Vision-based Autonomous Motion Control Method for Closed-chain Multi-legged Robot

doi:10.3901/JME.2024.19.062

Abstract

Abstract: Closed-chain multi-legged robots have strong environmental adaptability and high stability.They have the potential to move in complex outdoor environments.However, existing model-based control methods are difficult to use in actively obtaining terrain information and to walk autonomously in complex outdoor scenarios.Vision-based methods are able to acquire environmental information actively, but the robot has difficult in switching gaits autonomously according to the terrain.With regards to this, a vision-based closed-chain autonomous motion control method for multi-legged robots is proposed, and a mapping from visual images to gaits is established.The kinematic model and obstacle-crossing strategy are established through kinematic analysis and data fitting.The terrain passable domain delineation algorithm incorporating YOLACT++ is used to generate navigation paths in the passable domain based on the midpoint pixel sampling method.In order to validate the proposed method, a multi-terrain simulation scene and a real scene are built and experimented, and the results show that the closed-chained multi-legged robot is able to walk autonomously in unfamiliar environments and switch its gaits according to the terrain.

Key words: closed-chain multi-legged robot, image segmentation, obstacle surmounting strategy, motion planning

CLC Number:

TP242

DU Guofeng, ZHAO Meng, WU Jianxu, ZHANG Dong. Vision-based Autonomous Motion Control Method for Closed-chain Multi-legged Robot[J]. Journal of Mechanical Engineering, 2024, 60(19): 62-70.

References

[1] LEE Y H，LEE Y H，LEE H，et al. Development of a quadruped robot system with torque-controllable modular actuator unit[J]. IEEE Transactions on Industrial Electronics，2021，68(8)：7263-7273.
[2] VALSECCHI G，GRANDIA R，HUTTER M. Quadrupedal locomotion on uneven terrain with sensorized feet[J]. IEEE Robotics and Automation Letters，2020，5(2)：1548-1555.
[3] ZHANG C，AN H，WU C，et al. Contact force estimation method of legged-robot and its application in impedance control[J]. IEEE Access，2020，8：161175-161187.
[4] DING C，ZHOU L，LI Y，et al. A novel dynamic locomotion control method for quadruped robots running on rough terrains[J]. IEEE Access，2020，8：150435- 150446.
[5] FAHMI S，FOCCHI M，RADULESCU A，et al. STANCE：Locomotion adaptation over soft terrain[J]. IEEE Transactions on Robotics，2020，36(2)：443-457.
[6] MASTALLI C，HAVOUTIS I，FOCCHI M，et al. Motion planning for quadrupedal locomotion：Coupled planning，terrain mapping and whole-body control[J]. IEEE Transactions on Robotics，2022，36(6)：1635-1648.
[7] TENNAKOON E，PEYNOT T，ROBERTS J，et al. Probe-before-step walking strategy for multi-legged robots on terrain with risk of collapse[C]// IEEE International Conference on Robotics and Automation，Paris：IEEE，2020：5530-5536.
[8] DING L，XU P，LI Z，et al. Pressing and rubbing：Physics-informed features facilitate haptic terrain classification for legged robots[J]. IEEE Robotics and Automation Letters，2022，7(3)：5990-5997.
[9] HOMCHANTHANAKUL J，MANOONPONG P. Continuous online adaptation of bioinspired adaptive neuroendocrine control for autonomous walking robots[J]. IEEE Transactions on Neural Networks and Learning Systems，2022，33(5)：1833-1845.
[10] ZHONG B，ZHANG S，XU M，et al. On a CPG-based hexapod robot：AmphiHex-II with variable stiffness legs[J]. IEEE/ASME Transactions on Mechatronics，2018，23(2)：542-551.
[11] GANGAPURWALA S，GEISERT M，ORSOLINO R，et al. RLOC：terrain-aware legged locomotion using reinforcement learning and optimal control[J] IEEE Transactions on Robotics，2022，38(5)：2908-2927.
[12] RUDIN N，KOLVENBACH H，TSOUNIS V，et al. Cat-like jumping and landing of legged robots in low gravity using deep reinforcement learning[J]. IEEE Transactions on Robotics，2022，38(1)：317-328.
[13] BELLEGARDA G，IJSPEERT A. CPG-RL：learning central pattern generators for quadruped locomotion[J]. IEEE Robotics and Automation Letters，2022，7(4)：12547-12554.
[14] WANG J，HU C，ZHU Y. CPG-based hierarchical locomotion control for modular quadrupedal robots using deep reinforcement learning[J]. IEEE Robotics and Automation Letters，2021，4(6)：7193-7200.
[15] MAGANA O，BARASUOL V，CAMURRI M，et al. Fast and continuous foothold adaptation for dynamic locomotion through CNNs[J]. IEEE Robotics and Automation Letters，2019，4(2)：2140-2147.
[16] FAHMI S，BARASUOL V，ESTEBAN D，et al. ViTAL：vision-based terrain-aware locomotion for legged robots[J]. IEEE Transactions on Robotics，2023，39(2)：885-904.
[17] KIM D，CARBALLO D，CARLO J D，et al. Vision aided dynamic exploration of unstructured terrain with a small-scale quadruped robot[C]// IEEE International Conference on Robotics and Automation，Paris：IEEE，2020：2464-2470.
[18] BJELONIC M，HOMBERGER T，KOTTEGE N，et al. Autonomous navigation of hexapod robots with vision-based controller adaptation[C]// IEEE International Conference on Robotics and Automation，Singapore：IEEE，2017：5561-5568.
[19] FU Z，KUMAR A，AGARWAL A，et al. Coupling vision and proprioception for navigation of legged robots[C]// IEEE/CVF Conference on Computer Vision and Pattern Recognition，New Orleans，LA：IEEE，2022：17252-17262.
[20] DUDZIK T，CHIGNOLI M，BLEDT G，et al. Robust autonomous navigation of a small-scale quadruped robot in real-world environments[C]// International Conference on Intelligent Robots and Systems，Las Vegas，NV：IEEE，2020：3664-3671.
[21] RUAN Q，WU J，YAO Y. Design and analysis of a multi-legged robot with pitch adjustive units[J]. Chinese Journal of Mechanical Engineering. 2021，34(3)：17.
[22] BOLYA D，ZHOU C，XIAO F，et al. YOLACT++：better real-time instance segmentation[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence，2022，44(2)：1108-1121.