Computer vision-based path planning can play a crucial role in numerous technologically driven smart applications. Although various path planning methods have been proposed, limitations, such as unreliable three-dimensional (3D) localization of objects in a workspace, time-consuming computational processes, and limited two-dimensional workspaces, remain. Studies to address these problems have achieved some success, but many of these problems persist. Therefore, in this study, which is an extension of our previous paper, a novel path planning approach that combined computer vision, Q-learning, and neural networks was developed to overcome these limitations. The proposed computer vision-neural network algorithm was fed by two images from two views to obtain accurate spatial coordinates of objects in real time. Next, Q-learning was used to determine a sequence of simple actions: up, down, left, right, backward, and forward, from the start point to the target point in a 3D workspace. Finally, a trained neural network was used to determine a sequence of joint angles according to the identified actions. Simulation and experimental test results revealed that the proposed combination of 3D object detection, an agent-environment interaction in the Q-learning phase, and simple joint angle computation by trained neural networks considerably alleviated the limitations of previous studies.

Path planning for robot arms to reach a target and avoid obstacles has had a crucial role in manufacturing automation. Although many path planning algorithms, including RRT, APF, PRM, and RL-based, have been presented, they have many problems: a time-consuming process, high computational costs, slowness, non-optimal paths, irregular paths, failure to find a path, and complexity. Scholars have tried to address some of these issues. However, those methods still suffer from slowness and complexity. In order to address these two limitations, this paper presents a new hybrid path planning method that contains two separate parts: action-finding (active approach) and angle-finding (passive approach). In the active phase, the Q-learning algorithm is used to find a sequence of simple actions, including up, down, left, and right, to reach the target cell in a gridded workspace. In the passive phase, the joints angles of the robot arm, with respect to the found actions, are obtained by the trained neural network. The simulation and test results show that this hybrid approach significantly improves the slowness and complexity due to using the simplified agent-environment interaction in the active phase and simple computing the joints angles in the passive phase.

Intelligent robot arms are advanced robotic systems used in Industry 4.0 to perform complex tasks. Unlike conventional robot arms, which perform predefined tasks, intelligent robot arms have autonomy and can operate in changing environments, interact with other machines, and collaborate with humans. In this regard, adaptive path planning is crucial for intelligent robot arms, involving real-time environment monitoring and path generation to continuously update the robot’s trajectory based on changes in the surroundings. This paper presents an adaptive path planning method for intelligent robot arms to be used in dynamic environments. The proposed method is based on a hybrid active-passive approach and has been tested in a dynamic workspace simulation environment. The results indicate the ability of the proposed method to respond dynamically in a complex scenario where the target is fluctuating, and an obstacle is intentionally placed in the robot’s path. Additionally, real-time analysis results show that the method can be categorized as real-time path planning with less than 100 ms reaction time for grid sizes with less than 96×96×96 cells. This insight presents opportunities for the establishment of smart factories, smart homes, and smart cities, where the presence of intelligent robot arms in dynamic environments becomes essential.

Background: Demands for Precise positioning systems are dramatically increasing in various fields. Therefore, finding problems of these systems and solving them is very important. In this regard, scholars have tried to detect challenges with which precision positioners are faced. Some most important problems introduced by scholars are: short stroke, low dexterity, complicated structure, lack of compact size, low output force, low response, low resolution, etc. Methods: This paper presents an XY precision positioning system that the Impact Drive Mechanism (IDM) is used as the positioning system itself. This system is designed based on microscale steps of IDM which its Piezoelectric Actuator (PEA) is driven by saw-tooth shaped voltage pulses. Although IDM can solve aforementioned problems, it has nonlinear equations which may make a big challenge in its control. Therfore, this study presents a fuzzy controller that is capable to work well in such complex systems. Although using fuzzy controllers is very useful, it requires more computation time and cost. Thus, a square root controller is proposed in order to reduce them. Results: Results show that although the square root controller has the same result, it needs less computation time and cost than the fuzzy controller does. Conclusion: In conclusion, in this new design of the positioner, using a square root controller is more efficient and economical than the fuzzy one.

Background: Stick-slip actuators are commonly used in Nano/Micro precision positioning systems, but their control is challenging due to factors like nonlinear friction, PEA hysteresis, and uncertainty. Researchers have made efforts to address these challenges and documented their findings in articles and patents. Methods: This study introduces a novel vertical stick-slip actuator and proposes two different methods for overcoming the challenges associated with controlling it. The first method involves training an inverse model of the actuator using a supervised machine-learning algorithm to determine the optimal number of signals and peak voltage required for the saw-tooth signals in an open-loop controller. The second method is a closed-loop controller that utilizes the maximum allowable peak voltage unless the positioning error is smaller than the maximum step size. At this point, the neural network-based controller adjusts the peak voltage to a lower value, ensuring that the actuator reaches the desired position at the end of the final signal. Results: According to the results, both controllers perform effectively. The open-loop and closed-loop controllers exhibit a relative error of 1.59% and 0.4%, respectively, for an arbitrary desired position in the final position. Conclusion: In conclusion, the suggested controllers offer a practical solution to the controlling challenges faced by stick-slip positioners, which are essential in the advancement of Nano/Micro sciences.