Virtual-Sensor-Based Planetary Soil Classification with Legged Robots
The estimation of soil properties is crucial for legged robots during planetary exploration missions. A virtual-sensor-based soil classification approach for legged robots is proposed in this paper. Instead of installing extra force sensors on the foot of the robot, joint motion information from joint position sensors and current signals from joint motors on the leg are recorded and used as the dataset in classification. The collected data is decomposed using the Discrete Wavelet Transform and assigned a soil type by a Support Vector Machine (SVM). This approach is validated on a dataset acquired from a high-fidelity simulation model of a hexapod robot, and the classification accuracy of more than 90% was achieved. Different SVM models are used in classification for comparative analysis, and the contributions of the different signals to the classification performance are evaluated. Experimental results demonstrate that the proposed approach can estimate the soil properties with a good performance and rapid forecasting speed.
KeywordsSoil classification Virtual sensor Legged robots
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).
- 1.Janosi, Z., Hanamoto, B.: Analytical determination of drawbar pull as a function of slip for tracked vehicles in deformable soils: mechanics of soil-vehicle systems: Edizoni Minerva Technica. Editione Minerva Tecnica, Torino (1962)Google Scholar
- 3.Komizunai, S., Konno, S., Abiko, A., et al.: Slip characteristics identification for biped walking of a humanoid robot on sand. In: Proceedings of Eighth International Conference on Flow Dynamics, pp. 9–11 (2011)Google Scholar
- 6.Mrva, J., Faigl, J.: Feature extraction for terrain classification with crawling robots. Inf. Technol. Appl. Theory 1422, 179–185 (2015)Google Scholar
- 7.Degrave, J., Van Cauwenbergh, R., Wyffels, F., et al.: Terrain classification for a quadruped robot. In: 12th International Conference on Machine Learning and Applications, vol. 1, pp. 185–190. IEEE (2013)Google Scholar
- 13.Gonzalez de Santos, P., Garcia, E., Estremera, J.: Virtual sensors for walking robots. In: Gonzalez de Santos, P., Garcia, E., Estremera, J. (eds.) Quadrupedal Locomotion: An Introduction to the Control of Four-legged Robots, pp. 191–211. Springer, London (2006). https://doi.org/10.1007/1-84628-307-8_8CrossRefzbMATHGoogle Scholar
- 15.Tzanetakis, G., Essl, G., Cook, P.: Audio analysis using the discrete wavelet transform. In: Proceedings of Conference in Acoustics and Music Theory Applications, vol. 66 (2001)Google Scholar
- 16.Jaber, A.A., Bicker, R.: Industrial robot backlash fault diagnosis based on discrete wavelet transform and artificial neural network. Am. J. Mech. Eng. 4(1), 21–31 (2016)Google Scholar