| Ship-based fire control radar is weaponry on ships dedicated to track sea surface or aerial targets and provide target coordinates for weapon command and control systems. Since the 1940 s, ship-based fire control radar, constituting a complete weapon system together with ship-based weapons, has been highly valued by the navies in different countries. Its performance has been greatly improved in decades to make it essential in the ship-based combat system, in which the radar undertakes tasks including passive detection, long-range surveillance, searching, tracking, target indication and fire control guidance. According to the development requirements of the signal processor for ship-based fire control radar, the signal processing algorithms and specific project implementation are researched with the ADSP TS201 hardware platform in this paper.Firstly, signal processing patterns are separately introduced for search mode and tracking mode. Key signal processing algorithms involved are also presented in the order of signal processing such as pulse compression algorithm, MTD algorithm, CFAR detection algorithm, trace points aggregation algorithm and angle measurement algorithms of amplitude comparison monopulse. Computer simulation of these theoretical algorithms gives the MATLAB simulation results and detailed analysis which show the MATLAB simulation is correct. Low-altitude multi-frequency M/N detection algorithm is also briefly discussed in this paper because the detection of low-altitude targets may be affected multipath effects and a multifrequency method is reuired to determine the targets.Secondly, the paper introduces the system composition and hardware platforms of the signal processing processor, including the internal structure of the signal processing board, hardware resources, junction topology between the eight DSP chips and the performance of TS201 DSP chip. Data transmission and storage methods are accordingly analyzed. The processing time of main signal processing modules is also calculated and recorded, under the requirements of which the task allocation and timing diagram of each DSP chip are provided. The paper also gives the signal processing flow chart and corresponding DSP processing result analysis in DSP project implementation. DSP process proves to be correct through comparing the DSP results with the MATLAB simulation results.Finally, the paper discusses the upgrading methods of TS201 DSP assembler from fouraspects, namely: â‘ instructions in parallel, using X and Y calculation block at the sametime; â‘¡ using software pipelining; â‘¢ simultaneous application of double integer ALUand memory access of pair registers and four registers; â‘£ reducing the delays in theprogram. DSP program is written in C language and assembly language in theVisual DSP ++ 5.0 software environment. In order to improve the execution speed of theprogram, taking into account the readability, the main program is written in C language,and core subroutines are written in assembly language. Therefore the executionefficiency of assembly language is particularly important for the entire real-time signalprocessing system. The above four steps can be used for self-check in the process ofwriting assembler. The paper also gives an example of program optimization to comparethe program efficiencies previous and post CFAR algorithm optimization. The resultsshow that the software efficiency is increased by about 47.77% under the same DSPmemory. Finally, the paper also analyzes the junction communication problemencountered in project commissioning and gives the effective solution. |
