Research On Target Recognition Technologies Based On Wireless Communication Signals

Using wireless communication signals, also known as radio-frequency(RF) signal, for not only data communication, but also target detection and classification is an emerging technology that has a great potential in our daily life. In particularly, this technology can be used for target detection in foliage environment. How to select a specific RF signal as well as extracting its feature is a bottleneck issue for this technology. Thus, it is worthwhile to be investigated in detail from both academia and industry research point of view.In this dissertation, the target recognition techniques based on RF signals has been extensively investigated. The following outcomes have been achieved.In order to tackle the issues related to target detection in foliage environment, a novel detection method based on RF signal is proposed.Firstly, several different types of RF signals are compared in terms of their signal properties. It has been found that the ultra-wideband (UWB) signal is the most appropriate one. Secondly, the statistic properties of UWB signal,such as multi-path prorogation, is selected as the extracted features. Finally,different classification algorithms along with these selected features are proposed for target classification. Based on the collected data, the performance of the proposed methods have been extensively verified. The measured results have shown that the proposed methods can be successfully used for not only target detection but also classification between animal and human in a foliage environment.To minimize the impact on detection accuracy due to unwanted clutters,a high-order statistics (HOS) algorithm is proposed for feature extraction,which has a good immunity against background noise. First of all, 1-D diagonal slices of fourth-order cumulants are used to extract useful information from the received UWB signals. Then, different classification algorithms are used to verify the performance of this approach. Based on the collected data in a foliage environment, the proposed HOC-based feature extraction method has demonstrated a superior performance in terms of detection accuracy.Another critical issue related to the proposed method is potential impact on detection accuracy due to weather variations. Thus, a novel method for recognition of multi-target under multi-scenario based on optimized support vector machine (SVM) classifier is proposed. Since the performance of the conventional SVM classifier is strongly relied on parameters optimization,which can be trapped into a local minima, a hybrid differential evolution and flower pollination algorithm (DEFPA) is proposed. Using the collected data that has gained from different weather conditions, the proposed DEFPA approach has not only improved the efficiency of SVM classifier, but also classification accuracy as different weather conditions are considered.At the end of this dissertation, the summary of the overall work is presented and some possible future work is given.