Learning robot gait stability using neural networks as sensory feedback function for Central Pattern Generators

Santos-Victor, José; Ijspeert, Auke Jan

doi:10.1109/iros.2013.6696353

Cited by 56 publications

(37 citation statements)

References 14 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…This framework is a derivation of the system previously presented in [5], where it was used to control a quadruped walking on rough terrain. It is worth noting that the same system can be used on different robots with a different number of degrees of freedom and for both rhythmic and discrete tasks, with only very minor modifications.…”

Section: Model-free Control Frameworkmentioning

confidence: 99%

“…In that case the low level control was performed by a central pattern generator, and all joints were coupled except the abduction/adduction. The system presented in this paper is the discrete movement version of the one presented in [5], equivalent to having the amplitudes and phase of each oscillator set to 0 and using only their offset. Thus the components controlling the joint positions here are called integrators instead of oscillators since they do not effectively oscillate but output discrete trajectories in a smooth way.…”

Section: Model-free Control Frameworkmentioning

confidence: 99%

See 1 more Smart Citation

Model-based and model-free approaches for postural control of a compliant humanoid robot using optical flow

Kieboom

Santos-Victor

Ijspeert

2013

2013 13th IEEE-RAS International Conference on Humanoid Robots (Humanoids)

Self Cite

View full text Add to dashboard Cite

Abstract-Vision is a very rich sensor with a proven critical role in the control of balance. However, it is widely underused for robotics postural control. This paper presents and compares two approaches, one model-based and one model-free, to ensure stability of the COMAN compliant humanoid robot standing on a moving platform. The model-based approach uses inverse kinematics, while the model-free one relies on a neural network as mapping between sensors and actuators. The sensory information is composed of proprioceptive cues (gyroscope) and visual cues, used separately or together. We present methods of using vision as sensory input without relying on a particular object or feature of the scene, but only on the optical flow. The performance of both approaches are compared systematically in a realistic robotics simulator, for different movements of the platform and using different sensory cues. We aim to see if vision can replace proprioceptive sensors or be fused with them to improve the performance of the stabilizing controller. While both model-based and model-free approaches successfully stabilize the robot, the model-free approach shows better overall performance. Preliminary results on the real COMAN robot are shown.

show abstract

Section: Model-free Control Frameworkmentioning

confidence: 99%

Section: Model-free Control Frameworkmentioning

confidence: 99%

Model-based and model-free approaches for postural control of a compliant humanoid robot using optical flow

Kieboom

Santos-Victor

Ijspeert

2013

2013 13th IEEE-RAS International Conference on Humanoid Robots (Humanoids)

Self Cite

View full text Add to dashboard Cite

show abstract

“…Hence, rapid computation is a key parameter for implementing reactive stability control. Stability control of legged robots has been already investigated including over rough terrain, when slipping, under external perturbations, and other precarious scenarios [1][2][3][6][7][8][9][10][11][12][13]. While their solutions are satisfactory for the implemented robots, their control algorithms are specific to their robots and may be hard to generalize for use with other multi-legged robots under different types of perturbations.…”

Section: Introductionmentioning

confidence: 99%

Foot Force Based Reactive Stability of Multi-Legged Robots to External Perturbations

Agheli

Nestinger

2015

J Intell Robot Syst

View full text Add to dashboard Cite

Many environments and scenarios contain rough and irregular terrain and are inaccessible or hazardous for humans. Robotic automation is preferred in lieu of placing humans at risk. Legged locomotion is more advantageous in traversing complex terrain but requires constant monitoring and correction to maintain system stability. This paper presents a multi-legged reactive stability control method for maintaining system stability under external perturbations. Assuming tumbling instability and sufficient friction to prevent slippage, the reactive stability control method is based solely on the measured foot forces normal to the contact surface, reducing computation time and sensor information. Under external perturbations, the reactive stability control method opts to either displace the CG or the foot contacts of the robot based on the measured foot force distribution. Details describing the reactive stability control method are discussed including algorithms and an implementation example. An experimental demonstration of the reactive stability control method is presented. The experiment was conducted on a hexapod robot platform retrofitted with a tiny computer and force sensitive resistors to measure the foot forces. The experimental M. Agheli ( ) · S. S. Nestinger results show that the presented reactive stability control strategy prevents the robot from tipping over under external perturbation.

show abstract

“…Many animals use CPGs to assist in gait generation [3]. CPGs have been used for quadruped robots [4] to generate joint trajectories incorporating sensory feedback to stabilize locomotion [5], [6]. Most such work, however, focuses on the use of CPGs to design steady state forward gaits rather than turning gaits.…”

Section: Introductionmentioning

confidence: 99%

Stabilizing task-based omnidirectional quadruped locomotion with Virtual Model Control

Kuhlman

Hays

Sofge

et al. 2015

2015 IEEE International Conference on Robotics and Automation (ICRA)

View full text Add to dashboard Cite

Quadruped locomotion offers significant advantages over wheeled locomotion for small mobile robots operating in challenging terrain. Central pattern generators (CPGs), as found in the neural circuitry of many animals, may be used to generate joint trajectories for quadruped robots. However, basic CPG-based trajectories do not explicitly consider ground contact constraints, a particular concern during turning maneuvers when foot slip is most likely to occur. An alternative approach proposed here is to use task-based CPGs such that ground contact constraints are enforced and foot velocities are explicitly controlled, resulting in stable omnidirectional locomotion. Further, incorporating Virtual Model Control with the task-based CPG trajectories improves the stability of the quadruped in hardware experiments.

show abstract

Learning robot gait stability using neural networks as sensory feedback function for Central Pattern Generators

Cited by 56 publications

References 14 publications

Model-based and model-free approaches for postural control of a compliant humanoid robot using optical flow

Model-based and model-free approaches for postural control of a compliant humanoid robot using optical flow

Foot Force Based Reactive Stability of Multi-Legged Robots to External Perturbations

Stabilizing task-based omnidirectional quadruped locomotion with Virtual Model Control

Contact Info

Product

Resources

About