This paper presents our solution to the SHL recognition challenge 2023 which focuses on recognizing 8 transportation modes in a user-independent manner based on motion and GPS sensor data. Our team ZZL propose an ensemble framework based on fine multi-window feature engineering and overfitting prevention. Firstly, we extracted a large and diverse set of features in the feature engineering process, including incorporating OpenStreetMap data to better leverage location data, and introducing multiple time windows to extract long, medium, and short term aggregated features, providing rich feature inputs. Secondly, we proposed an ensemble framework that comprehensively utilizes different techniques to prevent overfitting, including data downsampling, fine-tuning data distribution, designed train-test splitting, and model integration. Moreover, we applied post-processing on the model predictions to smooth the predicted results. Finally, we achieve F1-score of 0.868 on validation dataset.
Detection of transient signals in a noisy environment is an important topic in radar, sonar, and communications. An optoelectronic system for detection of transient signals has been constructed. The detection scheme is based on the Gabor representation of a signal, which can be used to represent transient signals of unknown shape and arrival time. The transient signal and Gabor window function are written to the laser beam profile via either transparency or spatial light modulator. The Gabor coefficients are detected by a 2-D CCD array. The use of a liquid crystal television, allowing for real-time detection of signals is investigated. Experimental results for exponentially decaying signals, are presented which clearly indicate the signal frequency and arrival time.
A real-time optical cross correlator of very long digital or analog data streams and capable of operating in excess of 40 million samples per second is described. The overall system architecture relies on the implementation of the overlap-and-add algorithm. The algorithm has been appropriately modified to implement the cross correlation of two long signals by first computing the cross correlation of many shorter ones. The optical system utilizes a cw laser source, an acousto-optic cell, a liquid crystal display and a multichannel parallel photodiode array to implement the short correlations. A digital system is used to form the final correlation output. Preliminary results are described.