CAJ | 학술논문

针对目前移动机器人同时定位与地图创建(SLAM)研究中多采用激光雷达或超声环作为测距传感器导致系统复杂、成本高的问题,提出了一种利用舵机带动单超声传感器扫描的低成本设计方案.在高斯超声模型基础上,利用贝叶斯公式对栅格地图进行概率更新,并结合Sobel边缘检测算法提取特征点.实现了由不确定的移动机器人坐标系向固定的以环境特征点为原点的全局环境坐标系的转换及全局定位,为在相同环境下通过重复实验进行多地图融合研究奠定了基础.该低成本移动机器人设计的有效性通过实验得以验证.
침대목전이동궤기인동시정위여지도창건(SLAM)연구중다채용격광뢰체혹초성배작위측거전감기도치계통복잡、성본고적문제,제출료일충이용타궤대동단초성전감기소묘적저성본설계방안.재고사초성모형기출상,이용패협사공식대책격지도진행개솔경신,병결합Sobel변연검측산법제취특정점.실현료유불학정적이동궤기인좌표계향고정적이배경특정점위원점적전국배경좌표계적전환급전국정위,위재상동배경하통과중복실험진행다지도융합연구전정료기출.해저성본이동궤기인설계적유효성통과실험득이험증.
Laser radar and ultrasonic sensors ring are usually used as range sensors for mobile robot SLAM (simultaneous localization and mapping),while lead to the unavoidable problem of system complexity rising and high-cost.A low-cost mobile robot design using one ultrasonic sensor which is rotated by a steering gear is proposed.The probability of occupancy grid maps is updated based on Gaussian ultrasonic sensor model and Bayes formula.By implementing Sobel image edge detection algorithm,the environment map is generated and its feature points are extracted.The uncertain mobile robot coordinate is converted to the fixed global environment coordinate which selects the environmental feature points as the origin.The global localization are achieved and it paves the way for multi-maps fusion in the same environment by repeated experiments.The validity of the proposed low-cost mobile robot design is proved by experiments.