The study presents a method for dynamically configuring Model Predictive Controllers (MPC) to resolve local navigation issues based on the information represented in the semantic graph world model. The MPC problem’s constraints and objectives can be derived from a semantic graph world model that represents the environment’s layout, geometry, and area semantics. The researchers argue that by changing the area semantics, the layout, or the environment’s geometry as represented in the graph world model, the robot will behave differently in different areas in the environment. This approach drastically decreases computation time, while retaining task execution performance similar to an approach in which each robot always includes all other robots in its MPC computations.
Publication date: 2 Nov 2023
Project Page: https://arxiv.org/abs/2311.01180v1
Paper: https://arxiv.org/pdf/2311.01180
