Real life industrial data for environments of robotic cells represented in Pointcloud and Octomap - Dataset

Real life industrial data for environments of robotic cells represented in Pointcloud and Octomap

Dynamic collision map for robotic planning environment extracted from a high accuracy pointcloud.

Overview

This is a dataset to demonstrate the extracted octomap from a precise pointcloud from a real life shipyard. The dataset has been developed for the needs of the High-payload robots in shared space with humans use case during the integration and validation activities in BIS shipyard for the Mari4_YARD EU Project in 2024.

Dataset Details

The data has been generated in order to develop a Digital Twin, which is a replica of the physical world, and provides data to the robot, such as collision objects for the planner to construct a collision free trajectory and at the same time shares important information to the operator for the robots future trajectories and information of the state of the process, such as the state of the detection and the position of the detected part for the bin picking operations. This digital twin categorized as dynamic, due to the absence of 3D CAD models and can be alternated real time if such thing is needed. This dataset is an example of an Octomap generated by a pointcloud extracted from a sensor with accuracy of millimetres. The translation of the pointcloud to an Octomap has been achieved through an algorithm that takes as input a pointcloud and after an optimization process it generated the corresponding collision map.

Description

The scope of this dataset is to help anyone that wants to experiment with the development of collision maps, SLAM (simultaneous localization and mapping) operations or in any way that this dataset could help with a real life data extracted from a real life industrial environment.