Proprioceptive slip detection and state estimation of multi-legged robots in slippery scenarios

Abstract

Real-time slip detection and state estimation are crucial for locomotion control, facilitating posture adjustment and stability recovery of multi-legged robots moving on slippery terrain. However, existing proprioceptive methods rely on the fixed-contact assumption with fixed noise and suffer from low accuracy when multiple legs slip simultaneously. This paper proposes a novel proprioceptive approach for multi-legged robots moving in slippery scenarios to cope with slippage of multiple legs. In slip detection, the proprioceptive states of the robot are fed into a convolutional neural network to detect slip event(s) of the robot, enabling accurate identification of slipping legs even under simultaneous multi-leg slippage. For state estimation, an invariant extended Kalman filter is employed to fuse the motion information with the detected slip event(s) to obtain the robot state. By incorporating slip event(s) and foot velocity into the system motion equation of the filter, the proposed method better leverages leg odometry information and achieves more precise state estimation compared with existing methods. Simulations on a quadruped and a hexapod demonstrate the effectiveness and increased accuracy during multi-leg slippage. Experimental results for the quadruped robot show that the proposed approach achieves a 48% reduction in the root mean square error and a 47% reduction in the maximum error in velocity estimation under severe multi-leg slippage compared with the existing methods.

Graphical abstract

Keywords

multi-legged robot / slip detection / state estimation / simultaneous multi-leg slippage / proprioception

References

[1] Wooden D, Malchano M, Blankespoor K, Howardy A, Rizzi A A, Raibert M. Autonomous navigation for BigDog. In: Proceedings of 2010 IEEE International Conference on Robotics and Automation. Anchorage: IEEE, 2010, 4736–4741

[2] Raibert M , Blankespoor K , Nelson G , Playter R . BigDog, the rough-terrain quadruped robot. IFAC Proceedings Volumes, 2008, 41(2): 10822–10825

[3] Radosavovic I , Xiao T T , Zhang B K , Darrell T , Malik J , Sreenath K . Real-world humanoid locomotion with reinforcement learning. Science Robotics, 2024, 9(89): eadi9579

[4] Miki T , Lee J , Hwangbo J , Wellhausen L , Koltun V , Hutter M . Learning robust perceptive locomotion for quadrupedal robots in the wild. Science Robotics, 2022, 7(62): eabk2822

[5] Hoeller D , Rudin N , Sako D , Hutter M . ANYmal parkour: Learning agile navigation for quadrupedal robots. Science Robotics, 2024, 9(88): eadi7566

[6] Choi S , Ji G , Park J , Kim H , Mun J , Lee J H , Hwangbo J . Learning quadrupedal locomotion on deformable terrain. Science Robotics, 2023, 8(74): eade2256

[7] Nisticò Y , Fahmi S , Pallottino L , Semini C , Fink G . On slip detection for quadruped robots. Sensors, 2022, 22(8): 2967

[8] Teng S L, Mueller M W, Sreenath K. Legged robot state estimation in slippery environments using invariant extended kalman filter with velocity update. In: Proceedings of 2021 IEEE International Conference on Robotics and Automation. Xi’an: IEEE, 2021, 3104–3110

[9] Jenelten F , Hwangbo J , Tresoldi F , Bellicoso C D , Hutter M . Dynamic locomotion on slippery ground. IEEE Robotics and Automation Letters, 2019, 4(4): 4170–4176

[10] Catalano M G , Pollayil M J , Grioli G , Valsecchi G , Kolvenbach H , Hutter M , Bicchi A , Garabini M . Adaptive feet for quadrupedal walkers. IEEE Transactions on Robotics, 2022, 38(1): 302–316

[11] Park J , Kong D H , Park H W . Design of anti-skid foot with passive slip detection mechanism for conditional utilization of heterogeneous foot pads. IEEE Robotics and Automation Letters, 2019, 4(2): 1170–1177

[12] Bartsch S , Birnschein T , Römmermann M , Hilljegerdes J , Kühn D , Kirchner F . Development of the six-legged walking and climbing robot SpaceClimber. Journal of Field Robotics, 2012, 29(3): 506–532

[13] Spenko M J , Haynes G C , Saunders J A , Cutkosky M R , Rizzi A A , Full R J , Koditschek D E . Biologically inspired climbing with a hexapedal robot. Journal of Field Robotics, 2008, 25(4–5): 223–242

