Whole-body trajectory optimization for non-periodic dynamic motions on quadrupedal systems

Radulescu, Andreea; Havoutis, Ioannis; Caldwell, Darwin G.; Semini, Claudio

doi:10.1109/icra.2017.7989623

Cited by 7 publications

(5 citation statements)

References 23 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This offline planning is not event-based in nature and, therefore, needs more real-time responsiveness to react to external disturbances. Optimizationbased methods can compute feasible fall recovery solutions without the need to handcraft trajectories directly [17,18]. Model Predictive Control (MPC) [19] enables a system to make current decisions while considering their impact on the future through predictive modeling, which has shown promise in recent research on legged locomotion.…”

Section: Related Workmentioning

confidence: 99%

Dynamic Fall Recovery Control for Legged Robots via Reinforcement Learning

Li,

Pang,

Bai

et al. 2024

Biomimetics

View full text Add to dashboard Cite

Falling is inevitable for legged robots when deployed in unstructured and unpredictable real-world scenarios, such as uneven terrain in the wild. Therefore, to recover dynamically from a fall without unintended termination of locomotion, the robot must possess the complex motor skills required for recovery maneuvers. However, this is exceptionally challenging for existing methods, since it involves multiple unspecified internal and external contacts. To go beyond the limitation of existing methods, we introduced a novel deep reinforcement learning framework to train a learning-based state estimator and a proprioceptive history policy for dynamic fall recovery under external disturbances. The proposed learning-based framework applies to different fall cases indoors and outdoors. Furthermore, we show that the learned fall recovery policies are hardware-feasible and can be implemented on real robots. The performance of the proposed approach is evaluated with extensive trials using a quadruped robot, which shows good effectiveness in recovering the robot after a fall on flat surfaces and grassland.

show abstract

Section: Related Workmentioning

confidence: 99%

Dynamic Fall Recovery Control for Legged Robots via Reinforcement Learning

Li,

Pang,

Bai

et al. 2024

Biomimetics

View full text Add to dashboard Cite

show abstract

“…This offline planning is not event-based in nature and, therefore, needs more real-time responsiveness to react to external disturbances. Optimization-based methods can compute feasible fall recovery solutions without the need to handcraft trajectories directly 18,19 . Model Predictive Control (MPC) 20 enables a system to make current decisions while considering their impact on the future through predictive modeling, which has shown promise in recent research on legged locomotion.…”

Section: Related Workmentioning

confidence: 99%

Learning Agility and Adaptive Legged Locomotion via Curricular Hindsight Reinforcement Learning

Li,

Pang,

Bai

et al. 2024

Preprint

View full text Add to dashboard Cite

Agile and adaptive maneuvers such as fall recovery, high-speed turning, and sprinting in the wild are challenging for legged systems. We propose a Curricular Hindsight Reinforcement Learning (CHRL) that learns an end-to-end tracking controller that achieves powerful agility and adaptation for the legged robot. The two key components are (i) a novel automatic curriculum strategy on task difficulty and (ii) a Hindsight Experience Replay strategy adapted to legged locomotion tasks. We demonstrated successful agile and adaptive locomotion on a real quadruped robot that performed fall recovery autonomously, coherent trotting, sustained outdoor running speeds up to 3.45 m/s, and tuning speeds up to 3.2 rad/s. This system produces adaptive behaviors responding to changing situations and unexpected disturbances on natural terrains like grass and dirt.

show abstract

“…In [14], the authors made use of whole-body optimization procedures developed for the execution of non-periodic tasks on quadrupeds. To this aim, the authors defined the cost function according to the desired goal allowing the robot to reach the final position from different scenarios and postures.…”

Section: State Of the Artmentioning

confidence: 99%

Design a Fall Recovery Strategy for a Wheel-Legged Quadruped Robot Using Stability Feature Space

Castano

Zhou

Tsagarakis

2019

2019 IEEE International Conference on Robotics and Biomimetics (ROBIO)

View full text Add to dashboard Cite

In this paper, we introduced a conceptual analysis to select stability features when performing predefined and precise motions on robots. By analyzing the different stable poses named features and the possible transitions towards different ones, the introduced concept allows to design more predictable and suitable motions when performing particular tasks. As an example of how the concept can be applied we use it on the fall recovery of the quadruped robot CENTAURO. This robot, which is equipped with a custom hybrid wheellegged mobility system, have good intrinsic stability as other quadrupeds. However, the characteristics of the rough terrains where it might be deployed require complex maneuvers to cope with possible strong disturbances. To prevent and more importantly recover from falls, realignment of postural responses will not be adequate, and effective recovery procedures should be developed. This paper introduces the details of how the presented conceptual analysis provides and an effective fall recovery routine for CENTAURO based on a state machine. The performance of the proposed approach is evaluated with extensive simulation trials using the dynamic model of the CENTAURO robot showing good effectiveness in recovering the robot after fall on flat and inclined surfaces.

show abstract

Whole-body trajectory optimization for non-periodic dynamic motions on quadrupedal systems

Cited by 7 publications

References 23 publications

Dynamic Fall Recovery Control for Legged Robots via Reinforcement Learning

Dynamic Fall Recovery Control for Legged Robots via Reinforcement Learning

Learning Agility and Adaptive Legged Locomotion via Curricular Hindsight Reinforcement Learning

Design a Fall Recovery Strategy for a Wheel-Legged Quadruped Robot Using Stability Feature Space

Contact Info

Product

Resources

About