Find research datasets worth reusing

opencc-by-4.0Apr 2024View details →

CoCar NextGen: a Multi-Purpose Platform for Connected Autonomous Driving Research

Sample sensor data and vehicle information of our autonomous research vehicle CoCar NextGen. This repository contains a short sequence of sample data from all sensors (except radar) of our research platform. The data is collected in a ROSBag. Additionally, sensor postions are provided in a URDF file. Finally, basic vehicle parameters are given in a YAML file.

opencc-by-4.0Mar 2022View details →

PAsCAL Pilot 2 Autonomous Driving Training

The video documents and describes the activities of Pilot 2 "Autonomous Driving Training" of the PAsCAL vreal-world pilots. This pilot took place at the ACI-Sara Safe Driving Centre in Lainate, near Milan (Italy) on February 2022. It involved trainers, professional and experienced drivers to test the study on a new driving training methodology to encourage people to accept the transition from conventional cars with partial autonomous driving capabilities to fully autonomous cars.

opencc-by-nc-sa-4.0Apr 2023View details →

Consensual videos of potentially re-identifiable individuals recorded at the Autonomous Driving Test Area Baden-Württemberg (raw images with location and IMU data).

For the purpose of research on data intermediaries and data anonymisation, it is necessary to test these processes with realistic video data containing personal data. For this purpose, the <a href="http://treumoda.de">Treumoda</a> project, funded by the German Federal Ministry of Education and Research (BMBF), has created a dataset of different traffic scenes containing identifiable persons. This video data was collected at the <a href="https://taf-bw.de/">Autonomous Driving Test Area Baden-Württemberg</a>. On the one hand, it should be possible to recognise people in traffic, including their line of sight. On the other hand, it should be usable for the demonstration and evaluation of anonymisation techniques. The legal basis for the publication of this data set the consent given by the participants as documented in the file Consent.pdf (all purposes) in accordance with Art. 6 1 (a) and Art. 9 2 (a) GDPR. Any further processing is subject to the GDPR. We make this dataset available for non-commercial purposes such as teaching, research and scientific communication. Please note that this licence is limited by the provisions of the GDPR. Anyone downloading this data will become an independent controller of the data. This data has been collected with the consent of the identifiable individuals depicted. Any consensual use must take into account the purposes mentioned in the uploaded consent forms and in the privacy terms and conditions provided to the participants (see Consent.pdf). All participants consented to all three purposes, and no consent was withdrawn at the time of publication. KIT is unable to provide you with contact details for any of the participants, as we have removed all links to personal data other than that contained in the published images.

opencc-by-4.0Apr 2023View details →

Business process models for a ride fulfilment process in a ride-hailing company enabled by an autonomous driving system

The repository contains the business process models of a ride fulfilment process in a ride-hailing company enabled by an autonomous driving system. The models are depicted using BPMN.2.0 language and are extended with annotations for usage in the <a href="https://dpotool.cs.ut.ee/">DPO tool</a>. The models, which contain "Pleak" as a part of the title, are created using PE-BPMN language and some of them are annotated for usage in the <a href="https://pleak.io/">Pleak</a> tool set.  The repository also contains two .csv files which contain the results of the simple disclosure analysis ("6_RideFulfillment_PK_SecSharing_combined-Simple-Disclosure-results.csv") and the results of the leak-when analysis ("7_Pleak_RideFulfillment_PK_SecrSharing_BPMN_Leak_When-results.csv") in the ride fulfilment process. Both analyses have been conducted using the <a href="https://pleak.io/">Pleak</a> tool set and the business process models with respective names from this repository.

