). In Concerted Control, FMC is applied across multiple joints, enabling implicit coordination through the shared GRF signal, without the need for a centralized controller.
We evaluated the performance of Concerted Control on the simulated bipedal walker and demonstrated that it can generate stable walking gaits across a wide range of speeds, from 0.7 to 1.8m/s. Additionally, robustness was assessed through external angular momentum perturbation tests, which showed the gaits to be robust. By replicating key kinematic and kinetic characteristics of human walking, Concerted Control offers a promising framework for enhancing the control of mobile robots and assistive systems
Comparison of empirical and RL-based control for BATEX
No full textThe dataset includes EMG and GRF data recorded at Slow walking of 0.788m/s of 4 subjects and preferred walking speed of 3 subjects while they were wearing the exosuit BATEX. Each subject was recorded in 4 different settings (without the exosuit, with the exosuit without a control (so it is mainly the weight of the exo), using the exosuit with an empirically developed control and using the exosuit with en RL-based control).
The file names for the EMG data and the GRF data are structured by the subject number, the speed, the type of data, and the setting used to record that data.
Descriptions of how to use the Matlab code are provided in the Matlab files. First, run "GetStepsFromEMG" to create the average steps for each subject. With "CreateGraphs" the average step of all subjects is created as well as the plots
Supplementary data for publication "Algorithmic Planning of Ventilation Systems: Optimising for Life- Cycle Costs and Acoustic Comfort"
No full textThis dataset contains data for the paper "Algorithmic Planning of Ventilation Systems in Buildings: Optimising for Life-Cycle Costs and Acoustic Comfort". It contains a json file for the load cases used in the optimisations as well as the data to figure 11, 12, 13, 16 and figure 17 which contain the solutions.2.0.
LC-Quartier-Tool
No full textLC-Quartier ist ein lebenszyklusbasiertes Berechnungstool für Treibhausgasemissionen und kumulierten Energieaufwand für Wohnquartiere. Es basiert auf „synthetischen Quartieren“ für generische Berechnungen. Die hinterlegten Default-Daten können in Planungsprozessen sukzessiv durch Realdaten ersetzt werden bzw. durch Schnittstellen zu dem GMK® eingepflegt werden.
Das Tool kann sowohl für synthetische Quartiere, als auch reale Gebäude und Quartiere angewendet werden
ArgMining 2023 Shared Task Auxiliary Data NLPeer
No full textThis dataset comprises the auxiliary data for the 2023 Argumentation Mining Shared Task on Automatic Review Pragmatic Labeling in a domain-transfer setting. Please cite the original NLPeer paper when using this data