Constrained Multi-Objective Bayesian Optimization for Unmanned Ground Vehicle-Oriented Detection Models

Öz

Reliable perception of non-NATO armored vehicles is fundamental for Unmanned Ground Vehicle (UGV) operations in safety-critical, time-constrained environments. This study proposes a UGV-oriented framework integrating lightweight You Only Look Once (YOLO) architectures with a constrained multi-objective Bayesian optimization strategy. An original hybrid dataset of 10,640 ground-level images was constructed, featuring tanks, armored personnel carrier (APCs) main battle tank (MBT), self-propelled howitzers, and hard-negatives, excluding aerial views for domain consistency. Quantitative evaluation shows YOLOv9s achieves the highest accuracy reaching 97.33% mAP@50 and 0.8478 mAP@50–95, while maintaining a high recall of 92.11% and the highest Matthews Correlation Coefficient (MCC) score (0.8312). YOLO11s provides the highest sensitivity with a recall of 92.75%, whereas YOLOv5su delivers the lowest latency (13.82 ms) and highest throughput (72.4 FPS), highlighting critical trade-offs between detection accuracy and computational efficiency. To address accuracy-efficiency trade-offs, a Multi-Objective Tree-structured Parzen Estimator (MOTPE) based Bayesian optimization framework yielded Pareto-optimal configurations for Ambush, Reconnaissance, and Balanced mission modes. This approach enables adaptive, hardware-aware model selection while preserving mission-critical detection performance.

 

Anahtar Kelimeler

• Armored Vehicle Detection
• MOTPE
• Multi-objective Bayesian Optimization
• Real-time Object Detection
• Unmanned Ground Vehicles
• YOLO

İnsansız Kara Aracı Odaklı Algılama Modelleri için Kısıtlı Çok Amaçlı Bayes Optimizasyonu

Öz

NATO dışı zırhlı araçların güvenilir bir şekilde algılanması, güvenlik açısından kritik ve zaman kısıtlı ortamlarda insansız kara araçlarının (UGV) operasyonları için temel öneme sahiptir. Bu çalışma, hafif YOLO mimarilerini kısıtlı çok amaçlı Bayes optimizasyon stratejisiyle entegre eden UGV odaklı bir çerçeve önermektedir. Alan tutarlılığı için havadan görünümler hariç tutularak, tanklar, zırhlı personel taşıyıcıları, kendinden tahrikli obüsler ve sert negatifleri içeren 10.640 yer seviyesi görüntüsünden oluşan orijinal bir hibrit veri seti oluşturulmuştur. Nicel değerlendirme, YOLOv9s'nin %97,33 mAP@50 ve 0,8478 mAP@50–95 ile en yüksek doğruluğa ulaştığını, aynı zamanda %92,11'lik yüksek bir geri çağırma oranı ve en yüksek MCC puanını (0,8312) koruduğunu göstermektedir. YOLO11s, %92,75'lik bir geri çağırma oranıyla en yüksek hassasiyeti sağlarken, YOLOv5su en düşük gecikmeyi (13,82 ms) ve en yüksek işlem hızını (72,4 FPS) sunarak tespit doğruluğu ve hesaplama verimliliği arasındaki kritik dengeyi vurgulamaktadır. Doğruluk-verimlilik dengelerini ele almak için, MOTPE tabanlı bir Bayes optimizasyon çerçevesi, Pusu, Keşif ve Dengeli görev modları için Pareto-optimal konfigürasyonlar üretmiştir. Bu yaklaşım, görev açısından kritik tespit performansını korurken, uyarlanabilir, donanıma duyarlı model seçimini mümkün kılar.

