The present paper proposes a method to enhance the indoor navigation of an intelligent wheelchair through the integration of a semantic layer into the navigation framework. Semantic navigation, which refers to a robot’s ability to comprehend its surroundings, is of particular significance in the case of intelligent wheelchairs. To address the problem of negotiating narrow passages, the proposed solution employs a mapping layer that automatically detects and annotates these passages, as well as a planning layer that utilizes the resulting semantic information to guide lower-level planners. The feasibility of the proposed solution was evaluated through a series of experiments in both simulated and real-world scenarios.