人形機器人之高低地面的最佳踏點規劃與平衡行走控制

Optimizing Step Position Planning and Balancing Walking Control of High-low Ground for Humanoid Robot

廖怡晴(Yi-Qing Liao)

指導教授 ： 劉智誠

淡江大學/工學院/電機工程學系人工智慧機器人碩士班/碩士(2022年)

https://doi.org/10.6846/TKU.2022.00759

摘要

本論文基於粒子群最佳化法(Particle Swarm Optimization, PSO)規劃最佳的踏步位置，讓人形機器人可在高低落差之地面行走。在視覺系統中，本論文使用RGB-D 攝影機來取得色彩資訊以及深度資訊，透過影像前處理提取所需之地面資訊，也可以降低計算量，辨識前方木板之位置，確認人形機器人適合的踏步範圍。在踏步規劃中，透過粒子群最佳化法與木板高度，在適合的踏步範圍內搜尋人形機器人之下一步最佳踏步位置，接著立即修改人形機器人下一步的踏步軌跡，來達成人形機器人在高低落差之地面行走。最後本論文針對更新踏步位置後的雙足行走加入平衡控制，增加人形機器人在踏步時的穩定。在實驗結果中，本論文提出基於粒子群最佳化法之人形機器人最佳踏步位置規劃，使人形機器人能在高低落差之地面行走，其有效提升即時行走規劃的可行性，且加入平衡控制讓人形機器人可以穩定前進。

關鍵字

人形機器人 ； 雙足行走 ； 粒子群最佳化 ； 平衡控制 ； 視覺感測 ； 踏步規劃

並列摘要

In this thesis, optimizing step position is planned based on Particle Swarm Optimization (PSO), so that the humanoid robot can walk on the high-low ground. In the vision system, this paper uses RGB-D cameras to obtain color information and depth information, and extracts the required ground information through image pre-processing, which can also reduce the amount of calculation, identify the position of the front board, and confirm the feasible stepping region of the humanoid robot. In the footstep planning, through the particle swarm optimization method and the height of the board, search for the next best footstep position of the humanoid robot within the feasible stepping region, and then immediately modify the next step trajectory of the humanoid robot to achieve the humanoid robot walking on the ground with high-low drop. Finally, this thesis increases balance control to the bipedal walking after updating the stepping position to increase the stability of the humanoid robot when stepping. In the experimental results, this thesis proposes the optimal step position planning of humanoid robot based on particle swarm optimization method, so that the humanoid robot can walk on the ground with high and low drop, this method can effectively improve the feasibility of real-time walking planning. And adding balance control makes the humanoid robot to walk stably.

並列關鍵字

Humanoid Robot ； Biped Walking ； Particle Swarm Optimization ； Balance Control ； Visual Perception ； Footstep Planning

參考文獻

[1] M. Kim, D. Lim and J. Park, "Online walking pattern generation for humanoid robot with compliant motion control", IEEE International Conference on Robotics and Automation (ICRA), Montreal, QC, Canada, May 20-24 2019, pp. 1417–1422.

[2] S. Kuindersma, R. Deits, M. Fallon, A. Valenzuela, H. Dai, F. Permenter, T. Koolen, P. Marion and R. Tedrake, "Optimization-based locomotion planning estimation and control design for the atlas humanoid robot," Autonomous Robots, vol. 40, no. 3, Jul. 2015, pp. 429-455.

[3] I. Chestnutt, J. Kuffner, K. Nishiwaki, and S. Kagami, "Planning biped navigation strategies in complex environments," IEEE-RAS International Conference on Humanoid Robots, Munich, Germany, Oct. 2003.

[4] A. Hornung, A. Dornbush, M. Likhachev and M. Bennewitz, "Anytime search-based footstep planning with suboptimality bounds", 12th IEEE-RAS International Conference on Humanoid Robots, Osaka, Japan, Nov. 29-Dec. 1 2012, pp. 674-679.

[5] A. Stumpf, S. Kohlbrecher, D. C. Conner, and O. V. Stryk, “Open source integrated 3D footstep planning framework for humanoid robots,” 16th IEEE-RAS International Conference on Humanoid Robots, Cancun, Mexico, Nov. 15-17 2016, pp. 938–945.