opencc-by-4.0Apr 2023View details →

ADASIND: A Diverse Wide Angle Fisheye Camera Dataset for Autonomous Driving

This dataset comprises images collected from a fisheye lens on Indian roads to capture a variety of driving scenarios in a real-world environment, with a focus on semi-urban roads that don't have dividers.. The dataset includes 10,000 RGB images and has well labeled annotations for 2d bounding box, moving object detection, calibration of the fisheye lens containing parameters, ground truth of the RGB images and previous image of the corresponding frames. With this dataset, we would like to encourage the community to adapt computer vision models for the fisheye camera instead of using naive rectification.  The Adasind dataset was collected by recording a video of more than 10,000 seconds as a vehicle drove across the Silchar city, starting from the main gate of the National Institute of Technology Silchar during different time periods of the day considering traffic variations. To capture the video, we used a smartphone with a 64-megapixel camera, which was attached to a fisheye lens and a tripod stand. The smartphone was firmly held from the backside of the motor vehicle, ensuring that it was stable and wouldn't move during the recording process. The vehicle drove through the city at a steady pace, capturing video footage at approximately 30 frames per second (fps). The annotations for various tasks is available in their respective folder and described in details below: <ol> <li> 2d box annotations : This folder contains the object label, x-y coordinates, width and height of the various objects within the image in the .txt file. Here, the Yolov5 algorithm is used to perform object detection on a set of input images which uses a deep neural network to detect objects within an image. The files of this folder are in the same sequence as that of the original RGB image with the same file names. The use of 2D bounding box annotations has several benefits in computer vision, including object detection. 2D bounding boxes can be used to train object detection models, which are designed to identify the location and extent of objects within an image. </li> <li> Motion instance annotations : This folder contains the annotations for the moving objects which have relative motion with respect to the motion of the vehicle. Those objects create bias in specific tasks like depth estimation and have to be masked before feeding into the network. The files in the folder are in the same sequence as the original image and can be masked pixel wise. </li> <li> Calibration : The process of turning the 3D world into a 2D image is done by a camera and this process is said to be calibration which can be done using the camera parameters namely the intrinsic and extrinsic parameters. As of now only a single json file is dumped with the parametric values of the camera and users can make separate copies of the file for calibration tasks related to each original image. Together, the intrinsic and extrinsic parameters are used to model the relationship between the 3D world and the 2D image captured by the camera. This information is essential in many computer vision applications, such as 3D reconstruction, object tracking, and augmented reality. </li> <li> Ground Truth : This is a subfolder inside the "Motion instance annotations" folder that contains the correct labels or annotations for each image or object in a dataset. These labels are in the form of images and the filenames are the same as those of the original images. Ground truth labels are used during evaluation to assess the accuracy and performance of the machine learning or deep learning models. The accuracy of a model is measured by comparing its predicted labels to the ground truth labels, using evaluation metrics. </li> <li> RGB images: This folder contains the RGB images extracted from the video recording (30 fps) by considering one frame  per second with 30 frame interval. A total of  10,000 RGB images are there in the folder indexed sequentially as  appeared in the video recording.  </li> <li> Previous Image (not provided) :  The folder for the previous images corresponding to each original image may be prepared as per the requirement using the original images present in the "RGB image" folder.  The use of previous images is especially relevant in computer vision tasks that involve motion prediction or scene understanding, as it enables algorithms to better understand the dynamics of the scene and predict future states.   </li> </ol> The original images can be converted to 3d tensors with RGB channels or grayscale channels. The 2d box annotated values can be fed to the network for better understanding of the objects, the RGB image tensor values can be matrix multiplied with intrinsic parameters for converting from 3d to 2d and vice-versa, motion images can be multiplied with their respective original image so the moving objects is not taken into consideration, evaluation for the model with tensors from the ground truth can be done and at last previous image tensors can also be fetched along with the current image for upcoming predicted tasks.

opencc-by-nc-sa-4.0Oct 2023View details →

Consensual videos of potentially re-identifiable individuals recorded at the Autonomous Driving Test Area Baden-Württemberg (raw images recorded daytime).

For the purpose of research on data intermediaries and data anonymisation, it is necessary to test these processes with realistic video data containing personal data. For this purpose, the <a href="http://treumoda.de/">TreuMoDa</a> project, funded by the German Federal Ministry of Education and Research (BMBF), has created a dataset of different traffic scenes containing identifiable persons.This video data was collected at the <a href="https://taf-bw.de/">Autonomous Driving Test Area Baden-Württemberg</a>. On the one hand, it should be possible to recognise people in traffic, including their line of sight. On the other hand, it should be usable for the demonstration and evaluation of anonymisation techniques.The legal basis for the publication of this data set the consent given by the participants as documented in the file Consent.pdf (all purposes) in accordance with Art. 6 1 (a) and Art. 9 2 (a) GDPR. Any further processing is subject to the GDPR.We make this dataset available for non-commercial purposes such as teaching, research and scientific communication. Please note that this licence is limited by the provisions of the GDPR. Anyone downloading this data will become an independent controller of the data. This data has been collected with the consent of the identifiable individuals depicted.Any consensual use must take into account the purposes mentioned in the uploaded consent forms and in the privacy terms and conditions provided to the participants (see Consent.pdf). All participants consented to all three purposes, and no consent was withdrawn at the time of publication. KIT is unable to provide you with contact details for any of the participants, as we have removed all links to personal data other than that contained in the published images.

opencc-by-4.0Jun 2018View details →

Experimental Evaluation LTV-MPC Stability for Autonomous Driving

This video presents the experimental evaluation of LTV-MPC stability at Scania’s test track facilities located in Södertälje, Sweden, south of Stockholm.

opencc-by-4.0Sep 2024View details →

EMS3D-KITTI-Synthetic: A Synthetic 3D Dataset in KITTI Format with Balanced EMS Vehicle Distribution for Autonomous Driving AI Model Training

