This thesis titled "Comparative Study and Evaluation of Path Planning Algorithms in 2D Environment" submitted by Muhammad Mamunur Rashid,. in the Department of Software Engineering at Daffodil International University has been accepted as satisfactory in partial fulfillment of the requirements for the degree of Bachelor of Software Engineering and approvals for style and content. I declare that I wrote my thesis under the supervision of Mr.
They also declare that neither this thesis nor any part of it has been submitted elsewhere for the award of any degree. I would like to thank you very much for your support and understanding over the past eight months. Imtiaz-Ud-Din was my Assistant Professor of Artificial Intelligence in 9th semester at Daffodil International University.
His teaching style and enthusiasm for the subject made a strong impression on me and I have always carried positive memories of his classes with me. I must also thank my two teachers Mr. There are many uses of path planning algorithms in search-related applications, including robotic exploration, game design, artificial intelligence, traffic route navigation, network routing, and path analysis are most extensive. Comparative study of three path planning algorithms Dijkstra, Greedy BFS and A-Star algorithm to find out which algorithm provides the shortest path for a particular 2D environment with an obstacle.
Based on the simulation results, this literature finds that A* provides the best optimal path instead of the Dijkstra, Greedy BFS algorithm.
- Background
- Motivation of the Research
- Problem Statement
- Research Questions
- Research Objectives
- Research Scope
- Contribution
- Thesis Organization
There are many algorithms that can solve this problem, the most common and well-known ones are Dijkstra search, greedy BFS and A Star search algorithm. To identify all nodes surrounding it, a brute force approach would be to start with one node. Switch to one of the adjacent nodes and repeat this process until the goal node is found.
There will be so many obstacles and also a start and target node algorithm must figure out the path and path must be optimal. There are so many algorithms that can plan for path planning, but they are not always optimal. Optimal is the most adequate for UAV control problem. So need to find out which algorithm is best for optimality, although optimality depends on the environment.
To find out the best algorithm among Dijkstra search, greedy BFS and A Star search algorithm. The scope of this paper is to compare three algorithms (dijkstra, greedy BFS and A*) and all the algorithms are applied only on 10X10 grid environments. Establishment of A* algorithm is the best path planning algorithm in 2D static environment among dijkstar, pure heuristic and A* algorithms.
- Route Planning
- USV Avoiding Underwater Obstacles
- Performance comparison
- Demand control
- A* algorithm in Traffic Navigational System
- Evaluation of multiple algorithm
- Summery
From the problem of a critical power deficit due to low water levels in Zambia that has suppressed the economy. The government is also making efforts to overcome by diversifying into renewable solar energy, but it was not an optimal solution and also the current solar photovoltaic (PV) systems are not able to produce enough energy. In this literature, a system has been proposed to claim control in PV systems by using the Best First Search algorithm.
As a result, the proposed system is efficient to optimize the user's priorities and limit the energy consumed by finding the best combination of household appliances. To determine the best combination of these household appliances, the best-first search algorithm with the appropriate heuristic value (it's actually an A* search) is used. And it was also optimal and not the Breadth First Search algorithm to find the best combination of home appliances (Chimanga, 2016).
The A* algorithm is used to achieve high performance driving navigation by analyzing the capabilities of maturity, optimality, time complexity, and space complexity in a variant evaluation function (Jingang, 2010). Recently, researchers have been attracted to unmanned aerial vehicles (UAVs) because of their large potential civilian applications. One of the key issues for robot navigation is that the robot must have the ability to “detect and avoid”.
The most challenging part of the navigation is avoiding the noise while accounting for sensor noise, uncertainties in operating conditions, and real-time applicability. The purpose of this paper is the evaluation of widely known UAVs path planning and obstacle avoidance algorithm based on heuristic and non-heuristic method on three different global and local obstacle scenarios. And it will provide a clear vision to the reader for which algorithm is best for which obstacle scenarios.
In the above literature review there are many uses of dijkstar, greedy heuristic and A*.
- Heuristic Function
- The Dijkstra’s algorithm
- Greedy Best First Search or pure heuristic search Algorithm
- A* Algorithm
- Environment Setup
The adjacent node will store in open list and must select an adjacent node from the open list, the process will continue until the destination is found (Kang, 2008. But which adjacent node to select from the open list, proceeding with the algorithm First, set the starting node in the OPEN list and then calculate the cost function f (n).
Moving to the CLOSED list from the OPEN list, which node has the smallest cost function f (n). Determine all neighboring nodes of n and calculate the cost function f (n) for each neighboring node that is not in the CLOSED list. Associate the calculated costs with each neighbor list that is not in the OPEN or CLOSED list and move them to the OPEN list by placing pointers on n (n is the parent node).
To each adjacent that is already OPEN, associate the smaller of the cost values just calculated and the previous cost value. This algorithm first exposition by Peter Hart, Nils Nilsson and Bertram Raphael in 1968 (Peter E.Hart, 1968) (wikipedia, 2018). It is the combination of Dijkstra and Greedy's best first search. Go to one of the following nodes and repeat this process until the target node is found.
However, this method does not always guarantee the best path to the goal and is computationally intensive. For example, if the robot knows that the target is to the west, explore nodes west of your current location (Iram Noreen, 2016). If the robot knows the direction of the target robot, it can always try to go to the node in that direction.
To determine which node to select, A* uses the distance between the current node and the destination node and moves to the node with the smallest distance between neighboring nodes. A neighboring node can be determined by evaluating the cost to the target node from both nodes. The OPEN list stores all sequential paths that have yet to be explored, and the CLOSED list stores all paths that have been explored.
To trace the path from the goal to the starting node, it is used at the end, generating the optimal path. Redirected back to the starting node to obtain the full path once the destination node is reached, the parent nodes are found.
- Scenario Number: 1
- Scenario Number: 2
- Scenario Number: 3
- Scenario Number: 4
- Discussion
In the above figure and result proved that while obstacle is less, each algorithm will travel the same number of nodes, but while the obstacle density is more, A-star algorithm will show the more optimal than dijkstra and greedy BFS.
Findings
Recommendations for Future Works
A new path planning algorithm using a GNSS localization error map for UAVs in an urban area. A Greedy Approach in Path Selection for DFS-Based Maze Map Discovery Algorithm for an Autonomous Robot. Application of A* algorithm for real-time path replanning of an unmanned surface vehicle avoiding underwater obstacles.
