A Low-Cost Split Architecture ANFIS Framework for Teaching Autonomous Navigation and Obstacle Avoidance

Robotics education in Malaysia requires affordable platforms that enables students to understand intelligent and adaptive control without the computational complexity of conventional deep learning architecture. However, many existing approaches uses computational demanding models that are difficult to deploy on low-cost microcontrollers and can make relationship between sensor inputs, fuzzy rules and robot action difficult to understand. This study presents a hybrid Adaptive Neuro Fuzzy Inference System (ANFIS) framework for autonomous navigation and obstacle avoidance that separates model training from real time robot control. Expert-driving data are generated from simulated obstacle avoidance scenarios and used to learn Gaussian membership functions and first order Sugeno rule consequents in a Python/PyTorch environment. The resulting parameters are exported as static models to Arduino Nano ESP32, enabling millisecond scale onboard inference from three ultrasonic sensors position at the centre and at the both sides of the robot. The framework was validated using 20 steps obstacle encounter dataset representing five navigation phases which include cruising, obstacle detection, emergency manoeuvre, clearing and recovery. A two-rule Sugeno model reproduced he expert steering trajectory with a root mean square error (RSME) of 1.89°, corresponding to approximately 2.1% of the ±90° output range and achieved a Pearson correlation of r = 0.999 with the logged steering commands. Furthermore, centre sensor clearance exhibited a strong inverse correlation with steering magnitude (r = -0.94), conforming the dominant influence of frontal obstacle information during avoidance. The results shows that proposed framework can provide accurate real time fuzzy control on low-cost embedded hardware while allowing students to clearly observe how sensor data, learned fuzzy rules and control actions are connected. The framework therefore provides a practical and transparent platform for teaching intelligent robotic control.

Adaptive Neuro-Fuzzy Inference System (ANFIS); Autonomous Navigation; Obstacle Avoidance; Robotic Education

1. Betz, J., Zheng, H., Zang, Z., Sauerbeck, F., Walas, K., Dimitrov, V., Behl, M., Zheng, R., Biswas, J., Krovi, V., & Mangharam, R. (2022). Teaching Autonomous Systems Hands-On: Leveraging Modular Small-Scale Hardware in the Robotics Classroom. arXiv (Cornell University). [Google Scholar] [Crossref]

2. Klomp, M., Araffa, K., & Saunders-Smits, G. N. (2024). MIRTE: an affordable, open, mobile robot education platform. In Proc. 52nd Annu. Conf. SEFI*, Lausanne, Switzerland. [Google Scholar] [Crossref]

3. Sheh, R., Komsuoglu, H., & Jacoff, A. (2014, October). The open academic robot kit: Lowering the barrier of entry for research into response robotics. In 2014 IEEE International Symposium on Safety, Security, and Rescue Robotics (2014)(pp. 1-6). IEEE. [Google Scholar] [Crossref]

4. Rafai, A. N. A., Adzhar, N., & Jaini, N. I. (2022). A review on path planning and obstacle avoidance algorithms for autonomous mobile robots. Journal of Robotics, 2022(1), 2538220. [Google Scholar] [Crossref]

5. Kumar, A., Sahasrabudhe, A., & Nirgude, S. (2024). Fuzzy logic control for indoor navigation of mobile robots. arXiv preprint arXiv:2409.02437. [Google Scholar] [Crossref]

6. Zúñiga, J., Chamorro, W., Medina, J., Proaño, P., Díaz, R., & Chillán, C. (2024). On the performance comparison of fuzzy-based obstacle avoidance algorithms for mobile robots. Engineering Proceedings, 77(1), 23. [Google Scholar] [Crossref]

7. Souroulla, T., Hata, A., Terra, A., Özkahraman, Ö., & Inam, R. (2022). Model Compression for Resource-Constrained Mobile Robots. arXiv preprint arXiv:2207.10082. [Google Scholar] [Crossref]

8. Jang, J. S. (1993). ANFIS: adaptive-network-based fuzzy inference system. IEEE transactions on systems, man, and cybernetics, 23(3), 665-685. [Google Scholar] [Crossref]

9. Takagi, T., & Sugeno, M. (1985). Fuzzy identification of systems and its applications to modeling and control. IEEE transactions on systems, man, and cybernetics, (1), 116-132. [Google Scholar] [Crossref]

10. Gharajeh, M. S., & Jond, H. B. (2022). An intelligent approach for autonomous mobile robots path planning based on adaptive neuro-fuzzy inference system. Ain Shams Engineering Journal, 13(1), 101491. [Google Scholar] [Crossref]

11. Haider, M. H., Wang, Z., Khan, A. A., Ali, H., Zheng, H., Usman, S., ... & Zhi, P. (2022). Robust mobile robot navigation in cluttered environments based on hybrid adaptive neuro-fuzzy inference and sensor fusion. Journal of King Saud University-Computer and Information Sciences, 34(10), 9060-9070. [Google Scholar] [Crossref]

12. Yosif, Z., Mahmood, B., & Saeed, S. Z. (2023). Global and local robot navigation combination for mobile robot obstacle avoidance. Journal of Optimization and Decision Making, 2(2), 335-343. [Google Scholar] [Crossref]

13. Saleh, M. S., Al Mashhadany, Y. I., Alshaibi, M., Ameen, F. M., & Algburi, S. (2025). Optimal mobile robot navigation for obstacle avoidance based on ANFIS controller. Journal of Robotics and Control (JRC), 6(1), 484-492. [Google Scholar] [Crossref]

• Assessment of the Role of Artificial Intelligence in Repositioning TVET for Economic Development in Nigeria
• Teachers’ Use of Assure Model Instructional Design on Learners’ Problem Solving Efficacy in Secondary Schools in Bungoma County, Kenya
• “E-Booksan Ang Kaalaman”: Development, Validation, and Utilization of Electronic Book in Academic Performance of Grade 9 Students in Social Studies
• Analyzing EFL University Students’ Academic Speaking Skills Through Self-Recorded Video Presentation
• Major Findings of The Study on Total Quality Management in Teachers’ Education Institutions (TEIs) In Assam – An Evaluative Study