Bayesian Active Learning for Sim-to-Real Robotic Perception

Feng, Jianying; Lee, Jong-Seok; Durner, Maximilian; Triebel, Rudolph

doi:10.1109/iros47612.2022.9982175

Cited by 4 publications

(3 citation statements)

References 41 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…What motivates the given choice is the handling of the problem itself. The combination operation are to deal with having to combine the two different tasks per instance of an object detector, and the aggregation operation are to handle the multiple instances in an single image (Feng et al, 2022b). What remains is then the design of the information scores for both classification and regression tasks: U j,cls and U j,reg , respectively.…”

Section: Acquisition Functions For Query Generationmentioning

confidence: 99%

“…While deep ensembles (Lakshminarayanan et al, 2017) are another popular baseline, the suitability to active learning is limited due to the excessive training time. Here, the sampling strategy chosen from Feng et al (2022b), and therefore, the only difference between the baseline methods are the uncertainty estimates.…”

Section: Bayesian Active Learning For Field Roboticsmentioning

confidence: 99%

See 1 more Smart Citation

Virtual Reality via Object Pose Estimation and Active Learning: Realizing Telepresence Robots with Aerial Manipulation Capabilities

Lee¹,

Balachandran²,

Kondak³

et al. 2023

View full text Add to dashboard Cite

This paper presents a novel telepresence system for advancing aerial manipulation in dynamic and unstructured environments. The proposed system not only features a haptic device, but also a virtual reality (VR) interface that provides real-time 3D displays of the robot’s workspace as well as a haptic guidance to its remotely located operator. To realize this, multiple sensors, namely, a LiDAR, cameras, and IMUs are utilized. For processing of the acquired sensory data, pose estimation pipelines are devised for industrial objects of both known and unknown geometries. We further propose an active learning pipeline in order to increase the sample efficiency of a pipeline component that relies on a Deep Neural Network (DNN) based object detector. All these algorithms jointly address various challenges encountered during the execution of perception tasks in industrial scenarios. In the experiments, exhaustive ablation studies are provided to validate the proposed pipelines. Methodologically, these results commonly suggest how an awareness of the algorithms’ own failures and uncertainty (“introspection”) can be used to tackle the encountered problems. Moreover, outdoor experiments are conducted to evaluate the effectiveness of the overall system in enhancing aerial manipulation capabilities. In particular, with flight campaigns over days and nights, from spring to winter, and with different users and locations, we demonstrate over 70 robust executions of pick-andplace, force application and peg-in-hole tasks with the DLR cable-Suspended Aerial Manipulator (SAM). As a result, we show the viability of the proposed system in future industrial applications.

show abstract

Section: Acquisition Functions For Query Generationmentioning

confidence: 99%

Section: Bayesian Active Learning For Field Roboticsmentioning

confidence: 99%

Virtual Reality via Object Pose Estimation and Active Learning: Realizing Telepresence Robots with Aerial Manipulation Capabilities

Lee¹,

Balachandran²,

Kondak³

et al. 2023

View full text Add to dashboard Cite

show abstract

“…Most of these approaches aim to address the sample efficiency of RL methods during the learning process (Ghavamzadeh et al, 2015 ; Imani et al, 2018 ; Kamthe and Deisenroth, 2018 ). Meanwhile, Bayesian approaches have been used to quantify the discrepancies between real-world environments and simulations, facilitating sim-to-real policy transfer (Feng et al, 2022 ; Rothfuss et al, 2024 ). Other Bayesian approaches have also been developed in multi-agent and human-AI teaming to learn the intentions and preferences of teammates using partial data (Lin et al, 2024 ; Ravari et al, 2024a , b ; Zhang et al, 2024a , b ).…”

Section: Introductionmentioning

confidence: 99%

Bayesian reinforcement learning for navigation planning in unknown environments

Alali,

Imani

2024

Front. Artif. Intell.

View full text Add to dashboard Cite

This study focuses on a rescue mission problem, particularly enabling agents/robots to navigate efficiently in unknown environments. Technological advances, including manufacturing, sensing, and communication systems, have raised interest in using robots or drones for rescue operations. Effective rescue operations require quick identification of changes in the environment and/or locating the victims/injuries as soon as possible. Several techniques have been developed in recent years for autonomy in rescue missions, including motion planning, adaptive control, and more recently, reinforcement learning techniques. These techniques rely on full knowledge of the environment or the availability of simulators that can represent real environments during rescue operations. However, in practice, agents might have little or no information about the environment or the number or locations of injuries, preventing/limiting the application of most existing techniques. This study provides a probabilistic/Bayesian representation of the unknown environment, which jointly models the stochasticity in the agent's navigation and the environment uncertainty into a vector called the belief state. This belief state allows offline learning of the optimal Bayesian policy in an unknown environment without the need for any real data/interactions, which guarantees taking actions that are optimal given all available information. To address the large size of belief space, deep reinforcement learning is developed for computing an approximate Bayesian planning policy. The numerical experiments using different maze problems demonstrate the high performance of the proposed policy.

show abstract