{"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/following-high-level-navigation-instructions","title":"Following High-level Navigation Instructions on a Simulated Quadcopter with Imitation Learning","arxiv_id":"1806.00047","date":"2018-05-31","proceeding":null,"authors":["Valts Blukis","Nataly Brukhim","Andrew Bennett","Ross A. Knepper","Yoav Artzi"],"abstract":"We introduce a method for following high-level navigation instructions by\nmapping directly from images, instructions and pose estimates to continuous\nlow-level velocity commands for real-time control. The Grounded Semantic\nMapping Network (GSMN) is a fully-differentiable neural network architecture\nthat builds an explicit semantic map in the world reference frame by\nincorporating a pinhole camera projection model within the network. The\ninformation stored in the map is learned from experience, while the\nlocal-to-world transformation is computed explicitly. We train the model using\nDAggerFM, a modified variant of DAgger that trades tabular convergence\nguarantees for improved training speed and memory use. We test GSMN in virtual\nenvironments on a realistic quadcopter simulator and show that incorporating an\nexplicit mapping and grounding modules allows GSMN to outperform strong neural\nbaselines and almost reach an expert policy performance. Finally, we analyze\nthe learned map representations and show that using an explicit map leads to an\ninterpretable instruction-following model.","url_abs":"http://arxiv.org/abs/1806.00047v1","url_pdf":"http://arxiv.org/pdf/1806.00047v1.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":"following-high-level-navigation-instructions","repo_url":"https://github.com/lil-lab/drif","is_official":1,"mentioned_in_paper":0,"mentioned_in_github":0,"framework":"pytorch","reach":null}],"tasks":[{"task_slug":"imitation-learning","task_name":"Imitation Learning"},{"task_slug":"instruction-following","task_name":"Instruction Following"}],"methods":[{"method_slug":"speed","method_name":"SPEED"}],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1806.00047","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}