A 3D synthetic dataset in KITTI format, focused on emergency vehicles such as ambulances and police cars. The dataset was generated across 8 towns within the CARLA simulator and converted into the KITTI format, ensuring compatibility for direct use in AI model training for autonomous driving applications.

opencc-by-4.0Nov 2022View details →

PAsCAL WP6 Pilot 2 Autonomous Driving Training Video

A summary video documenting and introducing the activities of pilot 2 of the PAsCAL real-world pilots. The part of the pilot which is documented in the video took place at the Automobile Club d'Italia Vallelunga circuit in Rome, Italy between January 2022 and February 2022. The participants took part in a driving training for autonomous vehicles and specifically for advanced ADAS systems. Some took part in the training before and some after completing a driving course, where they had to overcome several different challenging conditions and manoeuvres.

opencc-by-4.0Jan 2023View details →

A traffic graph dataset for autonomous driving research: Commonroad-Nuplan-Dataset

We provide a standardized graph dataset for traffic based on the large-scale <a href="https://www.nuscenes.org/nuplan">NuPlan </a>v1.1 dataset, converted to a <a href="https://pytorch-geometric.readthedocs.io/en/latest/">PyTorch-Geometric </a>dataset using our <a href="https://commonroad.in.tum.de/tools/commonroad-geometric">CommonRoad-Geometric </a>tool. The dataset is collected from 3 megacities: Singapore, Boston and Pittsburgh.

opencc-by-4.0Feb 2023View details →

NexusStreets: a dataset combining human and autonomous driving behaviours

The NexusStreets dataset contains human and autonomous driving scenes. They are collected by monitoring a target vehicle that can be either autonomous or controlled by a human driver. Data is presented in the shape of: <ul> <li>sequences of JPEG images, one image per timestamp</li> <li>target vehicle state information for each timestamp</li> </ul> The dataset has been built on the CARLA simulator, thanks to Baidu Apollo and a Logitech G29 steering wheel for the autonomous and human drivings, respectively. The dataset consists of 520 scenes (260 pairs of mirrored scenarios) of 60 seconds each. The folders are organized as follows: <pre><code>. ├── ... ├── <driving mode> │ ├── <town> │ │ ├── <trial> │ │   │   └── ... │ │   └── ...  │   └── ... └── ...</code></pre> <ul> <li>driving mode: corresponds to the control modality of the target vehicle under test and can be either Baidu Apollo or manual driving;</li> <li>town: one of the five default maps in CARLA (e.g., Town01, Town02, etc);</li> <li>trial: 60 different trials per map, they differ in traffic and weather conditions (except Town04). Each trial records 60 seconds of simulation, logging 120 frames per video and an equal number of rows per CSV. In particular, each trial includes: <ul> <li>video: this folder groups the JPEG images;</li> <li>state_features.csv: reports the state information of the target vehicle for each frame;</li> <li>detection_features.csv: reports the 2D bounding box detections obtained from a pre-trained YOLOv7 detector.</li> </ul> </li> </ul>

opencc-by-4.0Mar 2023View details →

Driving environmental images from on-board camera and RTK-based localization for autonomous vehicles

This dataset contains images from an on-board frontal camera as well as information of the vehicle state, including: accurate global localization, heading, speed, acceleration and steering angle position. The data was captured from one of the vehicle instrumenhted vehicles of the <a href="https://autopia.car.upm-csic.es/">AUTOPIA group</a> at the surrounding of the <a href="https://www.car.upm-csic.es/">Centre for Automation and Robotics</a>, in Arganda del Rey (Madrid, Spain). The dataset is aimed to help users in developing and improving image-based road detection algorithms, as well as machine learning end-to-end approaches using driving information provided (steering angle, vehicle speed, etc.). The dataset has also been used to develop algorithms for online map adaptation based on computer vision. Please, refer to: "Artuñedo, A. (2020). Decision-making Strategies for Automated Driving in Urban Environments. In Springer Theses. Springer International Publishing. <a href="https://doi.org/10.1007/978-3-030-45905-5">https://doi.org/10.1007/978-3-030-45905-5</a>" Image information The sequence of images is provided in PNG format, and is named following the pattern: 'l_sA_nsB.png', where A and B are the second and nanosecond when the image was captured, so that A+B*1e-9 is the capture time. The images were captuted with the camera Bumblebee 2 0,8 MP Color FireWire 1394a 3,8mm (Sony ICX204) with the following technical features: <ul> <li>Sensor type: CCD</li> <li>Sensor format: 1/3"</li> <li>Pixel size 4.65 µm</li> <li>Focal length: 3.8mm</li> <li>Fames per second: 20 fps</li> <li>Resolution: 1024 x 768</li> </ul> Camera position The camera is placed at a height 1290 mm measured from the ground plane. With respect to the vehicle frame, it has a yaw of -2º and a pitch 5.5º, in order to focus the field of view in the road.   Vehicle localization The test vehicle is a prototype of Autonomous Vehicle developed by the <a href="https://autopia.car.upm-csic.es/">AUTOPIA reseach group</a> at the Centre for Automation and Robotics in Spain. Vehicle location mainly depends on a Trimble BX982 GNSS receiver using RTK. The GNSS antenna is installed near the rear axle of the vehicle, in the middle part of the vehicle. However, the location algorithm applies an EKF for combining GNSS measurements with different onboard sensors providing yaw rate, longitudinal acceleration and speed, steering wheel position and speed, etc. The vehicle localization file ('vehicle_data.csv') includes the following data at a sample rate of 20 Hz: <ul> <li>Time (s)</li> <li>UTM East (m)</li> <li>UTM North (m)</li> <li>Orientation (º)</li> <li>Speed (m/s)</li> <li>Acceleration(m/s^2)</li> <li>Steering wheel angle (º)</li> </ul> Disclaimer That the dataset comes "AS IS", without express or implied warranty and/or any liability exceeding mandatory statutory obligations. This especially applies to any obligations of care or indemnification in connection with the dataset. The dataset was created for our research purposes only and no quality assessment was done for the usage in products of any kind. We can therefore not guarantee for the correctness, completeness or reliability of the provided data set.

