受人观测启发的UUV前视声纳滤波方法

UUV依据前视声纳探测信息进行在线滚动路径规划。针对浅水环境下声纳探测信息受物理及声场环境干扰杂波较多的问题,提出了一种受人观测启发的前视声纳滤波方法。首先,建立了声纳视域模型,然后采用小波阈值法进行声纳信息的初步去噪;其次,引入仿人观测的真伪目标判别策略,提出了基于时间滑窗的探测目标模糊关联与真伪判别算法。湖试结果表明:该算法能够滤除浅水环境下声纳探测信息的杂波,可提供可信环境信息,并引导UUV避开岛屿、浮箱等典型障碍目标。

Forward looking sonar filtering method for UUV inspired by humanoid observation

UUV carries out online rolling path planning according to forward-looking sonar observation information. Aiming at the problem that the sonar detection information is heavily disturbed by physical and sound field environment interference clutters under shallow water environment, a forward-looking sonar filtering method inspired by humanoid observation is proposed. First of all, the sonar-vision model is built. Then, the wavelet threshold method is used to conduct the preliminary denoising of the sonar information. Next, the target authenticity discrimination strategy based on humanoid observation is introduced, and the fuzzy association and authenticity discrimination algorithm of the detection target based on time sliding window is proposed. Lake trial experiment was conducted, and the result shows that the proposed algorithm can filter the clutters contained in the forward looking sonar detection information under shallow water environment, provide credible environment information for UUV to carry out online path planning, and guide UUV to avoid typical obstacle targets such as islands and pontoons.