[14] Hu X H , Venkatesh A , Wan Y S , Zheng G L , Jawale N , Kaur N , Chen X , Birkmeyer P . Learning to detect slip through tactile estimation of the contact force field and its entropy properties. Mechatronics, 2024, 104: 103258

[15] Waters I , Wang L F , Jones D , Alazmani A , Culmer P . Incipient slip sensing for improved grasping in robot assisted surgery. IEEE Sensors Journal, 2022, 22(16): 16545–16554

[16] Waters I , Jones D , Alazmani A , Culmer P . Encouraging and detecting preferential incipient slip for use in slip prevention in robot-assisted surgery. Sensors, 2022, 22(20): 7956

[17] Burkhard N T , Cutkosky M R , Steger J R . Slip sensing for intelligent, improved grasping and retraction in robot-assisted surgery. IEEE Robotics and Automation Letters, 2018, 3(4): 4148–4155

[18] Focchi M, Barasuol V, Frigerio M, Caldwell D G, Semini C. Slip detection and recovery for quadruped robots. In: Bicchi A, Burgard W (eds). Robotics Research. Cham: Springer, 2018, 185–199

[19] Reinstein M , Hoffmann M . Dead reckoning in a dynamic quadruped robot based on multimodal proprioceptive sensory information. IEEE Transactions on Robotics, 2013, 29(2): 563–571

[20] Youm D, Oh H, Choi S, Kim H, Hwangbo J. Legged robot state estimation with invariant extended kalman filter using neural measurement network. 2024, arXiv preprint arXiv:2402.00366

[21] Buchanan R, Camurri M, Dellaert F, Fallon M. Learning inertial odometry for dynamic legged robot state estimation. In: Proceedings of 5th Conference on Robot Learning. London: PMLR, 2022, 1575–1584

[22] Hartley R , Ghaffari M , Eustice R M , Grizzle J W . Contact-aided invariant extended Kalman filtering for robot state estimation. The International Journal of Robotics Research, 2020, 39(4): 402–430

[23] Kim J H , Hong S , Ji G , Jeon S , Hwangbo J , Oh J H , Park H W . Legged robot state estimation with dynamic contact event information. IEEE Robotics and Automation Letters, 2021, 6(4): 6733–6740

[24] Wisth D , Camurri M , Fallon M . Robust legged robot state estimation using factor graph optimization. IEEE Robotics and Automation Letters, 2019, 4(4): 4507–4514

[25] Lu G L , Chen T , Rong X W , Zhang G T , Bi J , Cao J X , Jiang H , Li Y B . Whole‐body motion planning and control of a quadruped robot for challenging terrain. Journal of Field Robotics, 2023, 40(6): 1657–1677

[26] Neunert M , Stäuble M , Giftthaler M , Bellicoso C D , Carius J , Gehring C , Hutter M , Buchli J . Whole-body nonlinear model predictive control through contacts for quadrupeds. IEEE Robotics and Automation Letters, 2018, 3(3): 1458–1465

[27] Carius J , Ranftl R , Koltun V , Hutter M . Trajectory optimization with implicit hard contacts. IEEE Robotics and Automation Letters, 2018, 3(4): 3316–3323

[28] Sun P, Qiang J J, Qian L T, Luo X. Learning slip detection for agile locomotion of quadruped robots. In: 2023 IEEE International Conference on Robotics and Biomimetics (ROBIO). Koh Samui: IEEE, 2023, 1–6

[29] Lin T Y, Li T J, Tong W Z, Ghaffari M. Proprioceptive invariant robot state estimation. 2024, arXiv preprint arXiv:2311.04320

[30] Yang S R , Yang Q J , Zhu R , Zhang Z Y , Li C F , Liu H . State estimation of hydraulic quadruped robots using invariant-EKF and kinematics with neural networks. Neural Computing & Applications, 2024, 36(5): 2231–2244

[31] Qian L T, Lin W X, Chen J J, Xu Z H, Luo X. A novel bio-inspired optimal control strategy of heavy-duty robots considering leg momentum. In: Lan X G, Mei X S, Jiang C G, Zhao F, Tian Z Q (eds). Intelligent Robotics and Applications. Singapore: Springer Nature, 2025, 156–170

RIGHTS & PERMISSIONS

Higher Education Press