**4. Conclusions**

The paper presents a semi-control method of an electric wheelchair combined with an RGB-D camera system, a graphical user interface, and real environmental maps with natural landmarks, in which optimal path planning for the wheelchair navigation was determined. In particular, 2D grid maps were used for training to create the shortest paths to the targets, in which the virtual-real RL method using DQNs carried out the training process effectively. After training, disabled people may select the desired target on the interface-user map using EEG signals to reach it. Therefore, the semi-control wheelchair located itself based on natural landmarks during movement following the optimal path from the motion planner in the real indoor environment. With the proposed method for the optimal path based on DQNs, the semi-control wheelchair could operate well to reach the desired target with small errors compared to the simulation trajectory, as well as to the trajectory of the self-control user using an EEG system. It is obvious that, with our proposed optimal path trajectory and the semi-automatic control method, the semi-control wheelchair movement is more stable, safe, and takes less time for moving. As a result of the proposed method, this wheelchair control system can be developed to apply to more complex environments with obstacles in the future.

**Author Contributions:** Conceptualization, methodology, simulation, and writing original draft, B.-V.N.; methodology, supervision, validation and writing—review and editing, T.-H.N. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding.

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** Not applicable.

**Acknowledgments:** We would like to thank Ho Chi Minh City University of Technology and Education, Vietnam.

**Conflicts of Interest:** The authors declare no conflict of interest.