opencc-by-4.0Nov 2022View details →

READ: Large-Scale Neural Scene Rendering for \\ Autonomous Driving

With the development of advanced driver assistance systems~(ADAS) and autonomous vehicles, conducting experiments in various scenarios becomes an urgent need. Although having been capable of synthesizing photo-realistic street scenes, conventional image-to-image translation methods cannot produce coherent scenes due to the lack of 3D information. In this paper, a large-scale neural rendering method is proposed to synthesize the autonomous driving scene~(READ), which makes it possible to generate large-scale driving scenes in real time on a PC through a variety of sampling schemes. In order to effectively represent driving scenarios, we propose an ω rendering network to learn neural descriptors from sparse point clouds. Our model can not only synthesize photo-realistic driving scenes but also stitch and edit them. The promising experimental results show that our model performs well in large-scale driving scenarios.   Source Code: https://github.com/JOP-Lee/READ

opencc-by-4.0Jan 2020View details →

Livorno, Urban driving, autonomous speed adaptation approaching intersection

Scenario description: Test session for AD+connected car and connected cars approaching an intersection regulated by a "smart" traffic light with a stereocamera able to detect jaywalking. Session description: A "smart" traffic light (with stereocamera) sends SPaT and MAP messages describing the topology and the actual status of the traffic light. If a jaywalking occurrence is detected (pedestrian crossing with the red light) a DENM message is sent to warn vehicles of the presence of the pedestrian to other connected vehicles. The the hazard warning is also sent to the oneM2M platform on the cloud. An AD vehicle consumes the information and autonomously adapts its speed in order to cross the intersection without violating the traffic light phases, or even stop to avoid collision with pedestrian. The influence of other vehicles moving in front is considered too. Goal is to record data for the technical evaluation. Datasets descriptions: AUTOPILOT_Livorno_UrbanDriving_Vehicle_all: Data generated from the vehicle sensors This dataset refers to the vehicle datasets generated from the vehicle sensors during Urban Driving in Livorno. This includes the data coming from the CAN bus and GPS. It includes following kind of dataset: Vehicle: general data (speed, battery); PositioningSystem: data from GPS; VehicleDynamics: data about dynamic (acceleration...); LateralControl: steering and lane control data AUTOPILOT_Livorno_UrbanDriving_V2X_all: V2V messages during platooning sessions This dataset refers to the V2V messages exchanged between ITS stations (vehicles and RSUs) during the Urban Drining in Livorno. AUTOPILOT_Livorno_UrbanDriving_IoT_all: Data extracted from IoT oneM2M platform This dataset refers to messages exchanged by Urban Driving devices, applications and services across the oneM2M platform.

opencc-by-4.0Oct 2023View details →

A!ex autonomous car driving in manual mode dataset

Dataset from autonomous car A!ex (https://www.aalto.fi/en/news/invisible-powers-are-driving-cars-and-cranes-but-what-goes-on-behind-the-scenes-to-get-these) driven manually around Otaniemi.

opencc-by-4.0Jul 2023View details →

Dataset for "MarMot: Metamorphic Runtime Monitoring of Autonomous Driving Systems"

This is the dataset used in the paper "MarMot: Metamorphic Runtime Monitoring of Autonomous Driving Systems".

opencc-by-4.0May 2024View details →

Learn from Human-driving Accidents to Attack Autonomous Driving_Practical Traffic Flow Attacks on Decision-making

We provide some demo videos of attack patterns