In big data management platform of public security organ,an intelligent search engine based on PSO ( Particle Swarm Optimi-zation) algorithm is studied,which solves the current search engines' problem that cannot be very good for users to find their interested information and improves the search efficiency. Inverted sequence file segmentation storage method is adopted in big data management platform,significantly improving the efficiency of segmentation. In order to avoid a large number of data interaction between the client and the server,resulting in network congestion,use innovative data synchronization technology. The establishment of client database,gen-erate the client file tree,significantly reducing the amount of data communication between client and server. According to the characteris-tics of public security organ's big data,the PSO algorithm is introduced to intelligent search engine. Finally,develop a big data manage-ment platform and intelligent search engine for the public security organs. And through experiments verify the accuracy of search results and satisfaction of users.

In order to remove the bias,a noise ICA algorithm based the noise model is introduced and the multi-user detection method based this algorithm is deduced. The algorithms are simulated to compare the performance of the two kinds of multi-user detection methods which are based the normal ICA algorithm and the noise ICA algorithm. Simulation results indicate that noise ICA algorithm method has better performance.

In the spread spectrum communication systems,the conventional scheme of slide-correlation acquisition has a disadvantage that it needs a long time for the systems to acquire the synchronization with the received PN code sequence,especially when the sequence length is longer.In order to solve this problem,this paper proposes an improved scheme of code acquisition,which achieves PN code acquisition by three parallel circuits,the improvement is that the scheme isn't confined by the code length,but also effectively reduces false alarm probability,and the circuit is simple.The theoretical analysis and simulation results show the improved scheme shortens systemic acquisition time.

In the multi-path fast-fading channel,the independent component analysis(ICA) algorithm could not be convergent and the result would be deteriorative.In order to eliminate this disadvantage,dimension reduction is executed before ICA algorithm.The simulation results show that even in the case of small size of samples,the proposed method has better performance compared with multi-user detection based on ICA without dimension reduction.So the proposed method has much practicability.