DR-SPAAM: A Spatial-Attention and Auto-regressive Model for Person Detection in 2D Range Data

Jia, Dan; Hermans, Alexander; Leibe, Bastian

doi:10.1109/iros45743.2020.9341689

Cited by 31 publications

(21 citation statements)

References 23 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We also notice that state-of-the-art deep learning approaches [11], [24] clearly outperform the earlier, classical method [22] that only considers a single 2D laser scan, which is in line with results in the elderly care scenario of Beyer et al [11]. However, the method by Leigh et al [23], which is a non-DL approach but like DR-SPAAM [24] considers temporal information 2 , without retraining comes surprisingly close to both DL methods (−2% and −4% in peak-F1 over the combined sequences); for resource-constrained robots equipped with a 2D safety laser, this could be a reasonable choice over e.g. a resource-hungry DL-based detector if only medium-level detection accuracy is required.…”

Section: A 2d Laser Detectorsmentioning

confidence: 86%

“…Very recently, a more efficient version of DROW3x, dubbed DR-SPAAM [24], has been proposed, which we also consider in our evaluation. By replacing the expensive voting grid with a non-maximum suppression step and by using a recurrent update scheme instead of explicit odometrybased multi-scan alignment for temporal fusion, it achieves significantly higher frame rates and is suitable also for deployment on embedded platforms (e.g.…”

Section: A 2d Laser Detectorsmentioning

confidence: 99%

“…In the implementation by[24] that we used, DROW3x is run in a single-frame configuration (T=1) with non-maximum suppression instead of a voting grid to achieve fast real-time frame rates. It outperformed the original implementation[11] in terms of AP on our dataset 3.…”

mentioning

confidence: 99%

See 2 more Smart Citations

Cross-Modal Analysis of Human Detection for Robotics: An Industrial Case Study

Linder¹,

Vaškevičius²,

Schirmer³

et al. 2021

Preprint

View full text Add to dashboard Cite

Advances in sensing and learning algorithms have led to increasingly mature solutions for human detection by robots, particularly in selected use-cases such as pedestrian detection for self-driving cars or close-range person detection in consumer settings. Despite this progress, the simple question which sensor-algorithm combination is best suited for a person detection task at hand? remains hard to answer. In this paper, we tackle this issue by conducting a systematic cross-modal analysis of sensor-algorithm combinations typically used in robotics. We compare the performance of state-of-the-art person detectors for 2D range data, 3D lidar, and RGB-D data as well as selected combinations thereof in a challenging industrial use-case.We further address the related problems of data scarcity in the industrial target domain, and that recent research on human detection in 3D point clouds has mostly focused on autonomous driving scenarios. To leverage these methodological advances for robotics applications, we utilize a simple, yet effective multisensor transfer learning strategy by extending a strong imagebased RGB-D detector to provide cross-modal supervision for lidar detectors in the form of weak 3D bounding box labels.Our results show a large variance among the different approaches in terms of detection performance, generalization, frame rates and computational requirements. As our use-case contains difficulties representative for a wide range of service robot applications, we believe that these results point to relevant open challenges for further research and provide valuable support to practitioners for the design of their robot system.

show abstract

Section: A 2d Laser Detectorsmentioning

confidence: 86%

Section: A 2d Laser Detectorsmentioning

confidence: 99%

See 1 more Smart Citation

Cross-Modal Analysis of Human Detection for Robotics: An Industrial Case Study

Linder¹,

Vaškevičius²,

Schirmer³

et al. 2021

Preprint

View full text Add to dashboard Cite

show abstract

“…The path is about 20 meters long and the robot's controller is initialized with a linear DS to reach a goal 20 meters away from the onset position. Pedestrians are detected with a camera and LIDAR-based tracker published in [37] as can be seen in Fig. 20.…”

Section: Proof Of Concept: Outdoor Environmentmentioning

confidence: 99%

Avoiding Dense and Dynamic Obstacles in Enclosed Spaces: Application to Moving in Crowds

Huber,

Slotine,

Billard

2021

Preprint

View full text Add to dashboard Cite

This paper presents a closed-form approach to constrain a flow within a given volume and around objects. The flow is guaranteed to converge and to stop at a single fixed point. We show that the obstacle avoidance problem can be inverted to enforce that the flow remains enclosed within a volume defined by a polygonal surface. We formally guarantee that such a flow will never contact the boundaries of the enclosing volume and obstacles, and will asymptotically converge towards an attractor. We further create smooth motion fields around obstacles with edges (e.g. tables). Both obstacles and enclosures may be timevarying, i.e. moving, expanding and shrinking. The technique enables a robot to navigate within an enclosed corridor while avoiding static and moving obstacles. It was applied on an autonomous robot (QOLO) in a static complex indoor environment, and also tested in simulations with dense crowds. The final proof of concept was performed in an outdoor environment in Lausanne. The QOLO-robot successfully traversed a marketplace in the center of town in presence of a diverse crowd with a non-uniform motion pattern.

show abstract

“…Most recent developments [5,6,21,22] resorted to deeplearning techniques by applying 1D CNNs to range data, and no longer require motion-based filtering. The first of these approaches is the DROW detector [5].…”

Section: Related Workmentioning

confidence: 99%

2D vs. 3D LiDAR-based Person Detection on Mobile Robots

Jia¹,

Hermans²,

Leibe³

2021

Preprint

Self Cite

View full text Add to dashboard Cite

Person detection is a crucial task for mobile robots navigating in human-populated environments and LiDAR sensors are promising for this task, given their accurate depth measurements and large field of view. This paper studies existing LiDAR-based person detectors with a particular focus on mobile robot scenarios (e.g. service robot or social robot), where persons are observed more frequently and in much closer ranges, compared to the driving scenarios. We conduct a series of experiments, using the recently released JackRabbot dataset and the state-of-the-art detectors based on 3D or 2D LiDAR sensors (CenterPoint and DR-SPAAM respectively). These experiments revolve around the domain gap between driving and mobile robot scenarios, as well as the modality gap between 3D and 2D LiDAR sensors. For the domain gap, we aim to understand if detectors pretrained on driving datasets can achieve good performance on the mobile robot scenarios, for which there are currently no trained models readily available. For the modality gap, we compare detectors that use 3D or 2D LiDAR, from various aspects, including performance, runtime, localization accuracy, robustness to range and crowdedness. The results from our experiments provide practical insights into LiDAR-based person detection and facilitate informed decisions for relevant mobile robot designs and applications.

show abstract

DR-SPAAM: A Spatial-Attention and Auto-regressive Model for Person Detection in 2D Range Data

Cited by 31 publications

References 23 publications

Cross-Modal Analysis of Human Detection for Robotics: An Industrial Case Study

Cross-Modal Analysis of Human Detection for Robotics: An Industrial Case Study

Avoiding Dense and Dynamic Obstacles in Enclosed Spaces: Application to Moving in Crowds

2D vs. 3D LiDAR-based Person Detection on Mobile Robots

Contact Info

Product

Resources

About