The foot force of the legged robot plays a decisive role in the balance of the fuselage. Especially when walking on the irregular road surface and the less rigid road surface, the change of the foot end support force will change the attitude of the robot body, which will affect the stability. In addition, the change of foot force is also closely related to the flexibility of the movement of the leg of the robot. When the movement of the foot end has a transition from free space to constrained space, only the position control will not meet the requirements of the leg for the flexibility of the movement. This paper addresses the design of virtual sensors for terrain adaptation developed with the aims of simplifying the hardware of the legged robot or increasing the reliability of the sensorial information available. The virtual force sensor (VFS) is developed based on particle swarm optimization (PSO) BP neural network and can estimate the forces exerted by the feet from data extracted from joint-position, joint-velocity, and joint-torque, which are mandatory in all robotic systems. The force estimates are used to detect foot/ground contact. Several simulations carried out with the hexapod robot are reported to prove the efficacy of this method. This method simplifies the hardware of the robot, reduces design, construction and maintenance costs while enhancing the robustness of the robot and the reliability of its behavior.
Virtual force sensor Particle swarm optimization algorithm BP neural network Legged robot
This is a preview of subscription content, log in to check access.
This research was supported, in part, by the National Natural Science Foundation of China (No. 51875393) and by the China Advance Research for Manned Space Project (No. 030601).
Ding, L.: Foot-terrain interaction mechanics for legged robots: modeling and experimental validation. Int. J. Robot. Res. 32(13), 1585–1606 (2013)CrossRefGoogle Scholar
Zhang, J.P.: An ATPSO-BP neural network modeling and its application in mechanical property prediction. Comput. Mater. Sci. 163, 262–266 (2019)CrossRefGoogle Scholar
Wilcox, B.H.: ATHLETE: a cargo handling and manipulation robot for the moon. J. Field Robot. 24(5), 421–434 (2007)CrossRefGoogle Scholar
Estremera, J.: Neural virtual sensors for terrain adaptation of walking machines. J. Robot. Syst. 22(6), 299–311 (2005)CrossRefGoogle Scholar
Sharf, I.: Identification of contact dynamics parameters for stiff robotic payloads. IEEE Trans. Rob. 25(2), 240–252 (2009)CrossRefGoogle Scholar
Zhang, L.: Study of a new improved PSO-BP neural network algorithm. J. Harbin Inst. Technol. (New Ser.) 20(5), 206–212 (2013)zbMATHGoogle Scholar
Krzysztof, W.: Terrain classification and negotiation with a walking robot. J. Intell. Robot Syst. 78, 401–423 (2015)CrossRefGoogle Scholar
Lankarani, H.M.: A contact force model with hysteresis damping for impact analysis of multibody systems. J. Mech. Des. 112(3), 369–376 (1990)CrossRefGoogle Scholar
You, L.J.: Reconstruction and prediction of capillary pressure curve based on particle swarm optimization-back propagation neural network method. Petroleum 4, 268–280 (2018)CrossRefGoogle Scholar
Che, Z.H.: PSO-based back-propagation artificial neural network for product and mold cost estimation of plastic injection molding. Comput. Ind. Eng. 58, 625–637 (2010)CrossRefGoogle Scholar
Ren, C.: Optimal parameters selection for BP neural network based on particle swarm optimization: a case study of wind speed forecasting. Knowl.-Based Syst. 56, 226–239 (2014)CrossRefGoogle Scholar
Wang, H.S.: Cost estimation of plastic injection molding parts through integration of PSO and BP neural network. Expert Syst. Appl. 40, 418–428 (2013)CrossRefGoogle Scholar
Irawan, A.: Optimal impedance control based on body inertia for a hydraulically driven hexapod robot walking on uneven and extremely soft terrain. J. Field Robot. 28(5), 690–713 (2011)CrossRefGoogle Scholar
Silva, M.F.: Modelling and simulation of artificial locomotion systems. Robotica 23(5), 595–606 (2005)CrossRefGoogle Scholar