V.P. Baryaksheva1

1 Saint Petersburg State University of Aerospace Instrumentation (St. Petersburg, Russia)

1 vsvally@mail.ru

Implementation of unmanned systems in modern transport infrastructure today characterizes the socio-economic and scientific-technical level of the country's development. The most interesting task is to increase air cargo traffic with modernization of airport infrastructure. Purpose – increase the level of automation of cargo flow service and optimize the processes of aircraft service: to automate the processes of cargo movement on the territory of airports by introducing unmanned technologies.

A combined two-stage algorithm is proposed, which combines the stage of object detection based on the assignment of the central cross-section by two key points of the bounding box and the stage of predicting their trajectory based on the processing of the lidar data point cloud using the surrounding obstacles and terrain map data. A distinctive feature of the algorithm is the prediction of the likely trajectory of objects by estimating the minimum amount of object data observed in the scene. The accuracy and speed of the combined algorithm is ensured by using the original open codes of CenterPoint++ and MTR algorithms, confirmed by the verification of its work on the Waymo Open Dataset by widely used metrics: mAP, minADE and their modifications in comparison with existing models of predicting the trajectory of objects. Practical relevance – the research is aimed at improving the safety of vehicle and aircraft traffic on the airport territory, as well as at improving the safety of aircraft maintenance by service vehicles. The results of the research can be applied in active assistive collision warning systems of unmanned vehicles and autonomous robots.

Baryaksheva V.P. Algorithm for determining trajectory parameters of ground-based dynamic objects based on changes in the directions of central cross sections in a cloud of lidar measurement points. Achievements of modern radioelectronics. 2025. V. 79. № 8. P. 67–75. DOI: https://doi.org/10.18127/j20700784-202508-07 [in Russian]

1. Novosti aeroporta Pulkovo. URL: https://pulkovoairport.ru/about/press_center/news/53524 [in Russian]
2. Lobodinova O.O., Strotsev A.A. Model' formirovaniya dannykh navigatsionnykh datchikov dlya otsenki prostranstvennogo polozheniya ob"ektov. Uspekhi sovremennoy radioelektroniki. 2024. № 7 (78). S. 37–48. DOI: https://doi.org/10.18127/j20700784-202407-04. [in Russian]
3. Koshkarov A. S., Baryaksheva V. P., Dvornikov R.I. Sravnitel'nyy analiz metodov 3D-obnaruzheniya ob"ektov na testovom nabore dannykh Waymo Open Dataset. Sb. dokladov V Mezhdunar. nauch. konf. «Aerokosmicheskoe priborostroenie i ekspluatatsionnye tekhnologii», 4–20 aprelya 2024 g. SPb. S. 77–82. [in Russian]
4. Waymo Open Dataset. Waymo. URL: https://waymo.com/open
5. Yihan H., Zhuangzhuang D., Runzhou G., Wenxin S., Li H., Kun L., Qiang L. AFDetV2: Rethinking the Necessity of the Second Stage for Object Detection from Point Clouds. Proceedings of the AAAI Conference on Artificial Intelligence. 2022. V. 36. № 1. P. 969–979.
6. Yin T., Zhou X., Krähenbühl P. Center-based 3D Object Detection and Tracking. IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). 2021. P. 11784–11793.
7. Deng J., Shi S., Li P., Zhou W., Zhang Y., Li H. Voxel R-CNN: Towards High Performance Voxel-based 3D Object Detection. Proceedings of the AAAI Conference on Artificial Intelligence. 2021. V. 35. № 2. P. 1201–1209.
8. Bestugin A.R., Filin A.D., Kirshina I.A. Differential parameter method of aircraft flight simulation in aviation virtual electronic proving grounds. IOP Conference Series: Materials Science and Engineering. 2020. URL: https://www.elibrary.ru/download /elibrary_43321573_98641488.pdf
9. Gao X., Jia X., Li Y., Xiong H. Dynamic Scenario Representation Learning for Motion Forecasting With Heterogeneous Graph Convolutional Recurrent Networks. IEEE Robotics and Automation Letters. 2023. V. 8. № 5. P. 2946–2953.
10. Shi S., Jiang, L., Dai D., Schiele B. Motion Transformer with Global Intention Localization and Local Movement Refinement. 36th Conference on Neural Information Processing Systems (NeurlIPS 2022). URL: https://proceedings.neurips.cc/paper_files/paper/ 2022/file/2ab47c960bfee4f86dfc362f26ad066a-Paper-Conference.pdf
11. Piergiovanni A., Casser V., Ryoo M.S., Angelova A. 4D-Net for Learned Multi-Modal Alignment. IEEE/CVF International Conference on Computer Vision (ICCV). 2021. P. 15415–15425.
12. Koshkarov A.S., Semenova V.P. Razrabotka imitatsionnoy modeli lidara sistemy preduprezhdeniya stolknoveniy servisnogo transportnogo sredstva. Trudy MAI. 2023. № 128. URL: https://trudymai.ru/upload/iblock/44e/ eahh5vjvcq4ds21ez6nscha8qscxgbz5/ Koshkarov_Semenova_1.pdf?lang=ru&issue=128. [in Russian]