RoboCup 2019 AdultSize Winner NimbRo: Deep Learning Perception, In-Walk Kick, Push Recovery, and Team Play Capabilities

Rodríguez, Diego; Farazi, Hafez; Ficht, Grzegorz; Pavlichenko, Dmytro; Brandenburger, André; Hosseini, Mojtaba; Kosenko, Oleg; Schreiber, Michael; Missura, Marcell; Behnke, Sven

doi:10.1007/978-3-030-35699-6_51

Cited by 12 publications

(10 citation statements)

References 16 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…The Defense Advanced Research Projects Agency (DARPA) of the United States has organized multiple well-known challenges since 2004, covering applications like autonomous driving [5], [6] and subterranean exploration, 2 among others. Other competitions, like RockIn [7] or RoboCup [8], propose indoor scenarios to pursue reproducibility and repeatability in robotics benchmarks. MBZIRC is also a recently created competition to foster outdoor challenges involving ground and aerial robots.…”

Section: Related Workmentioning

confidence: 99%

“…There is a special Component implementing an abstraction layer to operate each UAV, called UAV Abstraction Layer (UAL). UAL is an open-source 8 library [37] that was developed in our lab to ease the integration of different types of UAV autopilots. Thus, UAL offers common interfaces to provide UAV positioning and the possibility of sending basic commands to the autopilot, such as take-off, landing, position or velocity controllers and so on.…”

Section: ) Ualmentioning

confidence: 99%

See 1 more Smart Citation

Experimental Evaluation of a Team of Multiple Unmanned Aerial Vehicles for Cooperative Construction

et al. 2021

View full text Add to dashboard Cite

This article presents a team of multiple Unmanned Aerial Vehicles (UAVs) to perform cooperative missions for autonomous construction. In particular, the UAVs have to build a wall made of bricks that need to be picked and transported from different locations. First, we propose a novel architecture for multi-robot systems operating in outdoor and unstructured environments, where robustness and reliability play a key role. Then, we describe the design of our aerial platforms and grasping mechanisms to pick, transport and place bricks. The system was particularly developed for the Mohamed Bin Zayed International Robotics Challenge (MBZIRC), where Challenge 2 consisted of building a wall cooperatively with multiple UAVs. However, our approach is more general and extensible to other multi-UAV applications involving physical interaction, like package delivery. We present not only our results in the final stage of MBZIRC, but also our simulations and field experiments throughout the previous months to the competition, where we tuned our system and assessed its performance. Multirobot systems, unmanned aerial vehicles, construction. INDEX TERMS

show abstract

Section: Related Workmentioning

confidence: 99%

Section: ) Ualmentioning

confidence: 99%

Experimental Evaluation of a Team of Multiple Unmanned Aerial Vehicles for Cooperative Construction

et al. 2021

View full text Add to dashboard Cite

show abstract

“…Rodriguez et al [2] use a convolutional neural network to not only detect circular objects, but also other pretrained objects in real time. Circular objects can be easily detected and differentiated from non-circular ones based on the shape of their contour.…”

Section: Related Workmentioning

confidence: 99%

Visually Guided Balloon Popping with an Autonomous MAV at MBZIRC 2020

Beul¹,

Bultmann²,

Andre³

et al. 2020

Preprint

Self Cite

View full text Add to dashboard Cite

Visually guided control of micro aerial vehicles (MAV) demands for robust real-time perception, fast trajectory generation, and a capable flight platform. We present a fully autonomous MAV that is able to pop balloons, relying only on onboard sensing and computing. The system is evaluated with real robot experiments during the Mohamed Bin Zayed International Robotics Challenge (MBZIRC) 2020 where it showed its resilience and speed. In all three competition runs we were able to pop all five balloons in less than two minutes flight time with a single MAV.

show abstract

“…Mobile robot navigation typically requires a robot to traverse a series of static and dynamic obstacles in the environment to reach desired target poses, e.g., by walking with pedestrians on sidewalks. Traditional methods tackle this problem by processing raw sensor information (e.g., RGB images or laser scans) in order to construct local maps for path planners [1][2][3]. Traditional approaches, however, lose expressivity with the increment of uncertainty and complexity of the environments mainly because of computational limitations associated with high-dimensional systems and real-time constraints.…”

Section: Introductionmentioning

confidence: 99%

Mapless Humanoid Navigation Using Learned Latent Dynamics

Brandenburger¹,

Rodríguez²,

Behnke³

2021

Preprint

Self Cite

View full text Add to dashboard Cite

In this paper, we propose a novel Deep Reinforcement Learning approach to address the mapless navigation problem, in which the locomotion actions of a humanoid robot are taken online based on the knowledge encoded in learned models. Planning happens by generating open-loop trajectories in a learned latent space that captures the dynamics of the environment. Our planner considers visual (RGB images) and non-visual observations (e.g., attitude estimations). This confers the agent upon awareness not only of the scenario, but also of its own state. In addition, we incorporate a termination likelihood predictor model as an auxiliary loss function of the control policy, which enables the agent to anticipate terminal states of success and failure. In this manner, the sample efficiency of the approach for episodic tasks is increased. Our model is evaluated on the NimbRo-OP2X humanoid robot that navigates in scenes avoiding collisions efficiently in simulation and with the real hardware.

show abstract

RoboCup 2019 AdultSize Winner NimbRo: Deep Learning Perception, In-Walk Kick, Push Recovery, and Team Play Capabilities

Cited by 12 publications

References 16 publications

Experimental Evaluation of a Team of Multiple Unmanned Aerial Vehicles for Cooperative Construction

Experimental Evaluation of a Team of Multiple Unmanned Aerial Vehicles for Cooperative Construction

Visually Guided Balloon Popping with an Autonomous MAV at MBZIRC 2020

Mapless Humanoid Navigation Using Learned Latent Dynamics

Contact Info

Product

Resources

About