Xavier Araújo

Conclusions

This dissertation presented a networking solution to increase the capacity and the lifetime of CSMA based WMSNs. From the analysis of the state of the art solutions, regarding these topics, it was possible to see the different existent approaches to achieve these goals. The strategy followed in this dissertation to surpass these problems, was the adoption of a multiple communication channel scheme. This strategy allowed to increase the capacity of the network, by reducing its interference levels, and it also allowed to extend the network lifetime, by using the channel assignment algorithm proposed in this dissertation, eTILIA. eTILIA is a dynamic, centralized and energy-aware channel assignment algorithm, based in the TILIA algorithm described in [2], which makes the channel assignment decision based on the energy levels of the nodes, trying to avoid the traffic forwarding through nodes with low energy level.
Besides eTILIA, this dissertation also proposed an architecture for a low-cost, solar powered video-surveillance systems, based in Wi-Fi, to cover large unconnected areas such as parks or beaches. This proposal was based on the WMSN concept, and covered the physical components, the protocol stack and the mode of operation of the network.
The evaluation of the algorithm was made through computer simulation, using the NS-3 tool. This approach allowed to verify and validate the eTILIA algorithm, without the deployment of a complete test-bed, which would be very costly and time-consuming. This experimental evaluation showed that eTILIA can increase the lifetime of a WMSN up to 4% when compared to TILIA, and up to 7% when compared to a random channel assignment procedure. Besides this, the experimental evaluation also showed that eTILIA can manage more efficiently the energy resources of the network, since it normally improves the mean node lifetime and significantly increases the number of alive nodes, at the moment of the network failure.
These are the most relevant conclusions of this dissertation, which proposed an algorithm that can lead to energetically improved WMSNs.

Future Work

In this dissertation the evaluation of eTILIA was made through computer simulation. As it is common knowledge, computer simulation can provide accurate results, but cannot be compared to real world testing. Computer simulation is based on mathematical models of the real world, and since these models are not perfect, the simulation results obtained can often be different from the results in real world conditions. Due to this, the first thing to do in order to improve this work, would be to test eTILIA in a real world situation. This would require the deployment and configuration of a WMSN. To do this it is possible to take advantage from the architecture described in Chapter 3, to deploy a low-cost, solar powered video-surveillance system. With this system it would be possible to obtain the required data to evaluate the eTILIA algorithm. Then, it could be done a statistical analysis of the data, in order to precisely characterize the behaviour of the eTILIA algorithm.
Another improvement to this work would be to upgrade eTILIA. One modification that could be made in the eTILIA algorithm would be to take into consideration, not only the energy level of the nodes, but also the ratio between the energy spent and the energy being harvested at a certain moment. This could bring advantages in terms of energy efficiency.