A Comparative Study Between Hopfield Neural Network and A* Path Planning Algorithms for Mobile Robot
TL;DR: This work compares one such algorithm namely the Hopfield neural network-based path planning algorithm with A* in a static environment and finds the A* algorithm fared better and theHopfield network showed promising results with scope for further reduction in its run time.
Abstract: Path planning is an important aspect of any mobile robot navigation to find a hazard-free path and an optimal path. Currently, the A* algorithm is considered to be one of the prominent algorithms for path planning in a known environment. However, with the rise of neural networks and machine learning, newer promising algorithms are emerging in this domain. Our work compares one such algorithm namely the Hopfield neural network-based path planning algorithm with A* in a static environment. Both the Hopfield network and the A* algorithm were implemented while minimizing the total run times of the programs. For this, both the algorithms were run in MATLAB environment and a set of mazes were then executed and their run times were compared. Based on the study, the A* algorithm fared better and the Hopfield network showed promising results with scope for further reduction in its run time.
Citations
More filters
TL;DR: The improved hybrid algorithm proposed in this study can effectively reduce distribution costs, shorten transportation routes, and improve optimization efficiency, indicating that the hybrid algorithm suggested is effective and practical when applied in the medical material distribution system.
Abstract: In order to explore the feasibility of using deep learning in the Wise Information Technology of 120 (WIT120) material distribution system under deep learning, in the study, first, the Hopfield neural network and simulated annealing (SA) algorithm are used to study the path optimization in the WIT120 material distribution system. Aiming at the poor optimization and robustness of the traditional Hopfield neural network and the low optimization efficiency of SA algorithm, in this study, a memory device is added to the SA algorithm, the distribution starting points of the two algorithms are fixed, and an improved hybrid algorithm is designed by combining Hopfield neural network and SA algorithm. Then, the improved hybrid algorithm of the traditional Hopfield neural network and SA algorithm is used in the actual medical material distribution system, and comparative analysis of the distribution route, cost, and scheduling time before and after using the proposed algorithm is performed. Besides, the difference between the proposed algorithm and other algorithms in path optimization and path planning efficiency is compared and analyzed. The results show that the improved hybrid algorithm proposed in this research reduces the total mileage of distribution by 46.9%, costs by 33.9%, and the time required for scheduling by 99% after applied in the medical material distribution system. It is also better than traditional analog algorithms and neural networks in distribution path optimization. The path optimization results show that the shortest path produced by the improved hybrid algorithm in this study is 27.85, which is superior to the traditional SA algorithm, Hopfield neural network, and genetic algorithm (the shortest paths obtained are 41.21, 44.63 and 36.48, respectively). Compared with other optimization algorithms, the path planning efficiency of the hybrid algorithm is the highest (the amount of distribution tasks completed in unit time is 22). In conclusion, the improved hybrid algorithm proposed in this study can effectively reduce distribution costs, shorten transportation routes, and improve optimization efficiency, indicating that the hybrid algorithm proposed in this study is effective and practical when applied in the medical material distribution system, which can provide reference value for medical material distribution and provide data support for the management of medical material distribution.
6 citations
TL;DR: In this article, the authors proposed an optimal control and ANN approach for a dynamic model of the multi-wheeled combat vehicle to generate the possible optimal paths that cover every part of the workspace.
Abstract: This paper presents a real-time path planning algorithm for autonomous multi-wheeled combat vehicles using Artificial Neural Network (ANN), Artificial Potential Fields (APFs) and optimal control theory. Real-time navigation of autonomous vehicles is a very complex problem and are crucial for many military operations. This paper proposes an optimal control and ANN approach for a dynamic model of the multi-wheeled combat vehicle to generate the possible optimal paths that cover every part of the workspace. Consequently, the obtained paths are used to train the proposed ANN model. The trained ANN has the capability to control the movement of combat vehicle in real time from any starting point to the desired goal position within the area of interest. The vehicle path is autonomously generated from the previous vehicle location parameter in terms of lateral velocity, heading angle and yaw rate of the vehicle. APF is proposed for preventing the vehicle from colliding with obstacles that represented in border destination. The effectiveness and efficiency of the proposed approach are demonstrated in the simulation results, which show that the proposed ANN model is capable of navigating the multi-wheeled combat vehicle in real time.
References
More filters
TL;DR: A real-time path-planning algorithm that provides an optimal path for off-road autonomous driving with static obstacles avoidance is presented and was applied to the autonomous vehicle A1, which won the 2010 Autonomous Vehicle Competition organized by the Hyundai-Kia Automotive Group in Korea.
Abstract: In this paper, a real-time path-planning algorithm that provides an optimal path for off-road autonomous driving with static obstacles avoidance is presented. The proposed planning algorithm computes a path based on a set of predefined waypoints. The predefined waypoints provide the base frame of a curvilinear coordinate system to generate path candidates for autonomous vehicle path planning. Each candidate is converted to a Cartesian coordinate system and evaluated using obstacle data. To select the optimal path, the priority of each path is determined by considering the path safety cost, path smoothness, and path consistency. The proposed path-planning algorithms were applied to the autonomous vehicle A1, which won the 2010 Autonomous Vehicle Competition organized by the Hyundai-Kia Automotive Group in Korea.
