Remote monitoring of geological situation at distant regions with complicated relief, including reliable detection of the rock falls and estimation of the scale of accidents, is of great importance for practical needs. In the paper, vibration data, caused by fall of sample objects and registered by cable sensors, are processed and analysed in order to construct some measure for the indirect evaluation of weight of fallen objects. Nonlinear models, which based on artificial neural networks (ANN), are proposed for the signal presentation. To improve and simplify the learning and classification abilities of the whole system, the aggregative learning method (ALM) is implemented. ALM features relatively low memory and computational resources needed for the procedure realisation, especially for data classification (recall) stage, in compare to conventional methods. The validity and efficiency of the proposed approach are tested through its application for rock fall detection and weight evaluation system using cable sensors and mobile communication network. Classification abilities of the proposed approach are shown useful for estimation of the fallen object weight. Characterised with high computational efficiency and simple decisionmaking procedure, the derived method can be useful for simple and reliable real-time monitoring system design.