Publicação
Indoor positioning system for wireless sensor networks
| Resumo: | Positioning technologies are ubiquitous nowadays. From the implementation of the global positioning system (GPS) until now, its evolution, acceptance and spread has been unanimous, due to the underlying advantages the system brings. Currently, these systems are present in many different scenarios, from the home to the movie theatre, at work, during a walk in the park. Many applications provide useful information, based on the current position of the user, in order to provide results of interest. Positioning systems can be implemented in a wide range of contexts: in hospitals to locate equipment and guide patients to the necessary resources, or in public spaces like museums, to guide tourists during visits. They can also be used in a gymnasium to point the user to his next workout machine and, simultaneously, gather information regarding his fitness plan. In a congress or conference, the positioning system can be used to provide information to its participants about the on-going presentations. Devices can also be monitored to prevent thefts. Privacy and security issues are also important in positioning systems. A user might not want to be localized or its location to be known, permanently or during a time interval, in different locations. This information is therefore sensitive to the user and influences directly the acceptance of the system itself. Concerning outdoor systems, GPS is in fact the system of reference. However, this system cannot be used in indoor environment, due to the high attenuation of the satellite signals from non-line-of-sight conditions. Another issue related to GPS is the power consumption. The integration of these devices with wireless sensor networks becomes prohibitive, due to the low power consumption profile associated with devices in this type of networks. As such, this work proposes an indoor positioning system for wireless sensor networks, having in consideration the low energy consumption and low computational capacity profile. The proposed indoor positioning system is composed of two modules: the received signal strength positioning module and the stride and heading positioning module. For the first module, an experimental performance comparison between several received signal strength based algorithms was conducted in order to assess its performance in a predefined indoor environment. Modifications to the algorithm with higher performance were implemented and evaluated, by introducing a model of the effect of the human body in the received signal strength. In the case of the second module, a stride and heading system was proposed, which comprises two subsystems: the stride detection and stride length estimation system to detect strides and infer the travelled distance, and an attitude and heading reference system to provide the full three-dimensional orientation stride-by-stride. The stride detection enabled the identification of the gait cycle and detected strides with an error percentage between 0% and 0.9%. For the stride length estimation two methods were proposed, a simplified method, and an improved method with higher computational requirements than the former. The simplified method estimated the total distance with an error between 6.7% and 7.7% of total travelled distance. The improved method achieved an error between 1.2% and 3.7%. Both the stride detection and the improved stride length estimation methods were compared to other methods in the literature with favourable results. For the second subsystem, this work proposed a quaternion-based complementary filter. A generic formulation allows a simple parameterization of the filter, according to the amount of external influences (accelerations and magnetic interferences) that are expected, depending on the location that the device is to be attached on the human body. The generic formulation enables the inclusion/exclusion of components, thus allowing design choices according to the needs of applications in wireless sensor networks. The proposed method was compared to two other existing solutions in terms of robustness to interferences and execution time, also presenting a favourable outcome. |
|---|---|
| Autores principais: | Silva, Hélder David Malheiro |
| Assunto: | Indoor positioning wireless sensor networks received signal strength body effect propagation model stride and heading system attitude and heading reference system sensor fusion posicionamento em ambiente interior redes de sensores sem fios potência de sinal recebido efeito do corpo modelo de propagação navegação inercial pedestre orientação, fusão sensorial Engenharia e Tecnologia::Engenharia Eletrotécnica, Eletrónica e Informática |
| Ano: | 2016 |
| País: | Portugal |
| Tipo de documento: | tese de doutoramento |
| Tipo de acesso: | acesso aberto |
| Instituição associada: | Universidade do Minho |
| Idioma: | inglês |
| Origem: | RepositóriUM - Universidade do Minho |
| Resumo: | Positioning technologies are ubiquitous nowadays. From the implementation of the global positioning system (GPS) until now, its evolution, acceptance and spread has been unanimous, due to the underlying advantages the system brings. Currently, these systems are present in many different scenarios, from the home to the movie theatre, at work, during a walk in the park. Many applications provide useful information, based on the current position of the user, in order to provide results of interest. Positioning systems can be implemented in a wide range of contexts: in hospitals to locate equipment and guide patients to the necessary resources, or in public spaces like museums, to guide tourists during visits. They can also be used in a gymnasium to point the user to his next workout machine and, simultaneously, gather information regarding his fitness plan. In a congress or conference, the positioning system can be used to provide information to its participants about the on-going presentations. Devices can also be monitored to prevent thefts. Privacy and security issues are also important in positioning systems. A user might not want to be localized or its location to be known, permanently or during a time interval, in different locations. This information is therefore sensitive to the user and influences directly the acceptance of the system itself. Concerning outdoor systems, GPS is in fact the system of reference. However, this system cannot be used in indoor environment, due to the high attenuation of the satellite signals from non-line-of-sight conditions. Another issue related to GPS is the power consumption. The integration of these devices with wireless sensor networks becomes prohibitive, due to the low power consumption profile associated with devices in this type of networks. As such, this work proposes an indoor positioning system for wireless sensor networks, having in consideration the low energy consumption and low computational capacity profile. The proposed indoor positioning system is composed of two modules: the received signal strength positioning module and the stride and heading positioning module. For the first module, an experimental performance comparison between several received signal strength based algorithms was conducted in order to assess its performance in a predefined indoor environment. Modifications to the algorithm with higher performance were implemented and evaluated, by introducing a model of the effect of the human body in the received signal strength. In the case of the second module, a stride and heading system was proposed, which comprises two subsystems: the stride detection and stride length estimation system to detect strides and infer the travelled distance, and an attitude and heading reference system to provide the full three-dimensional orientation stride-by-stride. The stride detection enabled the identification of the gait cycle and detected strides with an error percentage between 0% and 0.9%. For the stride length estimation two methods were proposed, a simplified method, and an improved method with higher computational requirements than the former. The simplified method estimated the total distance with an error between 6.7% and 7.7% of total travelled distance. The improved method achieved an error between 1.2% and 3.7%. Both the stride detection and the improved stride length estimation methods were compared to other methods in the literature with favourable results. For the second subsystem, this work proposed a quaternion-based complementary filter. A generic formulation allows a simple parameterization of the filter, according to the amount of external influences (accelerations and magnetic interferences) that are expected, depending on the location that the device is to be attached on the human body. The generic formulation enables the inclusion/exclusion of components, thus allowing design choices according to the needs of applications in wireless sensor networks. The proposed method was compared to two other existing solutions in terms of robustness to interferences and execution time, also presenting a favourable outcome. |
|---|