Reinforced Whale Optimizer for Ground Robotics : A Hybrid Framework

Abstract

View Video Presentation: https://doi.org/10.2514/6.2023-1601.vid The paper presents a multi-objective optimization technique for the exploration of unknown space. In the context of robotics, exploration employees construction of the surrounding map using sensor information. Conventionally, for space exploration, the optimization is performed utilizing a single optimization technique with a particular objective function. Utilizing a mono objective function with a specific task of optimizing one particular aspect although simplifies and fasten the optimization process, but adversely effects the map accuracy and exploration depth. Realizing this vital aspect, this paper introduces an optimization technique with multi-objective functions which are simultaneously optimized. This not maximizes the search process but also increases the map accuracy. The proposed framework termed as Reinforced Whale Algorithm (RWO), is based on bio-inspired Whale Optimizer. It starts with the initialization of the whale's population, which are referred to as way-points. These way-points are assumed to be constant once they are set in the initial stage/iterations. The next step involves the position update from the non-dominated way-points catered by the robots. The waypoints are optimized by the algorithm. The performance matrices are carefully analyzed through extended simulations mimicking different conditions environment. After determining the isolated benefits of the algorithm, the results efficacy is then demon started by comparing its performance with three contemporary optimization algorithms namely Coordinated Multi-Robot Exploration (CME), and conventional Whale Optimizer (WO), and Arithmetic Optimizer (AO) integrated with CME. Results indicates that the proposed algorithm greatly enhance the optimization process by enhancing the explored area and reducing the search time.