Anahtar Kelimeler

• İnsansız Kara Araçları
• YOLO
• Zırhlı Araç Tespiti
• Çok Amaçlı Bayes Optimizasyonu
• MOTPE
• Gerçek Zamanlı Nesne Tespiti

Kaynakça

1. [1] Ersü, C., Petlenkov, E., & Janson, K. (2024). A systematic review of cutting-edge radar technologies: Applications for unmanned ground vehicles (UGVs). Sensors, 24(23), 7807. https://doi.org/10.3390/s24237807
2. [2] Wang, X., Guo, Y., & Gao, Y. (2024). Unmanned autonomous intelligent system in 6G non-terrestrial network. Information, 15(1), 38. https://doi.org/10.3390/info15010038
3. [3] Munasinghe, I., Perera, A., & Deo, R. C. (2024). A comprehensive review of UAV–UGV collaboration: Advancements and challenges. Journal of Sensor and Actuator Networks, 13(6), 81. https://doi.org/10.3390/jsan13060081
4. [4] Wang, Y., Li, J., Yang, X., & Peng, Q. (2025). UAV–ground vehicle collaborative delivery in emergency response: A review of key technologies and future trends. Applied Sciences, 15(17), 9803. https://doi.org/10.3390/app15179803
5. [5] Nistorescu, C. V. (2024). The role of heavy armour in modern warfare. Romanian Military Thinking, (3), 42–59. https://doi.org/10.55535/RMT.2024.3.03
6. [6] Ma, Z., Xiong, J., Gong, H., & Wang, X. (2024). Adaptive depth graph neural network-based dynamic task allocation for UAV–UGVs under complex environments. IEEE Transactions on Intelligent Vehicles. https://doi.org/10.1109/TIV.2024.3457493
7. [7] Amorim, J. S., Neto, A. F., Chaves, R. S., Zachi, A. R., Gouvêa, J. A., Andrade, F. A., & Pinto, M. F. (2025). Collaborative inspection of solar panel farms using YOLOv5-based computer vision and UGV–UAV integration. Journal of Intelligent & Robotic Systems, 111(2), 66. https://doi.org/10.1007/s10846-025-02265-w
8. [8] Sun, T., Feng, B., Huo, J., Xiao, Y., Wang, W., Peng, J., … Liu, L. (2024). Artificial intelligence meets flexible sensors: Emerging smart flexible sensing systems driven by machine learning and artificial synapses. Nano-Micro Letters, 16(1), 14. https://doi.org/10.1007/s40820-023-01235-x

