Grid-scan-based multi-hop localization algorithm for wireless sensor networks

Due to single reference information priority, poor network topology adaptability, low positioning accuracy and high computational complexity, the multi-hop localization approaches for wireless sensor networks have many limitations in application. In view of this, a novel grid-scan-based multi-hop localization algorithm (GSML) is proposed. We construct a more realistic weighted constrained model for multi-hop localization and give the rule of weight assignment. Based on the intersections of bounding square rings, the normal nodes can estimate their feasible regions as small as possible with only a few floating-point operations. On that basis, the approximately optimal values of normal nodes´ coordinates can be obtained through a lightweight grid-scan procedure and further improved by the collaboration between neighboring nodes. Extensive simulations show that GSML outperforms the existing typical schemes in aspects mentioned above.