275 citations
TL;DR: It has been verified that the proposed method has better performance in convergence speed, solution variation, dynamic convergence behavior, and computational efficiency than the path planning method based on the genetic algorithm with elitist model.
134 citations
Journal Article•
TL;DR: This work revisits the benefits, in terms of travelling time, of path planning in marine environments showing spatial variability, and focuses on the application to a real environment of such techniques with autonomous underwater glider properties as the mobile platform.
Abstract: Autonomous Underwater Vehicles (AUVs) usually operate in ocean environments characterized by complex spatial variability which can jeopardize their missions. To avoid this, planning safety routes with minimum energy cost is of primary importance. This work revisits the benefits, in terms of travelling time, of path planning in marine environments showing spatial variability. By means of a path planner presented in a previous paper, this work focuses on the application to a real environment of such techniques. Extensive computations have been carried out to calculate optimal paths on realistic ocean environments, based on autonomous underwater glider properties as the mobile platform. Unlike previous works, the more realistic and applied case of an autonomous underwater glider surveying the Western Mediterranean Sea is considered. Results indicate that substantial energy savings of planned paths compared to straight line trajectories are obtained when the current intensity and the vehicle speed are comparable. Conversely, the straight line path betwe en starting and ending points can be considered an optimum path when the current speed does not exceed half of the vehicle velocity. In both situations, benefits of path planning seem dependent also on the spatial structure of the current field.
127 citations
02 May 1993
TL;DR: An algorithm is presented for planning the path of an object (robot) through an obstacle cluttered space by using a modified A* method for searching through the free space, with a large speed increase for a slight degradation in path length optimality.
Abstract: An algorithm is presented for planning the path of an object (robot) through an obstacle cluttered space by using a modified A* method for searching through the free space. The A* method can be extremely slow for large numbers of cells to be searched. The alternative presented is to use trial vectors that span several cells to aid in the planning. The fine resolution of the obstacle mapping is maintained. A loose search is performed on the fine grid. The paths that are found using this method are slightly sub-optimal. The speed of the searching algorithm is significantly increased. Solution times for three-dimensional problems are typically below 100 milliseconds, while the traditional A* search for the same problem is several minutes. The trade-off is a large speed increase for a slight degradation in path length optimality. Two-dimensional problems are solved very quickly. >
124 citations
01 Jan 2013
TL;DR: It is concluded that various works of research have been successfully applied PSO to solve the mobile robot path planning problem, due to its simplicity and efficiency in navigating large search spaces for optimal solutions.
Abstract: This study investigates the application of Modified Particle Swarm Optimization (MPSO) to the problem of mobile robot navigation to determine the shortest feasible path with the minimum time required to move from a starting position to a target position in working environment with obstacles. In this study, MPSO is developed to increase the capability of the optimized algorithms for a global path planning. The proposed algorithms read the map of the environment which expressed by grid model and then creates an optimal or near optimal collision free path. The effectiveness of the proposed optimized algorithm for mobile robot path planning is demonstrated by simulation studies. The programs are written in MATLAB R2012a and run on a computer with 2.5 GHz Intel Core i5 and 6 GB RAM. Improvements presented in MPSO are mainly trying to address the problem of premature convergence associated with the original PSO. In the MPSO an error factor is modelled to ensure the PSO converges. MPSO try to address another problem which is the population may contain many infeasible paths; a modified procedure is carrying out in the MPSO to solve the infeasible path problem. The results demonstrate that this algorithm have a great potential to solve the path planning with satisfactory results in terms of minimizing distance and execution time. Using the heuristic approach, the mobile robot can navigate safely among the obstacles without hitting them and reach the predefined target point. These techniques are also helpful for the solution of the local minima problem. Researchers have been seeking for more efficient ways to solve this problem, in the following section, the recent works on robot's navigation and path planning using particle swarm are reviewed. (4) provided an overview of the research progress in path planning of a mobile robot for off-line as well as on-line environments. Commonly used classic and evolutionary approaches of path planning of mobile robots have been discussed, and showed that the evolutionary optimization algorithms are computationally efficient. Also, challenges involved in developing a computationally efficient path planning algorithm are addressed. (5) proposed a modified particle swarm optimization algorithm for the robot path planning in dynamic environment, two parameters of particle-distribution-degree and particle dimension-distance are introduced into the proposed algorithm in order to avoid premature convergence. (6) provided an intelligent approach for the navigation of a mobile robot in unknown environments, the navigation problem becomes an optimization problem, and then it is solved by PSO algorithm. Based on position of goal, an evaluation function for every particle in PSO is calculated. It's assumed that Robot can detect only obstacles in a limited radius of surrounding with its sensors. Environment is supposed to be dynamic and obstacles can be fixed or movable. (7) proposed an Immune Particle Swarm Optimization (IPSO) algorithm for path planning of the mobile robot which based on the biological mechanism of the immune system. They compared the simulation results with PSO optimization results. They concluded that the optimal path and the execution time based on IPSO algorithm are reduced separately, and the improved PSO algorithm enhances the convergence speed and robustness of time-varying parameters. From the above literature review for the recently published paper, it is concluded that various works of research have been successfully applied PSO to solve the mobile robot path planning problem, due to its simplicity and efficiency in navigating large search spaces for optimal solutions.
28 citations