9. [9] Bayoudh, K., Knani, R., Hamdaoui, F., & Mtibaa, A. (2022). A survey on deep multimodal learning for computer vision: Advances, trends, applications, and datasets. The Visual Computer, 38(8), 2939–2970. https://doi.org/10.1007/s00371-021-02166-7
10. [10] Tahir, N. U. A., Zhang, Z., Asim, M., Chen, J., & ELAffendi, M. (2024). Object detection in autonomous vehicles under adverse weather: A review of traditional and deep learning approaches. Algorithms, 17(3), 103. https://doi.org/10.3390/a17030103
11. [11] Liu, Q., Li, Z., Yuan, S., Zhu, Y., & Li, X. (2021). Review on vehicle detection technology for unmanned ground vehicles. Sensors, 21(4), 1354. https://doi.org/10.3390/s21041354
12. [12] Guo, Y., He, Y., Zhang, H., & Ma, J. (2025). YOLO-CAM: A lightweight UAV object detector with combined attention mechanism for small targets. Remote Sensing, 17(21), 3575. https://doi.org/10.3390/rs17213575
13. [13] Yépez-Ponce, D. F., Montalvo, W., Guamán-Gavilanes, X. A., & Echeverría-Cadena, M. D. (2025). Route optimization for UGVs: A systematic analysis of applications, algorithms and challenges. Applied Sciences, 15(12), 6477. https://doi.org/10.3390/app15126477
14. [14] Ali, M. L., & Zhang, Z. (2024). The YOLO framework: A comprehensive review of evolution, applications, and benchmarks in object detection. Computers, 13(12), 336. https://doi.org/10.3390/computers13120336
15. [15] Zou, Z., Chen, K., Shi, Z., Guo, Y., & Ye, J. (2023). Object detection in 20 years: A survey. Proceedings of the IEEE, 111(3), 257–276. https://doi.org/10.1109/JPROC.2023.3238524
16. [16] Dilmi, W., El Ferik, S., Ouerdane, F., Khaldi, M. K., & Saif, A. W. A. (2025). Technical aspects of deploying UAV and ground robots for intelligent logistics using YOLO on embedded systems. Sensors, 25(8), 2572. https://doi.org/10.3390/s25082572
17. [17] Wang, L., Wang, H., Letchmunan, S., Xiao, R., Ahmed, O. H., & Liu, Z. (2025). A systematic literature review of lightweight YOLO models for object detection. PeerJ Computer Science, 11, e3357. https://doi.org/10.7717/peerj-cs.3357
18. [18] Ramos, L. T., & Sappa, A. D. (2025). A decade of you only look once (YOLO) for object detection: A review. IEEE Access, 13, 192747–192794. https://doi.org/10.1109/ACCESS.2025.3630988
19. [19] Apostolidis, K. D., & Papakostas, G. A. (2025). Delving into YOLO object detection models: Insights into adversarial robustness. Electronics, 14(8), 1624. https://doi.org/10.3390/electronics14081624
20. [20] Mehmood, F., Ahmad, S., & Whangbo, T. K. (2023). An efficient optimization technique for training deep neural networks. Mathematics, 11(6), 1360. https://doi.org/10.3390/math11061360
21. [21] Sánchez, M. S., Ortiz, M. C., & Sarabia, L. A. (2016). A useful tool for computation and interpretation of trading-off solutions through Pareto-optimal front in the field of experimental designs for mixtures. Chemometrics and Intelligent Laboratory Systems, 158, 210–217. https://doi.org/10.1016/j.chemolab.2016.09.007
22. [22] Wang, C. Y., Yeh, I. H., & Mark Liao, H. Y. (2024, September). Yolov9: Learning what you want to learn using programmable gradient information. In European conference on computer vision (pp. 1-21). Cham: Springer Nature Switzerland. https://doi.org/10.1007/978-3-031-72751-1_1
23. [23] Jocher, G., Qiu, J., Liu, M., Lyu, S., Akyon, F. C., & Kalfaoglu, M. E. (2026). Ultralytics YOLO26: Unified Real-Time End-to-End Vision Models. https://doi.org/10.48550/arXiv.2606.03748
24. [24] Ultralytics. (n.d.). Ultralytics YOLO11 documentation and model zoo. https://docs.ultralytics.com Accessed 18 February 2026
25. [25] Roboflow, Inc. (n.d.). Roboflow Universe: Open-source computer vision datasets. https://universe.roboflow.com Accessed 17 February 2026
26. [26] Khorshidi, M. S., Merigó, J. M., Beydoun, G., Furht, B., & Khoshgoftaar, T. M. (2026). A bibliometric analysis of the first decade of the Journal of Big Data. Journal of Big Data. https://doi.org/10.1186/s40537-026-01412-w
27. [27] Fuentes, A. (2026). Machine Learning Models for Early Identification of At-Risk Students. International Journal of Sustainability and Advanced Integrated Research, 2(2), 2934-2939. https://doi.org/10.65339/ijsair.V2.I2.527
