| Island ecosystems have simple biological chains,low natural resilience,and environment fragile.If they are destroyed,it is very difficult to recover.In order to preserve the island ecosystem,the utilization of USV proves to be an effective approach due to its ability to provide precise observations of the island ecosystem over extended periods.However,the specific geographic conditions of islands and reefs,with variations in tidal levels,necessitate manual setting of searching paths for most USVs is used in island observation.Therefore,it is essential to conduct research on USV path planning algorithm for observing islands and reefs affected by tides to facilitate long-term planning and sustainable development of these environments.Firstly,this thesis presents the development of a island tidal environment map for USV path planning.In large scale environments,the conventional approach to environmental modeling presents a challenge in achieving coordinated modeling accuracy with the number of map raster.To address this issue,this study utilizes a variable scale raster map model.This model establishes different grid scales for the USV no-navigation area,risk area and safe area,based on water depth data of islands.Furthermore,the tidal environment map of islands and reefs is created in accordance with the tidal level change data,which varies with time.Experimental findings indicate that the variable scale raster map model significantly reduces the number of raster maps while maintaining map accuracy in risk areas.Secondly,this study involves the application of a region segmentation method to decompose the observation paths of islands and reefs,which exhibit a strong spatio-temporal coupling effect.Due to the strong spatio-temporal coupling effect caused by the task access sequence,the selection of entry and exit points in the task area,and the dynamic changes of tides that occurred during the observation of islands and reefs in tidal environments,it is practically difficult to work out the solution straightaway.To solve this problem,the region segmentation method is proposed that can be used to divide the area to be observed around the island and reef into a set of path nodes surrounding them from the temporal and spatial perspectives.In addition,to address the issue that ant colony algorithm is prone to local optimum in the task of island and reef observation,while preventing path intersection and repeated observation in the planned path,the probability of entry and exit point selection is factored into the state transition rule.Apart from that,in order to find the solution to the path of island and reef observation,the local optimization strategy is developed in accordance with the spatial characteristics of points within the island and reef.Experimental results demonstrate that the improved ant colony algorithm is more effective in selecting the entry and exit points and is capable of adapting to tidal changes.Finally,an improved artificial potential field algorithm is applied to achieve local path planning for island observation mission.In the process of local path planning under the island and reef environment,the artificial potential field algorithm tends to ignore the issue of feasible paths in the concave area of islands and reefs.In light of this,the island and reef potential field forces are constructed in this study by taking into account the safe water depth boundary of the island and reef for expanding the observation area in the concave area of the island and reef.Meanwhile,the current artificial potential field algorithms ignore continuity when solving local minimum problems,which causes path oscillations and the sudden changes in path angle at the local minimum,thus making it difficult to meet the constraints on USV motion.To solve this problem,the smoothness of the path is improved in this study by introducing the correction force to avert the local minimum.Experimental findings demonstrate the effectiveness of the proposed algorithm. |
