Adaptive path planning for cognitive robot arms using AI and computer vision

Abstract

Smart robot arms are a big step forward in robotics, especially in Industry 4.0. These robots are different from regular ones because they can do complicated tasks really well. They can work on their own and are good at adapting to new situations. They can also work well with other machines and even with people. One important thing for these robots is their ability to plan their path as their end- effector moves toward the target point. This means they continuously monitor the workspace in real time and change their route if objects change. So they can keep working smoothly even if the situation around them changes. In this thesis, a new way for smart robot arms to plan their path has been developed. The method contains two main parts, including environment monitoring and path generation. In the first part, a combined YOLO object detection algorithm and neural network are used to detect and determine the 3D coordinates of each object in the workspace in real time. Additionally, in the path generation section, a path is updated based on the location of objects so that the robot can reach the target as well as avoid obstacles. To update the path in real time, we used a hybrid approach using D* Lite and neural networks. This method is made especially for robot arms that work in dynamic environments. It has been tested a lot in a computer program that simulates different situations. The results are very good and important. In a difficult situation where the target place keeps changing, and an obstacle is in the robot's way on purpose, this new method can quickly find a new way. This means the robot can change its path right away to avoid obstacles or reach a moving target. The research shows that this method is super-fast, taking less than 100 milliseconds to react in some cases. This speed is really important in situations where quick decisions can stop accidents or make things work much better, like for human-robot collaboration purposes. This research has big effects, not just for robots, but for many other things too. When we use smart robot arms in places that change a lot, like factories, homes, and cities, it can make a big difference. In smart factories, these robots can help make things faster and without stopping much. In smart homes, they can help people with different jobs, making life easier and more comfortable. Also, in smart cities, these intelligent robots can help with things like managing trash, public services, and healthcare, changing how these services work. Having these adaptable robots in these places is really important.