Modeling and Multi Source Signal Processing of High Precision Perception Sensing System in Intelligent Robots

Abstract

With the development of intelligent robot technology, the demand for high-precision perception of complex environment is increasing. This article focuses on the modeling of high-precision sensing system, aiming at improving the accuracy and adaptability of robot environmental perception. Compared with the traditional modeling method based on physical characteristics, the data-driven modeling method is emphasized. In this article, a fusion algorithm of convolutional neural network (CNN) and long-term memory network (LSTM) is used. CNN is responsible for processing visual signals, extracting local features through convolution kernel sliding, and constructing feature hierarchy through multi-layer convolution and pooling. LSTM processes auditory and other time series signals, and solves the long-term dependence problem with special gating structure. Experiments show that data-driven modeling is lower than traditional methods in average error and maximum error, and the fusion algorithm of CNN and LSTM improves the multi-source signal processing effect, and the accuracy of target recognition reaches 88%. This verifies the effectiveness of the proposed modeling and multi-source signal processing strategy, and provides strong support for robot complex environment perception and decision-making.