A hybrid recurrent neural network and optimization framework for intelligent mobile robot navigation in smart manufacturing

In the third dimension prospective of production industries taking on smart manufacturing principles, the integration of automation and digitalization revolutionizes conventional processes, unlocking heightened productivity and operational efficiency. This endeavour implicates coordinating unified interactions among machines and human operators, capitalizing on their unique strengths and capabilities. In this study, a multi-objective optimal navigation system tailored for mobile robots operating in dynamic surroundings, leveraging hybrid optimization algorithms. Primarily, introduce the modified animal's migration optimization (MAMO) algorithm, which measures obstacle state data essential for dynamic obstacles. This facilitates proactive collision avoidance, thus minimizing unnecessary disruptions. Consequently, deep features are extracted from all feasible paths spanning the target region connecting the origin and destination points. These path features are then subjected to the chaos locust search (CLS) algorithm, which determines multiple paths to consider. In addition, the hypercube search with recurrent neural network (HS-RNN) is employed to locate the optimal path while removing redundant alternatives among the multiple choices, thereby refining the path planning process. Simulation outcomes highlight the better performance of the proposed system in optimal path generation compared to alternative approaches, validating its efficacy in dynamic environments.