A Semidefinite Programming Approach for Obstacle Prediction and Localization

Localization of things is becoming increasingly important in many of today’s applications, especially in applications using wireless sensor networks (WSNs) for communication. Here, the most readily available radio information is the received signal strength (RSS). However, accurate localization of nodes based on the RSS in non-line-of-sight (NLOS) scenarios and in the presence of obstacles is challenging. Known approaches often suffer from model mismatch and strong multipath in NLOS. Inevitable noise and shadowing caused by obstacles significantly affect the accuracy of distance estimation, often making localization a very difficult task. Here, we present a robust semi-definite programming (SDP) method using RSS-values in noisy NLOS environments, which significantly improves positioning accuracy and allows for obstacle prediction.