The Research Of Radar Intelligence Coding Technology In Parallel And Implementation On Android

The reconnaissance radar equipment is playing an important role in modern war.Aviation radar, shipboard radar and portable radar is togethered to be a large network. Dueto complexities of the battlefield environment，the radar transmitted the intelligenceinformation can only be through wireless channel. However, the Radar intelligenceinformation is very large, and the wireless channel bandwidth is limited. Therefore, thetransmission will be slower. The main means of solving this problem is to compress theintelligence data, but the compression will also takes time. In order to reduce thecompression time and send the intelligence data to the command and control center foranalysis processing as soon, in this paper, the parallel processing capabilities of multi-coreprocessor are exploited to compress the radar information before transmittion.Firstly, in this paper, the fundamental principle of compression theory, which called thelimit entropy, is described. The arithmetic coding and decoding algorithm is expounded,At the same time, the fork-join model、parallel program tructure and task schedulingcriterions of OpenMP are introduced in detail. Because of Android’s openness, easiness forkernel pruning, and excellent performance in the mobile terminal equipment, it is selected asthe portable radar terminal system. The Android NDK advantage in computing is comparedwith SDK, and the programming environment of NDK is built.Secondly, arithmetic coding is processed in parallel making use of its owncharacteristics, and an alternative method is proposed computing with large numbers insteadof normalization. Then, the encoding output rules and decoding rules are put forth. Theparallel programming is accomplished to compress data based on the theory of parallelprefix sum, but the implemention is resultful on the premise of that the CPU has more thanfour cores.Thirdly, arithmetic coding is processed in parallel using the data decomposition tocompress the radar intelligence data. The implementation of integer arithmetic coding isfocused on, and the move-to-front probality model, which is inefficient, in WNC algorithmis replaced by improved BIT structure, so that the arithmetic coding will be more efficient.Finally, OpenMP has been not applicated in the Android system yet, because thecompilation environment does not support. This paper is committed to use OpenMP onAndroid successfully. Then, two methods using OpenMP technology on Android system areprovided. The arithmetic coding is transplanted On Android supporting OpenMPsuccessfully and the test result is given in the end. As a matter of fact, it is very efficient for the compression of radar intelligence data.