Second-Order Cone Relaxation and Adaptive Weights Enhanced DV-Hop Localization Algorithm in WSN

To tackle the issue that the DV-Hop localization algorithm exhibits poor localization performance in non-uniform networks and the non-convexity of the established localization model leads to the inability to solve convex optimization directly. Therefore, this article puts forward an improved DV-Hop localization approach. First, an adaptively weighted localization model is established to mitigate the adverse effects of non-uniform node distribution on positioning accuracy. Subsequently, through the introduction of auxiliary variables, the initial optimization model is converted into a linear objective function accompanied by non-convex constraints. Then, the non-convex constraints are reformulated as convex constraints using second-order cone relaxation, followed by an iterative solution process employing the interior point method. Under the same network environment configuration, Simulation experiment indicates the effectiveness of the presented algorithm in localization accuracy and enhances by 38.94%, 31.05%, 23.01%, 10.55% and 10.78% over DV-Hop, PSO-DV-Hop, GSM-DV-Hop, CVX-DV-Hop and MinMax-DV-Hop in square network, respectively.