Font Size: a A A

Research On Full Coverage Path Planning And Tracking Technology For USV

Posted on:2021-04-17Degree:MasterType:Thesis
Country:ChinaCandidate:D R ZhaoFull Text:PDF
GTID:2392330602489083Subject:Power electronics and electric drive
Abstract/Summary:PDF Full Text Request
Area full-coverage navigation technology is widely used in the field of resource survey,information collection and intelligent cleaning in the marine field.It is an important part of the control technology for unmanned surface vehicle(USV).In practical applications,the marine environment is intricate and complex.The design purpose of this navigation technology is to enable USV to increase the area coverage and reduce the path repetition rate as much as possible while ensuring navigation safety.This article takes the navigation control system of the whole area coverage of the USV as the research object,and studies the path planning and path tracking control technology of the USV under the actual sea background.In view of the characteristics of the subject,this paper specifically divides the system into three core links:the establishment of a sea area information environment model,global coverage path planning,and precise path tracking control technology.The main work of this article is as follows:1.Build a grid map for the actual sea application background.In this paper,the grid method is used to divide the mission area into sub-areas,then the environmental information of each grid is determined according to the aerial image,a series of expansion treatments are performed on the obstacles,and the mission sea area is finally divided into navigable areas and obstacle areas to complete the model construction of the mission area.2.According to the characteristics of the area full-coverage path planning task,a improved path planning algorithm based on the original biological inspired neural network(BINN)algorithm is designed.In this paper,by proposing a new best decision formula and redefining the BINN model,it effectively solves the difficulty in decision-making due to the same activity value in both directions caused by the original BINN algorithm and the area cannot be covered due to the low activity value,and it will eventually reach 100%coverage of the whole region.At the same time,by designing high excitation options in the external input of the neuron,it is possible to perform priority coverage search on a particular sea area,which further enhances the practicality of the algorithm in this paper.3.Design the path tracking algorithm according to the complex and changeable characteristics of the marine environment.Based on the grid coverage sequence determined above,this paper proposes a joint control algorithm with ADRC controller as the core and fuzzy controller adjustment parameters,which can effectively prevent trajectory deviation,improve navigation efficiency,and enhance system robustness.4.Construct a holonomic system to complete the simulation experiment verification for real complex sea area.According to the timeline design algorithm flow,a complete USV full area-coverage navigation control system is constructed by combining the three core links mentioned above,and the effectiveness and superiority of the system are verified by simulation experiments conducted in the actual sea area.After a series of targeted simulation experiments,the results show that the environmental modeling method used in this paper can effectively describe all environmental information and locate accurately.The path planning algorithm designed can greatly reduce the path repetition rate and effectively reduce the number of turn s of ensuring regional coverage.Under the control of the path tracking algorithm,the USV can still accurately follow the planned path under the interference of wind,waves and currents with small deviation.All of these reflect the effectiveness,superiority and practicability of the overall system designed in this paper.
Keywords/Search Tags:area full-coverage path planning, path tracking, improved biological inspired neural network, F-ADRC, environment model based on aerial image
PDF Full Text Request
Related items