| With the development of science and the improvement of people's quality of life in recent years,mobile robot is one of the research fields of robotics.Mobile robots can not only discover areas that humans cannot reach,but also help people complete boring or heavy tasks in daily life.As one of the research fields of robotics,path planning has been the focus of scholars' research,and some important research results have been achieved so far.The embedded platform is widely used in the development platform of mobile robots due to its special features,low power consumption and small size.At the same time,due to the small core of the embedded platform,its processing power,storage space and computing resources are relatively limited.So,improving the performance of path planning is particularly important in embedded environments.The A-star algorithm is a heuristic search algorithm which is widely used to solve the static global path planning problem because of its simplicity and convenience,strong operability,and high accuracy.However,the traditional A-star algorithm still has the disadvantages of long paths and many inflection points,especially when there are many obstacles because of its node search strategy.In this paper,aiming at the inadequacy of the traditional A-star algorithm,an improved A-star algorithm which extended the search neighborhoods and weighted the heuristic function was proposed.First,a two-dimensional environment model is established by the point cloud information obtained by vision sensors.Then,an improved heuristic function is proposed to a weighted measure.Next,in order to improve the efficiency of path planning,the minimum binary heap is applied to the data storage structure of list OPEN of A-star algorithm.Finally,the traditional eight neighborhoods of the A-star algorithm are expanded upon multilayers in order to add multiple search directions.At the same time,do the research on the impact of different extended layers on the path length and path planning time,to determine the path planning optimization strategies.Based on the consideration above,the improvement A-star algorithm is compared with the traditional A-star algorithm by simulation experiments.The experiment results show that compared with the traditional A-star algorithm,the improved A-star algorithm has better performance on the path length,smoothness and pathfinding time,and can be applied to situations with many obstacles.This paper proposes a path planning solution based on the algorithm above,and applies it to the Jetson Nano embedded platform.It uses the point cloud information taken by sensors to carry out environment model and establishes a two-dimensional environment model with grid method.Finally,using the current coordinates of the robot as a starting point and a preset end point to realize global path planning.The map and the movement path of the robot is dynamically updated during the real-time movement of the robot. |
