Boston University
Comparative analysis of predictive control methods for mobile robots in dynamic environments
Abstract
dc:description.abstractThe focus of this thesis is to evaluate the performance of Model Predictive Control (MPC) and Model Predictive Contouring Control (MPCC) for path following and obstacle avoidance tasks in congested environments. Autonomous robots need to safely and efficiently traverse environments with static and dynamic obstacles of varying degrees while respecting dynamic constraints. In this work, a unicycle model was used to formulate the controllers within a receding horizon optimization framework. These two controllers differ in that the standard MPC for this task is a trajectory tracking problem, while MPCC is formulated to minimize contouring and lag errors relative to progress along a geometric reference path. These optimal control methods were implemented in simulation and evaluated by varying obstacle density and congestion scenarios, with the key performance metrics being task success, path following accuracy, safety margin behavior and qualitative failure modes. Results show significant tradeoffs in aggressiveness, path adherence, and robustness under different levels of congestion. In addition to quantitative comparisons, this work highlights scenario dependent failure modes that are not immediately obvious from pure success metrics alone. Qualitative findings show that although MPCC does not suffer from the time and geometry coupling of the standard MPC path tracking and is more conservative and safe, it has a higher computational cost. These findings ultimately clarify the explicit trade-off between the formulations for robotic navigation tasks that involve tight geometric constraints and obstacle interaction.
Degree
thesis:*- Name thesis:degree_name
- Master of Science
- Level thesis:degree_level
- masters
- Discipline thesis:degree_discipline
- Mechanical Engineering
- Grantor
- Boston University
- Year dc:date.issued
- 2026
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Amgalanbat, Usukhbayar
- Advisor dc:contributor.advisor
-
- Baillieul, John
- ORCID dc:identifier.orcid
Subjects
dc:subject × 5Rights
dc:rights- Statement dc:rights
-
- Attribution 4.0 International
- Licence dc:rights.uri
- Language dc:language.iso
- en_US
Identifiers
dc:identifier.*- Handle dc:identifier.uri
- https://hdl.handle.net/2144/52977