To address the limitations of the Sparrow Search Algorithm (SSA) in autonomous vehicle path planning, such as insufficient optimization accuracy, low efficiency, and susceptibility to local optima, an Improved Sparrow Search Algorithm (ISSA) is proposed. This approach initializes the population using a quadratic compound chaotic map to enhance initial population diversity. An adaptive inertia weight mechanism based on the Logistic function is integrated into the position update of discoverers to balance global and local search capabilities. Additionally, a nonlinear perturbation factor-based cosine algorithm is introduced to improve the position update of followers, enhancing the search speed and accuracy, and improving the ability to escape local optima. Experimental results demonstrate significant optimization effects in terms of both optimization accuracy and efficiency, validating the effectiveness of the ISSA in autonomous vehicle path planning.