Ensuring Consistent Local System State in Future Outdoor Lighting Networks

From Master Projects
Jump to: navigation, search


About Ensuring Consistent Local System State in Future Outdoor Lighting Networks

  • This project has been fulfilled.
  • This project fits in the following Bachelor programs: {{#arraymap:|, |xXx|bachelorproject within::xXx|,}}
  • This project fits in the following masterareas: {{#arraymap:Parallel and Distributed Computer Systems|, |xXx|project within::xXx|,}}


Description

Advances in sensing and communication technologies have enabled the move from traditional outdoor lighting systems to intelligent systems. This move is motivated by the problems in traditional outdoor lighting systems. Such problems include light pollution, high energy costs, inability to adapt to their environment, and contribution to CO2 emissions. Some efforts have already been made to improve traditional outdoor lighting systems resulting in solutions like Philips Starsense, Dynadimmer, and Lumimotion. However, they do have shortcomings. In this work, we assume that future wireless outdoor lighting systems (traffic-adaptive) will use cheap and unreliable sensors. This unreliability can lead to an inconsistent local system state in terms of lighting behavior. Therefore, there is a need to investigate and develop traffic-adaptive protocols and algorithms that avoid such inconsistency in lighting behavior. The main question that we are addressing is how to achieve application-level reliability in a traffic-adaptive outdoor lighting system that relies on unreliable sensors. In other words, how to achieve correct actuation behavior for street lamps using unreliable sensing input? We propose five different traffic-adaptive algorithms that use local information and information from the neighboring light poles to take well-informed control decisions for actuating street lamps. Additionally, we propose a simple behavior-based distributed reputation system and study its impact on system performance. Finally, we evaluate our proposed algorithms in different scenarios and present the results.