Useful Signal Discrimination Based on Echolocation Path Analysis Using Machine Learning Methods

Introduction. When measuring the distance with a laser rangefinder, interference along the path of the beam can significantly affect measurement accuracy. Classical filtering algorithms, which rely on the characteristics of the received signal, are unable to reliably distinguish the useful signals from natural interference and targets. In this work, we propose an algorithm for classifying objects based on processing echolocation path signals obtained by vertical laser rangefinder sensing. The main task is to distinguish useful signals against the background of natural interference, such as atmospheric aerosols, haze, and clouds. The problem is solved by a method based on isolating intensity peaks on an echolocation path, followed by their classification using machine learning methods. Owing to the preprocessing of the signal, the algorithm is compatible with systems having different parameters of the emitter and the receiving channel.

Aim. To develop an algorithm for identifying and classifying objects on an echolocation path obtained by vertical sensing using a laser rangefinder.

Materials and methods. Vertical sensing was carried out using an experimental stand, which includes a laser rangefinder with a radiation source of 100 mJ and a wavelength of 1064 nm. This made it possible to collect data in various meteorological conditions. The sample included echolocation traces obtained from different objects with different characteristics, such as monolithic structures, lattice structures, and natural interference encountered during vertical sensing. A comparative analysis of the following algorithms was conducted: logistic regression, random forest, gradient boosting, and a neural network algorithm. The metrics accuracy and F1-score were chosen to evaluate the prediction quality of the models. The models were trained on a dataset containing cloud structures, with a split into training and test sets.

Results. The gradient boosting model demonstrated performance comparable with that of the neural network algorithm, achieving an F1-score of 0.89 on the test set. This makes it suitable for deployment in resource-limited systems without compromising predictive performance.

Conclusion. The results confirm the effectiveness of the algorithm for useful signal discrimination under interference conditions, which is important for geodesy, navigation, and satellite sensing.

1. Muzal M., Zygmunt M., Knysak P., Drozd T., Jakubaszek M. Methods of Precise Distance Measurements for Laser Rangefinders with Digital Acquisition of Signals. Sensors. 2021, vol. 21, iss. 19, art. no. 6426. doi: 10.3390/s21196426

2. Vilner V., Volobuev V., Laryushin A., Ryabokul A. Accuracy of Measurements of the Pulse Laser Rangefinder. Photonics. 2013, no. 3, pp. 42–60. (In Russ.)

3. Pisarev I. S., Uhov A. A. Klassifikatsiya tselei na osnove analiza lazernogo izlucheniya s ispol'zovaniem novoi metodiki razmetki grafikov [Classification of Targets Based on Laser Radiation Analysis Using a New Graph Marking Technique]. The 80th Scientific and Technical Conf. of the St Petersburg NTO RES named after A.S. Popov, dedicated to Radio Day: collection of reports. SPbSETU. Saint Petersburg, 2025, p. 32–35. (In Russ.)

4. Romano F., Cimini D., Di Paola F., Gallucci D., Larosa S., Nilo S. T., Ricciardelli E., Iisager B. D., Hutchison K. The Evolution of Meteorological Satellite CloudDetection Methodologies for Atmospheric Parameter Retrievals. Remote Sens. 2024, vol. 16, iss. 14, art. no. 2578. doi: 10.3390/rs16142578

5. Manakitsa N., Maraslidis G. S., Moysis L., Fragulis G. F. A Review of Machine Learning and Deep Learning for Object Detection, Semantic Segmentation, and Human Action Recognition in Machine and Robotic Visionю Technologies. 2024, vol. 12, iss. 2, art. no. 15. doi: 10.3390/technologies12020015

6. Pinto A. M., Rocha L. F., Moreira A. P. Object Recognition Using Laser Range Finder and Machine Learning Techniques. Robotics and Computer-Integrated Manufacturing. 2013, vol. 29, iss. 1, pp. 12–22. doi: 10.1016/j.rcim.2012.06.002

7. Grollius S., Grosse S., Ligges M., Grabmaier A. Optimized Interference Suppression for TCSPC LiDAR. IEEE Sensors J. 2022, vol. 22, no. 24, pp. 24094–24101. doi: 10.1109/JSEN.2022.3216810

8. Reguiegue M., Chouireb F. Automatic day Time Cloud Detection Over Land and Sea from MSG SEVIRI Images Using Three Features and Two Artificial Intelligence Approaches. Signal, Image and Video Processing. 2018, vol. 12, pp. 189–196. doi: 10.1007/s11760-017-1145-0

9. Seo H., Yoon H., Kim D., Kim J., Kim S.-J., Chun J.-H. Direct TOF Scanning LiDAR Sensor With Two-Step Multievent Histogramming TDC and Embedded Interference Filter. IEEE J. of Solid-State Circuits. 2021, vol. 56, no. 4, pp. 1022–1035. doi: 10.1109/JSSC.2020.3048074

10. Kurihana T., Moyer E. J., Foster I. T. AICCA: AI-Driven Cloud Classification Atlas. Remote Sensing. 2022, vol. 14, iss. 22, art. no. 5690. doi: 10.3390/rs14225690

11. Rossow W. B., Schiffer R. A. ISCCP Cloud Data Products. Bull. Amer. Meteorol. Soc. 1991, vol. 71, pp. 2–20.

12. Rybakova G. V. Oblaka i ikh transformatsiya [Clouds and Their Transformation]. Ed by I. V. Kuzhevskaya. Tomsk, Publishing House of Tomsk State University, 2020, 234 p. (In Russ.)

13. Vilner V., Laryushin A., Ryabokul A. Optoelectronic Altimeter-Speed Meters Based on Semiconductor Lasers for UAVs. Power engineering: research, equipment, technology. 2015, no. 5–6, pp. 127–133. doi: 10.30724/1998-9903-2015-0-5-6-127-133

14. Pisarev S., Uhov A. A. Development of a Photodetector Device Based on an Avalanche Photodiode with a Thermal Compensation System. IEEE 3rd Intern. Conf. on Problems of Informatics, Electronics and Radio Engineering, Novosibirsk, Russia, 15–17 Nov. 2024. IEEE, 2024, pp. 560–563. doi: 10.1109/PIERE62470.2024.10804907

15. Qun Hao, Jie Cao, Yao Hu, Yunyi Yang, Kun Li, Tengfei Li. Differential Optical-Path Approach to Improve Signal-to-Noise Ratio of Pulsed-Laser Range Finding. Optics Express. 2014, vol. 22, iss. 1, pp. 563–575. doi: 10.1364/OE.22.000563