This repository is a placeholder for an upcoming conference submission. The main purpose of this repository is to provide a convenient way for reviewers and other interested parties to access our work once it's updated.
Conference: 11th European Conference on Mobile Robots (2023)
Abstract: Autonomous industrial mobile robots need advanced perception capabilities to operate safely and human-compliantly in shared working environments. To achieve this high-level understanding of the mobile robots' surroundings, this paper investigates Multi-Task Learning approaches to process multiple tasks simultaneously and potentially improve the generalization performance. Our work alleviates the scarcity of datasets that are relevant for industrial settings by introducing and making publicly available a simulated warehouse dataset covering semantic segmentation, depth estimation and surface normals estimation tasks. We collect and examine numerous MTL task-balancing techniques for industrial mobile robot perception. Our experiments show that the performance of those approaches is very dependent on the considered dataset, which further highlights the value of introducing new relevant datasets.
To access and download the dataset this private link can be used: WarehouseSIM dataset --- In case of acceptance, a public url linking the dataset to this paper will be created.
Affiliation details will be provided upon submission.
- Paper Submission Deadline:
17 April 20231 May 2023 - Author Notification: 15 June 2023
- Camera-Ready Deadline: 30 June 2023
- Conference Dates: 4-7 September 2023
This repository will include:
- Source code
- Link to simulated dataset
- Samples of dataset
- Pre-trained models
For any questions or comments, please feel free to reach out:
- Dimitrios Arapis dimara@dtu.dk