{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/deep-active-localization","title":"Deep Active Localization","arxiv_id":"1903.01669","date":"2019-03-05","proceeding":null,"authors":["Sai Krishna","Keehong Seo","Dhaivat Bhatt","Vincent Mai","Krishna Murthy","Liam Paull"],"abstract":"Active localization is the problem of generating robot actions that allow it\nto maximally disambiguate its pose within a reference map. Traditional\napproaches to this use an information-theoretic criterion for action selection\nand hand-crafted perceptual models. In this work we propose an end-to-end\ndifferentiable method for learning to take informative actions that is\ntrainable entirely in simulation and then transferable to real robot hardware\nwith zero refinement. The system is composed of two modules: a convolutional\nneural network for perception, and a deep reinforcement learned planning\nmodule. We introduce a multi-scale approach to the learned perceptual model\nsince the accuracy needed to perform action selection with reinforcement\nlearning is much less than the accuracy needed for robot control. We\ndemonstrate that the resulting system outperforms using the traditional\napproach for either perception or planning. We also demonstrate our approaches\nrobustness to different map configurations and other nuisance parameters\nthrough the use of domain randomization in training. The code is also\ncompatible with the OpenAI gym framework, as well as the Gazebo simulator.","url_abs":"http://arxiv.org/abs/1903.01669v1","url_pdf":"http://arxiv.org/pdf/1903.01669v1.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"deep-active-localization","repo_url":"https://github.com/montrealrobotics/dal","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"pytorch","reach":null}],"tasks":[{"task_slug":"openai-gym","task_name":"OpenAI Gym"},{"task_slug":"reinforcement-learning","task_name":"Reinforcement Learning"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":null,"mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}