2026-09-21T15:12:47Z https://www.ideals.illinois.edu/oai-pmh

oai:www.ideals.illinois.edu:2142/129224 2025-10-20 col_2142_5131 col_2142_8888 com_2142_5130 com_2142_8887 com_2142_234

application/pdf en text Submission original under an indefinite embargo labeled 'Open Access'. The submission was exported from vireo on 2025-10-19 without embargo terms The student, Keyi Shen, accepted the attached license on 2025-05-06 at 23:20. The student, Keyi Shen, submitted this Thesis for approval on 2025-05-06 at 23:29. This Thesis was approved for publication on 2025-05-09 at 15:23. DSpace SAF Submission Ingestion Package generated from Vireo submission #21852 on 2025-10-19 at 18:09:34 Long-horizon motion planning with branch-and-bound and neural dynamics Shen, Keyi 2025-05-09 Zhang, Huan Li, Yunzhu Robotic Manipulation Model-based Planning Neural Dynamics Models Branch-and-bound Method eng Neural-network-based dynamics models learned from observational data have shown strong predictive capabilities for scene dynamics in robotic manipulation tasks. However, their inherent non-linearity presents significant challenges for effective planning. Current planning methods, often dependent on extensive sampling or local gradient descent, struggle with long-horizon motion planning tasks involving complex contact events. In this paper, we present a GPU-accelerated branch-and-bound (BaB) framework for motion planning in manipulation tasks that require trajectory optimization over neural dynamics models. Our approach employs a specialized branching heuristics to divide the search space into subdomains, and applies a modified bound propagation method, inspired by the state-of-the-art neural network verifier α,β-CROWN, to efficiently estimate objective bounds within these subdomains. The branching process guides planning effectively, while the bounding process strategically reduces the search space. Our framework achieves superior planning performance, generating high-quality state-action trajectories and surpassing existing methods in challenging, contact-rich manipulation tasks such as non-prehensile planar pushing with obstacles, object sorting, and rope routing in both simulated and real-world settings. Furthermore, our framework supports various neural network architectures, ranging from simple multilayer perceptrons to advanced graph neural dynamics models, and scales efficiently with different model sizes. 2025-05 Text https://hdl.handle.net/2142/129224 Copyright 2025 Keyi Shen Electrical & Computer Eng Electrical & Computer Engr University of Illinois Urbana-Champaign M.S. Thesis