Learning from Demonstration: Apple Picking Robot

Due to the natural variation in the scene it is difficult to program a fixed motion trajectory for a robot to harvest fruit. Wageningen University & Research demonstrates a Learning from Demonstration (LfD) approach to make it easier for robots to perform tasks such as harvesting apples. This research is part of the Data Driven High Tech Knowledge Based programme and the Synergia project. . The research presented in this video has been carried out by Robert van de Ven, Angelo Mencarelli and Jochen Hemming. Read more in the publications: van de Ven, R., Leylavi Shoushtari, A., Nieuwenhuizen, A., Kootstra, G.& van Henten, E. J. (2024): Using Learning from Demonstration (LfD) to perform the complete apple harvesting task. Computers and Electronics in Agriculture.224, 109195. https://doi.org/10.1016/j.compag.2024.109195 van de Ven, R., Nieuwenhuizen, A., van Henten, E. J., Kootstra, G. (2024): DualLQR: Efficient Grasping of Oscillating Apples using Task Parameterized Learning from Demonstration. arXiv preprint arXiv:2409.16957, 2024. https://doi.org/10.48550/arXiv.2409.16957