Dynamic Locomotion Through Online Nonlinear Motion Optimization for Quadrupedal Robots

Bellicoso, C. Dario; Jenelten, Fabian; Gehring, Christian; Hutter, Marco

doi:10.1109/lra.2018.2794620

Cited by 165 publications

(207 citation statements)

References 21 publications

Supporting

Mentioning

192

Contrasting

Unclassified

Order By: Relevance

“…Different approaches on legged robots apply simple CPG approaches for temporal coordination [Kalakrishnan et al, 2010;Ijspeert, 2014]. While these excel at spatial coordination in rough terrain, recently Bellicoso et al, [2018] argued that-for their case of a quadruped robot-climbing and changing environmental situations also require adaptation on the temporal level which appeared difficult when dealing with fixed gaits.…”

Section: Control Of the Movement Of An Individual Legmentioning

confidence: 99%

Decentralized Control of Insect Walking – a simple neural network explains a wide range of behavioral and neurophysiological results

Schilling

Cruse

2019

Preprint

View full text Add to dashboard Cite

AbstractControl of walking with six or more legs in an unpredictable environment is a challenging task, as many degrees of freedom have to be coordinated. Generally, solutions are proposed that rely on (sensory-modulated) CPGs, mainly based on data from neurophysiological studies. Here, we are introducing a sensor based controller operating on artificial neurons, being applied to a (simulated) hexapod robot with a morphology adapted to Carausius morosus. We show that such a decentralized solution leads to adaptive behavior when facing uncertain environments which we demonstrate for a large range of behaviors – slow and fast walking, forward and backward walking, negotiation of curves and walking on a treadmill with various treatment of individual legs. This approach can as well account for these neurophysiological results without relying on explicit CPG-like structures, but can be complemented with these for very fast walking.

show abstract

Section: Control Of the Movement Of An Individual Legmentioning

confidence: 99%

Decentralized Control of Insect Walking – a simple neural network explains a wide range of behavioral and neurophysiological results

Schilling

Cruse

2019

Preprint

View full text Add to dashboard Cite

show abstract

“…Kalakrishnan et al [11] demonstrated robust locomotion over challenging terrain with a quadrupedal robot: to date this remains the state-of-the-art for rough terrain locomotion. Recently, Bellicoso et al [12] demonstrated dynamic gaits, smooth transitions between them, and agile outdoor locomotion with a similar controller design. Yet despite their attractive properties, modular designs have limitations.…”

Section: Introductionmentioning

confidence: 99%

Learning agile and dynamic motor skills for legged robots

et al. 2019

Self Cite

View full text Add to dashboard Cite

Legged robots pose one of the greatest challenges in robotics. Dynamic and agile maneuvers of animals cannot be imitated by existing methods that are crafted by humans. A compelling alternative is reinforcement learning, which requires minimal craftsmanship and promotes the natural evolution of a control policy. However, so far, reinforcement learning research for legged robots is mainly limited to simulation, and only few and comparably simple examples have been deployed on real systems. The primary reason is that training with real robots, particularly with dynamically balancing systems, is complicated and expensive. In the present work, we report a new method for training a neural network policy in simulation and transferring it to a state-of-the-art legged system, thereby we leverage fast, automated, and cost-effective data generation schemes. The approach is applied to the ANYmal robot, a sophisticated medium-dog-sized quadrupedal system. Using policies trained in simulation, the quadrupedal machine achieves locomotion skills that go beyond what had been achieved with prior methods: ANYmal is capable of precisely and energy-efficiently following high-level body velocity commands, running faster than ever before, and recovering from falling even in complex configurations.

show abstract

“…Fortunately, research in traditional legged locomotion offers solutions to bridge this gap. The quadrupedal robot ANYmal (without wheels) performs highly dynamic motions using MPC [16], [17] and TO [18], [19] approaches. Impressive results are shown by MIT Cheetah, which performs blind locomotion over stairs [20] and jumps onto a desk with the height of 0.76 m [21].…”

Section: A Related Workmentioning

confidence: 99%

“…The gravito-inertial wrench [28] is given by f gi = m · (g −r COM ) ∈ R 3 and m gi = m · r COM × (g −r COM ) −l COM ∈ R 3 , where m is the mass of the robot, l COM ∈ R 3 is the angular momentum of the COM, and g ∈ R 3 is the gravity vector. In contrast to [11], [16], the line coefficients d(t) = [p(t) q(t) r(t)] T that describe an edge of a support polygon depend on time t, since the contact points of wheeled-legged robots continue to move even when a leg is in contact, unlike conventional legged robots. The ZMP inequality constraint is sampled over the time horizon t f with a fixed sampling time ∆t = t k − t k−1 .…”

Section: Generalization Of Zmp Inequality Constraintmentioning

confidence: 99%

“…Each QP problem including the optimization problem in Section III is solved using QuadProg++ [33], which internally implements the Goldfarb-Idnani active-set method [34]. To maintain a positive definite Hessian Q in (1) and to ensure the convexity of the resulting QP problem, a regularizer ρ is added to its diagonal elements, e.g., ρ = 10 −8 as in [16].…”

Section: A Implementationmentioning

confidence: 99%

See 1 more Smart Citation

Rolling in the Deep – Hybrid Locomotion for Wheeled-Legged Robots Using Online Trajectory Optimization

Bjelonic

Sankar

Bellicoso

et al. 2020

IEEE Robot. Autom. Lett.

Self Cite

View full text Add to dashboard Cite

Wheeled-legged robots have the potential for highly agile and versatile locomotion. The combination of legs and wheels might be a solution for any real-world application requiring rapid, and long-distance mobility skills on challenging terrain. In this paper, we present an online trajectory optimization framework for wheeled quadrupedal robots capable of executing hybrid walking-driving locomotion strategies. By breaking down the optimization problem into a wheel and base trajectory planning, locomotion planning for high dimensional wheeled-legged robots becomes more tractable, can be solved in real-time on-board in a model predictive control fashion, and becomes robust against unpredicted disturbances. The reference motions are tracked by a hierarchical whole-body controller that sends torque commands to the robot. Our approach is verified on a quadrupedal robot that is fully torquecontrolled, including the non-steerable wheels attached to its legs. The robot performs hybrid locomotion with different gait sequences on flat and rough terrain. In addition, we validated the robotic platform at the Defense Advanced Research Projects Agency (DARPA) Subterranean Challenge, where the robot rapidly maps, navigates, and explores dynamic underground environments.

show abstract

Dynamic Locomotion Through Online Nonlinear Motion Optimization for Quadrupedal Robots

Cited by 165 publications

References 21 publications

Decentralized Control of Insect Walking – a simple neural network explains a wide range of behavioral and neurophysiological results

Decentralized Control of Insect Walking – a simple neural network explains a wide range of behavioral and neurophysiological results

Learning agile and dynamic motor skills for legged robots

Rolling in the Deep – Hybrid Locomotion for Wheeled-Legged Robots Using Online Trajectory Optimization

Contact Info

Product

Resources

About