CAJ | 학술논문

为了实现在高相似度环境中移动机器人精确高效的自定位与建图，提出了一种基于粒子群优化（ PSO）的Rao-Blackwellized粒子滤波同步定位与地图构建（ SLAM）算法。利用激光扫描数据校正里程计信息，得到多模态的似然函数，克服相似环境对机器人定位的影响；利用粒子群优化算法提高常规粒子滤波器的估计性能，使得高似然采样集向各个后验概率密度分布取值极大的区域运动，同时保持低似然粒子多样性，从而在一定程度上克服粒子贫乏问题，并且显著地降低精确定位所需的粒子数。对所提算法与Gmapping算法在MIT数据集上进行仿真对比实验，结果表明了该算法的可行性和有效性。
위료실현재고상사도배경중이동궤기인정학고효적자정위여건도，제출료일충기우입자군우화（ PSO）적Rao-Blackwellized입자려파동보정위여지도구건（ SLAM）산법。이용격광소묘수거교정리정계신식，득도다모태적사연함수，극복상사배경대궤기인정위적영향；이용입자군우화산법제고상규입자려파기적고계성능，사득고사연채양집향각개후험개솔밀도분포취치겁대적구역운동，동시보지저사연입자다양성，종이재일정정도상극복입자빈핍문제，병차현저지강저정학정위소수적입자수。대소제산법여Gmapping산법재MIT수거집상진행방진대비실험，결과표명료해산법적가행성화유효성。
In order to realize a precious and efficient Simultaneous Localization and Mapping ( SLAM) of a mobile robot in the environments with high similarities, a Rao-Blackwellized particle filter SLAM algorithm based on Particle Swarm Optimization ( PSO) was proposed. By correcting odometer information with the laser scanning data and acquiring a multimode likehood function, the influences of the similar environment to the robot localization were overcome. The performance of the generic particle filter was improved by PSO. Particles in high likelihood sampling set moved to the region with posterior probability distribution maximum value, meanwhile the algorithm maintained the multiplicity of the low likehood particles. The improverishment of the particle filter was overcome partly and the number of required particles was reduced observably. The simulation experiments were performed for the comparison of the proposed algorithm with Gmapping algorithm on public MIT datasets, and the feasibility and effectiveness of the algorithm are validated.