AUV path planning method based on improved sparrow search algorithm

To address the challenges of complex underwater environments, particularly the uncertainties in ocean currents, this study proposes an improved sparrow search algorithm (ISSA) for autonomous underwater vehicles (AUVs) path planning. The goal is to enhance the efficiency and robustness of path planning by minimizing navigation time and improving path stability in uncertain conditions.

The proposed ISSA incorporates several key enhancements to the classic sparrow search algorithm (SSA). First, a vector analysis method is developed to evaluate interval responses, allowing the algorithm to effectively handle uncertainties in ocean currents. By modeling the uncertain ocean currents as intervals, the algorithm can accurately calculate the energy consumption and navigation times for different paths. Second, the ISSA introduces segmented learning and quantum mechanisms to improve global search capabilities. These mechanisms enable the algorithm to dynamically adjust its search strategy by learning from both elite and marginal individuals within the population, thereby enhancing diversity and preventing premature convergence. Third, a Cauchy-Gaussian mechanism is integrated into the update formula to balance global exploration and local exploitation during the search process. Finally, the population size is dynamically updated using Thompson sampling, allowing the algorithm to adaptively allocate computational resources based on the complexity of the environment.

Simulation results demonstrate that the ISSA significantly outperforms the original SSA and other state-of-the-art algorithms, such as particle swarm optimization (PSO), differential evolution (DE), artificial bee colony (ABC), and whale optimization algorithm (WOA). Specifically, the ISSA reduces the average maximum navigation time by 49.88% compared to the original SSA and decreases the failure rate in extreme ocean current conditions by 10.6%. The ISSA also exhibits superior convergence speed, achieving near-optimal paths in approximately 20 iterations, while other algorithms require around 40 iterations to approach the global optimum. Moreover, the ISSA shows a lower average fitness value, indicating better optimization performance and path planning efficiency.

The ISSA demonstrates strong global search capabilities and robustness in dynamic and uncertain underwater environments, making it a promising solution for AUV path planning. The improvements in convergence characteristics and the ability to handle complex ocean currents highlight the algorithm's potential for practical applications. Future work will focus on further optimizing the computational efficiency of ISSA and exploring its application in more diverse and challenging underwater scenarios.