MAVNet: An Effective Semantic Segmentation Micro-Network for MAV-Based Tasks

Nguyen, Ty; Shivakumar, Shreyas S.; Miller, Ian D.; Keller, James F.; Lee, Elijah S.; Zhou, Alex; Özaslan, Tolga; Loianno, Giuseppe; Harwood, Joseph H.; Wozencraft, Jennifer M.; Taylor, Camillo J.; Kumar, Vijay

doi:10.1109/lra.2019.2928734

Cited by 39 publications

(24 citation statements)

References 34 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We achieve artifact recognition using the ERFNet [19] deep learning based semantic segmentation network. We selected this network architecture because of its compact and efficient design, allowing fast inference on our compute hardware [20].…”

Section: Object Detection a Vision Based Detectionmentioning

confidence: 99%

Mine Tunnel Exploration Using Multiple Quadrupedal Robots

Miller

Cohen

Kulkarni

et al. 2020

IEEE Robot. Autom. Lett.

Self Cite

View full text Add to dashboard Cite

Robotic exploration of underground environments is a particularly challenging problem due to communication, endurance, and traversability constraints which necessitate high degrees of autonomy and agility. These challenges are further enhanced by the need to minimize human intervention for practical applications. While legged robots have the ability to traverse extremely challenging terrain, they also engender further inherent challenges for planning, estimation, and control.In this work, we describe a fully autonomous system for multi-robot mine exploration and mapping using legged quadrupeds, as well as a distributed database mesh networking system for reporting data. In addition, we show results from the DARPA Subterranean Challenge (SubT) Tunnel Circuit demonstrating localization of artifacts after traversals of hundreds of meters. To our knowledge, these experiments represent the first fully autonomous exploration of an unknown GNSS-denied environment undertaken by legged robots.

show abstract

Section: Object Detection a Vision Based Detectionmentioning

confidence: 99%

Mine Tunnel Exploration Using Multiple Quadrupedal Robots

Miller

Cohen

Kulkarni

et al. 2020

IEEE Robot. Autom. Lett.

Self Cite

View full text Add to dashboard Cite

show abstract

“…This work simulates the perimeter defense based on Unmanned Aerial Vehicles (UAVs). UAVs are deployed in various space such as power plant [11], penstock [12], forest [13], or disaster sites [14], which are good perimeter defense applications.…”

Section: Theory Practicementioning

confidence: 99%

Defending a Perimeter from a Ground Intruder Using an Aerial Defender: Theory and Practice

Lee¹,

Shishika²,

Loianno³

et al. 2021

Preprint

Self Cite

View full text Add to dashboard Cite

The perimeter defense game has received interest in recent years as a variant of the pursuit-evasion game. A number of previous works have solved this game to obtain the optimal strategies for defender and intruder, but the derived theory considers the players as point particles with first-order assumptions. In this work, we aim to apply the theory derived from the perimeter defense problem to robots with realistic models of actuation and sensing and observe performance discrepancy in relaxing the first-order assumptions. In particular, we focus on the hemisphere perimeter defense problem where a ground intruder tries to reach the base of a hemisphere while an aerial defender constrained to move on the hemisphere aims to capture the intruder. The transition from theory to practice is detailed, and the designed system is simulated in Gazebo. Two metrics for parametric analysis and comparative study are proposed to evaluate the performance discrepancy.

show abstract

“…For semantic image segmentation, efficient architectures for inference onboard UAVs have mostly been proposed for specific applications, such as UAV tracking and visual inspection [1] or weed detection for autonomous farming [12]. The DeepLab v3+ architecture [13] shows state-of-the-art performance on large, general datasets while including elements of the MobileNet architectures such as depthwiseseparable convolutions for efficient computation.…”

Section: Related Workmentioning

confidence: 99%

“…The combination of all these sensor modalities enables a complete and detailed interpretation of the environment. A semantic map aids inspection tasks [1], perception-aware path planning [2], and increases robustness and accuracy of simultaneous localization and mapping (SLAM) through the exclusion of dynamic objects during scan matching [3].…”

Section: Introductionmentioning

confidence: 99%

Real-Time Multi-Modal Semantic Fusion on Unmanned Aerial Vehicles

Bultmann¹,

Quenzel²,

Behnke³

2021

Preprint

View full text Add to dashboard Cite

Unmanned aerial vehicles (UAVs) equipped with multiple complementary sensors have tremendous potential for fast autonomous or remote-controlled semantic scene analysis, e.g., for disaster examination.In this work, we propose a UAV system for real-time semantic inference and fusion of multiple sensor modalities. Semantic segmentation of LiDAR scans and RGB images, as well as object detection on RGB and thermal images, run online onboard the UAV computer using lightweight CNN architectures and embedded inference accelerators. We follow a late fusion approach where semantic information from multiple modalities augments 3D point clouds and image segmentation masks while also generating an allocentric semantic map.Our system provides augmented semantic images and point clouds with ≈ 9 Hz. We evaluate the integrated system in realworld experiments in an urban environment. * : equal contribution.

show abstract

MAVNet: An Effective Semantic Segmentation Micro-Network for MAV-Based Tasks

Cited by 39 publications

References 34 publications

Mine Tunnel Exploration Using Multiple Quadrupedal Robots

Mine Tunnel Exploration Using Multiple Quadrupedal Robots

Defending a Perimeter from a Ground Intruder Using an Aerial Defender: Theory and Practice

Real-Time Multi-Modal Semantic Fusion on Unmanned Aerial Vehicles

Contact Info

Product

Resources

About