28. [28] Chicco, D., & Jurman, G. (2020). The advantages of the Matthews correlation coefficient (MCC) over F1 score and accuracy in binary classification evaluation. BMC Genomics, 21, Article 6. https://doi.org/10.1186/s12864-019-6413-7
29. [29] Everingham, M., Van Gool, L., Williams, C. K. I., Winn, J., & Zisserman, A. (2010). The Pascal visual object classes (VOC) challenge. International Journal of Computer Vision, 88(2), 303–338. https://doi.org/10.1007/s11263-009-0275-4
30. [30] Redmon, J., Divvala, S., Girshick, R., & Farhadi, A. (2016). You only look once: Unified, real-time object detection. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR) (pp. 779–788). https://doi.org/10.1109/CVPR.2016.91
31. [31] Bochkovskiy, A., Wang, C.-Y., & Liao, H.-Y. M. (2020). YOLOv4: Optimal speed and accuracy of object detection. arXiv preprint arXiv:2004.10934. https://doi.org/10.48550/arXiv.2004.10934
32. [32] Jocher, G., Chaurasia, A., & Qiu, J. (2020). YOLOv5 by Ultralytics
33. [Computer software]. GitHub. https://github.com/ultralytics/yolov5 Accessed 10 February 2026
34. [33] Palaniappan, D., Jain, R., Premavathi, T., Parmar, K., Ghribi, W., Ahmed, A. M., & Ahmad, N. (2025). Yolo in healthcare: A comprehensive review of detection architectures, domain applications, and future innovations. IEEE Access, 13, 145714–145735. https://doi.org/10.1109/ACCESS.2025.3599358
35. [34] Weng, S., Wang, H., Wang, J., Xu, C., & Zhang, E. (2025). YOLO-SRMX: A Lightweight Model for Real-Time Object Detection on Unmanned Aerial Vehicles. Remote Sensing, 17(13), 2313. https://doi.org/10.3390/rs17132313
36. [35] Zhong, Z., Yun, L., Cheng, F., Chen, Z., & Zhang, C. (2024). Light-YOLO: A lightweight and efficient YOLO-based deep learning model for mango detection. Agriculture, 14(1), 140. https://doi.org/10.3390/agriculture14010140
37. [36] Ioannou, G., Forti, N., Millefiori, L. M., Carniel, S., Renga, A., Tomasicchio, G., ... & Braca, P. (2024). Underwater inspection and monitoring: Technologies for autonomous operations. IEEE Aerospace and Electronic Systems Magazine, 39(5), 4-16. https://doi.org/10.1109/MAES.2024.3366144
38. [37] Pióro, Ł., Sychowiec, J., Kanciak, K., & Zieliński, Z. (2024). Application of attribute-based encryption in military internet of things environment. Sensors, 24(18), 5863. https://doi.org/10.3390/s24185863
39. [38] Kannan, P. N., Vishal, J., Preethivkanth, V., Kalimuthu, K., & Ramya, S. (2025, June). Smart Weed Detection Using Deep Learning for Sustainable Agriculture. In 2025 Second International Conference on Networks and Soft Computing (ICNSoC) (pp. 1-6). IEEE. https://doi.org/10.1109/ICNSoC66817.2025.00087
40. [39] Fang, W., & Chen, W. (2025). TBF-YOLOv8n: A Lightweight Tea Bud Detection Model Based on YOLOv8n Improvements. Sensors, 25(2), 547. https://doi.org/10.3390/s25020547
41. [40] Huang, Y., Liu, Z., Zhao, H., Tang, C., Liu, B., Li, Z., ... & Qiao, X. (2025). YOLO-YSTs: An improved YOLOv10n-based method for real-time field pest detection. Agronomy, 15(3), 575. https://doi.org/10.3390/agronomy15030575
42. [41] Hossain, M., Pritom, M. M., & Khan, A. A. S. (2024, October). YOLOv5-Driven Decision Fusion Approach for Aiding Automated Tomato Harvesting: Detection, Maturity Classification, Peduncle Estimation and Picking Point Determination. In 2024 International Conference on Innovations in Science, Engineering and Technology (ICISET) (pp. 1-6). IEEE. https://doi.org/10.1109/ICISET62123.2024.10941363
43. [42] Wang, A., Qian, W., Li, A., Xu, Y., Hu, J., Xie, Y., & Zhang, L. (2024). NVW-YOLOv8s: An improved YOLOv8s network for real-time detection and segmentation of tomato fruits at different ripeness stages. Computers and Electronics in Agriculture, 219, 108833. https://doi.org/10.1016/j.compag.2024.108833
44. [43] Alshehri, M., Wu, T., Almujally, N. A., AlQahtani, Y., Hanzla, M., Jalal, A., & Liu, H. (2025). UAV-based intelligent traffic surveillance using recurrent neural networks and Swin transformer for dynamic environments. Frontiers in Neurorobotics, 19, 1681341. https://doi.org/10.3389/fnbot.2025.1681341
