The article addresses the development of a multisensor intelligent system for detecting explosive objects during excavator and bulldozer operations. An analysis of existing methods for detecting explosive objects is carried out, and the feasibility of implementing such a system in Ukraine is substantiated. The concept and architecture of the proposed system are presented, artificial intelligence algorithms are defined, and simulation results are reported. The system integrates a ground-penetrating radar (GPR) operating in the 400–900 MHz range, a magnetogradient sensor, an infrared (IR) sensor, and an inertial measurement unit (IMU). Data processing is performed using convolutional neural networks (CNNs) for GPR signals, long short-term memory (LSTM) networks for magnetic signals, and Bayesian fusion for integrating multisensor data. Simulation results indicate a detection probability of 95–97% with a false alarm rate of 8–10%, a probing depth of up to 1.5 m, and a processing time of 40–50 ms. Further research should focus on experimental field validation of the system, expanding the sensor set through the inclusion of acoustic and chemical detectors, adapting artificial intelligence algorithms to different soil types in Ukraine, and integrating the system with geographic information systems for automated mapping of cleared areas
Levkivskyi et al. (Tue,) studied this question.