Body attenuation and path loss exponent estimation for RSS-based positioning in WSN

• Autor(a) principal: Silva, Hélder David Malheiro
• Data de Publicação: 2017
• Outros Autores: Afonso, José A., Rocha, Luís Alexandre Machado
• Tipo de documento: Artigo
• Idioma: eng
• Título da fonte: Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos)
• Texto Completo: http://hdl.handle.net/1822/46566
• topic: Received signal strength; Body effect; Propagation model; Wireless sensor networks; Positioning; CRLB; Engenharia e Tecnologia::Engenharia Eletrotécnica, Eletrónica e Informática; Science & Technology

The influence of the human body in antenna systems has significant impact in the received signal strength (RSS) of wireless transmissions. Accounting for body effect is generally considered as being able to improve position estimation based on RSS measurements. In this work we perform several experiments with a wireless sensor network, using a sensor node equipped with an inertial measurement unit (IMU), in order to obtain the relative orientation between the sensor node and multiple anchor nodes. A model of the RSS attenuation induced by the body was created using experimental measurements in a controlled environment and applied to a real-time positioning system. A path loss exponent (PLE) estimation method using RSS information from neighbor anchors was also implemented and evaluated. Weighted centroid localization (WCL) algorithm was the positioning method used in this work. When the sensor node was placed on the user’s body, accounting for body effect produced negligible improvements (6%) in the best-case scenario and consistently degraded accuracy under real conditions, whether the node was placed on the user’s body (in the order of 3%), 10 cm away (from 14% to 35%) or 20 cm away from the body (from 42% to 105%) for results in the 70th percentile. The PLE estimation method showed improvements (in the order of 11%) when the sensor node is further away from the body. Results demonstrate that the distance between sensor node and the body has an extremely important influence on the accuracy of the position estimate.