Although the Basic RRT algorithm is considered a traditional search method, it has been widely used in the field of robot path planning (manipulator and mobile robot), especially in the past decade. This algorithm has many features that give it superiority over other methods. On the other hand, the Basic RRT suffers from a bad convergence rate (it takes a long time until finding the goal point), especially in environments with cluttered obstacles, or whose targets are located in narrow passages. Many studies have discussed this problem in recent years. This paper introduces an improved method called (Hybrid RRT-A*) to overcome the shortcomings of the original RRT, specifically slow convergence and cost rate. The heuristic function of A-star algorithm is combined with RRT to decrease tree expansion and guide it towards the goal with less nodes and time. Various experiments have been conducted with different environment scenarios to compare the proposed method with the Basic RRT and A-star under the same conditions, which have shown remarkable performance. The time consumed to find the path of the worst one of these scenarios is about 4.9 seconds, whereas it is 18.3 and 34 for A-star and RRT, respectively.
Path planning is an essential concern in robotic systems, and it refers to the process of determining a safe and optimal path starting from the source state to the goal one within dynamic environments. We proposed an improved path planning method in this article, which merges the Dijkstra algorithm for path planning with Potential Field (PF) collision avoidance. In real-time, the method attempts to produce multiple paths as well as determine the suitable path that’s both short and reliable (safe). The Dijkstra method is employed to produce multiple paths, whereas the Potential Field is utilized to assess the safety of each route and choose the best one. The proposed method creates links between the routes, enabling the robot to swap between them if it discovers a dynamic obstacle on its current route. Relating to path length and safety, the simulation results illustrate that Dynamic Dijkstra-Potential Field (Dynamic D-PF) achieves better performance than the Dijkstra and Potential Field each separately, and going to make it a promising solution for the application of robotic automation within dynamic environments.
Face recognition is the technology that verifies or recognizes faces from images, videos, or real-time streams. It can be used in security or employee attendance systems. Face recognition systems may encounter some attacks that reduce their ability to recognize faces properly. So, many noisy images mixed with original ones lead to confusion in the results. Various attacks that exploit this weakness affect the face recognition systems such as Fast Gradient Sign Method (FGSM), Deep Fool, and Projected Gradient Descent (PGD). This paper proposes a method to protect the face recognition system against these attacks by distorting images through different attacks, then training the recognition deep network model, specifically Convolutional Neural Network (CNN), using the original and distorted images. Diverse experiments have been conducted using combinations of original and distorted images to test the effectiveness of the system. The system showed an accuracy of 93% using FGSM attack, 97% using deep fool, and 95% using PGD.