45. [44] Cao, L., Xiao, W., Hu, Z., Li, X., & Wu, Z. (2025). Detection of Citrus Huanglongbing in Natural Field Conditions Using an Enhanced YOLO11 Framework. Mathematics, 13(14), 2223. https://doi.org/10.3390/math13142223
46. [45] Priambodo, A. R., Junaidi, A., & Al Haromainy, M. M. (2025). Performance Evaluation of YOLOv5su and SVM With HOG Features for Student Attendance Face Recognition. bit-Tech, 8(2), 2182-2192. https://doi.org/10.32877/bt.v8i2.3215
47. [46] Yao, L., Zhou, J., & Li, L. (2024). Improved YOLOv5su model for detecting peach leaf curl disease. Transactions of the Chinese Society of Agricultural Engineering, 40(14), 109-117. https://doi.org/10.11975/j.issn.1002-6819.202403104
48. [47] Sapkota, R., Meng, Z., Churuvija, M., Du, X., Ma, Z., & Karkee, M. (2026). Comprehensive performance evaluation of yolov12, yolo11, yolov10, yolov9 and yolov8 on detecting and counting fruitlet in complex orchard environments. Agriculture Communications, 100125. https://doi.org/10.1016/j.agrcom.2026.100125
49. [48] Huang, Y., Huang, H., Qin, F., Chen, Y., Zou, J., Liu, B., ... & Qiao, X. (2024). YOLO-IAPs: A Rapid Detection Method for Invasive Alien Plants in the Wild Based on Improved YOLOv9. Agriculture, 14(12), 2201. https://doi.org/10.3390/agriculture14122201
50. [49] Ibrahim, M., Ahmdia, A., Shbo, M., Sarhan, A., Hafez, A. T., & Abdallah, A. E. (2026). Evolution of Path-Planning and Control Methods for Unmanned Aerial Vehicles: A Comprehensive Review. International Journal of Aeronautical and Space Sciences, 1-41. https://doi.org/10.1007/s42405-026-01205-8
51. [50] Dasu, B. T., Reddy, M. V., Kumar, K. V., Chithaluru, P., Ahmed, N., & Abd Elminaam, D. S. (2026). A self-attention driven multi-scale object detection framework for adverse weather in smart cities. Scientific Reports. https://doi.org/10.1038/s41598-025-31660-4
52. [51] Peter, E., Ang, L. M., Seng, K. P., & Srivastava, S. (2026). Recent Advances in Deep Learning for SAR Images: Overview of Methods, Challenges, and Future Directions. Sensors, 26(4), 1143. https://doi.org/10.3390/s26041143
53. [52] Sheltami, T., Ahmed, G., Ghaleb, M., & Mahmoud, A. (2026). UAV Path Planning and Trajectory Optimization: A Comprehensive Survey. Arabian Journal for Science and Engineering, 51(1), 105–145. https://doi.org/10.1007/s13369-025-10971-8
54. [53] Rahmani, S., Aghalar, H., Jebreili, S., & Goli, A. (2024). Optimization and computing using intelligent data-driven approaches for decision-making. In I. Ali, U. M. Modibbo, A. L. Bolaji, & H. Garg (Eds.), Optimization and computing using intelligent data-driven approaches for decision-making (pp. 90-176). CRC Press. https://doi.org/10.1201/9781003536796-6
55. [54] Han, Z., Chen, P., Zhou, B., & Yu, G. (2025). Real-Time Navigation of Unmanned Ground Vehicles in Complex Terrains With Enhanced Perception and Memory-Guided Strategies. IEEE Transactions on Vehicular Technology, 74(3), 3723–3735. https://doi.org/10.1109/TVT.2024.3500002
56. [55] Kabir, R., Watanobe, Y., Ding, D., Islam, M. R., & Naruse, K. (2025). A Comprehensive Survey on Advanced Data Science Platforms for Cyber-Physical Systems, Digital Twins, and Robotics. IEEE Access, 13, 177269–177304. https://doi.org/10.1109/ACCESS.2025.3619776
57. [56] Akiba, T., Sano, S., Yanase, T., Ohta, T., & Koyama, M. (2019). Optuna: A next-generation hyperparameter optimization framework. In Proceedings of the 25th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining (pp. 2623–2631). https://doi.org/10.1145/3292500.3330701
58. [57] Öztaş, D. M., Salehi, H., Demir, Ö., Acici, K., Güzel, M. S., Eraslan, E. C., ... & Ekinci, F. (2026). Balancing accuracy and efficiency in fine-grained image classification: a systematic comparison of CNN, vision transformer, and hybrid architectures. Plant Systematics and Evolution, 312(4), 30. https://doi.org/10.1007/s00606-026-02005-z
59. [58] Altay, H. T., Demir, Ö., Ekinci, F., Güzel, M. S., Kumru, E., Akata, I., ... & Sevindik, M. (2026). A comparative analysis of single-and dual-backbone deep learning architectures with explainable AI for cherry leaf disease classification. Scientific Reports. https://doi.org/10.1038/s41598-026-50104-1