GAPLE: Generalizable Approaching Policy LEarning for Robotic Object Searching in Indoor Environment

Ye, Xin; Lin, Zhe; Lee, Joon-Young; Zhang, Jianming; Zheng, Shibin; Yang, Yezhou

doi:10.1109/lra.2019.2930426

Cited by 24 publications

(33 citation statements)

References 28 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…With the rise of deep learning techniques, some works apply more complex methods to find objects. In [17] and [18] the search problem is approached through a deep reinforcement learning model that learns action policies to reach the target object. In [17] the target object has been previously seen and the semantic segmentation and the depth information are computed in each robot observation.…”

Section: Related Workmentioning

confidence: 99%

“…In [17] and [18] the search problem is approached through a deep reinforcement learning model that learns action policies to reach the target object. In [17] the target object has been previously seen and the semantic segmentation and the depth information are computed in each robot observation. Druon et al [19] propose a visual navigation method based on the context information of previously detected objects to calculate the similarity to the target object.…”

Section: Related Workmentioning

confidence: 99%

See 1 more Smart Citation

Searching for Objects in Human Living Environments based on Relevant Inferred and Mined Priors

Hernandez

Durner

Gómez

et al. 2021

2021 European Conference on Mobile Robots (ECMR)

View full text Add to dashboard Cite

Service robots performing tasks in human environments constantly face changes due to the dynamic of the environments. Such robots need to reason about their surrounding for a better understanding of it. Besides, it is important to demonstrate capabilities that potential users would find useful, thus validating the development of such systems. One of these capabilities is to help a person to find what she or he is looking for. This mundane task of searching for an object is highly relevant in showing the non-expert user that a robot can understand the world. In this paper, we propose an efficient search strategy to find target objects that have not been seen before, based on the reasoning about in which scenes and with which objects they co-occur. Our method consists of an inference process based on a Conditional Random Field (CRF), that fuses the information about other previously detected objects, the semantic floor map, and the object-object/-room relations, to build a prediction map with the most promising locations for an unseen object. To validate our work, comparative experiments in simulated environments have been performed, demonstrating the efficiency of our proposed search strategy.

show abstract

Section: Related Workmentioning

confidence: 99%

Section: Related Workmentioning

confidence: 99%

Searching for Objects in Human Living Environments based on Relevant Inferred and Mined Priors

Hernandez

Durner

Gómez

et al. 2021

2021 European Conference on Mobile Robots (ECMR)

View full text Add to dashboard Cite

show abstract

“…With the rise of deep learning techniques, some works apply more complex methods to find objects. In the approaches of Ye et al [188,189], the search problem is addressed through a deep reinforcement learning model that learns action policies to reach the target object. In [188] the target object has been previously seen and the semantic segmentation and the depth information are computed in each robot observation.…”

Section: Related Workmentioning

confidence: 99%

“…In the approaches of Ye et al [188,189], the search problem is addressed through a deep reinforcement learning model that learns action policies to reach the target object. In [188] the target object has been previously seen and the semantic segmentation and the depth information are computed in each robot observation. Druon et al [27] propose a visual navigation method based on previously detected objects' context information to calculate the similarity to the target object.…”

Section: Related Workmentioning

confidence: 99%

Adding Uncertainty to an Object Detection System for Mobile Robots

Hernandez

Gómez

Crespo

et al. 2017

2017 6th International Conference on Space Mission Challenges for Information Technology (SMC-IT)

View full text Add to dashboard Cite

I would like to dedicate these lines to express my gratitude to all the people who have helped me in a way or another to achieve this goal that is so important to me. First, I would like to thank my tutor and director, Dr. Ramón Barber, for introducing me to the exciting world of research, for all his support, guidance, trust, and patience over these years of my Ph.D. studies. Thank you for giving me the opportunity to work with you. I also thank Dr. Oscar Martínez, for all the advice and technical talks and for allowing me to visit his laboratory in Sweden. I would like to thank the University Carlos III of Madrid and all the Robotics Lab members I met during these years. Thank you for your support, professionalism, and friendship; It is an honor to belong to this great research group.I would like to especially thank my lab partner Clara Gómez. Thank you for being such a great partner and friend over these years, for your support, professionalism, your dedication at work. Thank you for being my partner in conferences and research visits. Definitely, the Ph.D. would not have been the same without you.I would like to thank Dr. Zoltan Marton and his team for giving me the opportunity to work with them at the DLR institute in Germany. I would also like to thank Prof. Robert Babuška for welcoming me to his lab at the CTU in Prague. Especially thanks to Erik Derner for being a great lab partner, for our interesting discussions, for his professionalism and friendship. I also thank the people I met at Örebro University, especially Eduardo Gutiérrez; thank you for your support and friendship.Finally and most importantly, I would like to thank my family. Thank you for all your support, your love, and for always being there. Despite the more than 7000 km. that separate us, I am always with you, and I know that you are with me celebrating this achievement. I leave for last, the person I must thank for having embarked on this path. I want to thank my husband, Jesús. Thank you for supporting me and helping me accomplish this goal, for being my partner on this roller coaster of life, for always believing in me, and for having a word of encouragement when I felt like I could not continue. Without you, I am sure I would not have made it. For you all my love and my gratitude.To all my sincere thanks.

show abstract

“…Latest reinforcement learning studies on these tasks mainly focus on prior knowledge about targets [27,30,8]. Other studies focus on learning representation on the environment [46] or about the targets only implicitly [44]. These methods lead to insufficient understanding of the goal, and sample-inefficiency and generalization problems arise.…”

Section: Introductionmentioning

confidence: 99%

Goal-Aware Cross-Entropy for Multi-Target Reinforcement Learning

Kim¹,

Lee²,

Kim³

et al. 2021

Preprint

View full text Add to dashboard Cite

Learning in a multi-target environment without prior knowledge about the targets requires a large amount of samples and makes generalization difficult. To solve this problem, it is important to be able to discriminate targets through semantic understanding. In this paper, we propose goal-aware cross-entropy (GACE) loss, that can be utilized in a self-supervised way using auto-labeled goal states alongside reinforcement learning. Based on the loss, we then devise goal-discriminative attention networks (GDAN) which utilize the goal-relevant information to focus on the given instruction. We evaluate the proposed methods on visual navigation and robot arm manipulation tasks with multi-target environments and show that GDAN outperforms the state-of-the-art methods in terms of task success ratio, sample efficiency, and generalization. Additionally, qualitative analyses demonstrate that our proposed method can help the agent become aware of and focus on the given instruction clearly, promoting goal-directed behavior.

show abstract

GAPLE: Generalizable Approaching Policy LEarning for Robotic Object Searching in Indoor Environment

Cited by 24 publications

References 28 publications

Searching for Objects in Human Living Environments based on Relevant Inferred and Mined Priors

Searching for Objects in Human Living Environments based on Relevant Inferred and Mined Priors

Adding Uncertainty to an Object Detection System for Mobile Robots

Goal-Aware Cross-Entropy for Multi-Target Reinforcement Learning

Contact Info

Product

Resources

About