Coverage optimization of wireless sensor networks (WSNs) faces challenges such as uneven node distribution and vulnerability to coverage blind spots. This paper introduces and improves the Northern Goshawk Optimization (NGO) algorithm: the Logistic chaotic map is adopted to initialize the population for enhanced ergodicity, a nonlinear dynamic weight is introduced to balance global exploration and local exploitation, and a Gaussian–Lévy hybrid mutation mechanism is integrated to strengthen the ability to escape from local optima. Experiments on standard test functions show that the improved algorithm (INGO) can stably approach the theoretical optimal values for both unimodal and multimodal functions. The convergence speed and solution accuracy are significantly superior to those of the original NGO, with a smaller standard deviation and stronger robustness. INGO is applied to the coverage optimization of 2D and 3D WSNs, with coverage rate as the fitness function, and the optimal node deployment coordinates are output through iterative optimization. Simulation results show that INGO achieves a best coverage rate of 98.32% in the 2D scenario, which is 7.72 percentage points higher than the 90.6% of NGO. In the 3D scenario, the best coverage rate reaches 72.32%, 6.78 percentage points higher than the 65.54% of NGO. Meanwhile, INGO yields more uniform node deployment and effectively reduces coverage blind spots. Its convergence curve is smooth and oscillation-free in the late iteration stage, and the stability is significantly better than that of NGO. With proper settings of population size and iteration times, INGO can achieve better coverage performance, providing a reliable technical solution for the efficient deployment of wireless sensor networks in complex environments.
Wang et al. (Sun,) studied this question.