Conclusion

In recent years, wireless technology has experienced significant developments. More and more wireless applications are realized in daily life.

Oriented to the wireless technology, the thesis covers three research areas:

wireless communication technology, Bluetooth; wireless architecture, wireless sensor network; wireless application, indoor positioning system. From the basic wireless communication technology to the actual application, the thesis provides a complete solution for constructing the next-generation positioning systems based on scattered wireless sensors.

As a wireless communication technology, Bluetooth faces a time delay problem. Before communications, Bluetooth devices must form a network. Its specification provides piconet only including up to 8 devices and not sufficient for applications. Another kind of network is defined and named scatternet, which connects piconets together to form extended network. Inside a scatternet, communications are carried out through paths, formed a number of hops between devices. If a path includes too many hops, time delay is generated.

Moreover, routing process to seek a path also increases the time delay.

To shorten the time delay, a new Bluetooth scatternet structure Solidring is designed in the research. The most significant characteristics of Solidring are

that it is a 3D (3-dimension) structure and organized regularly. Because of its 3D structure, Solidring significantly decreases the number of hops in communication paths and shortens the time delay. As Solidring is formed regularly, the location of each device can be easily predicted. As a result, the routing algorithm is simplified and the time delay is shortened. The two characteristics shorten the time delay remarkably.

Simulations are carried out to compare Solidring with two conventional structures, tree and ring. Two parameters of average path length and average time delay are evaluated. The simulation results show that the average path length in Solidring is only 50%-90% of those in tree and ring. The average time delay in Solidring keeps no more than 50% of those in tree and ring. As a short summary, a new Bluetooth scatternet structure Solidring is proposed in the thesis and it is proved that Solidring shortens the time delay remarkably.

Wireless sensor network organizes and manages the network composed by scattered wireless sensors. It provides regulations of how the wireless sensors form a network and what sensors should do. Besides highly efficient data transmission, wireless sensor network requires low power and high stability.

Low power makes the network work longer, while high stability ensures it to achieve high performance during the whole lifetime. However, conventional architectures cannot suffice the two requirements simultaneously. The research concentrates on designing a new architecture to keep stability and prolong lifetime.

A new architecture LESCS is proposed in the thesis. LESCS is the abbreviation of low-energy static clustering scheme, expected to keep stability

and prolong lifetime for wireless sensor network. Two architectures, static clustering and dynamic clustering are referred for designing LESCS. Essentially, LESCS is a kind of static clustering, as the clusters are never reformed after all sensors distributed to clusters initially, keeping the stability. Inside each cluster, one sensor is assigned as cluster head, which is changeless. Another sensor is assigned as gateway to transmit data for all sensors inside the cluster. Cluster head reassigns sensors as gateway frequently to balance energy consumptions of sensors and prolong the lifetime, similar to dynamic clustering. As a result, LESCS keeps stability and prolong lifetime for wireless sensor network.

In the simulations, LESCS is compared with the conventional static clustering and dynamic clustering. The results show that LESCS outperforms the two in term of lifetime. The average lifetime of all sensors in LESCS keeps up to 150% of that in the conventional dynamic clustering. Considering some applications need all sensors keep working, the lifetime of the first failing sensor is also compared. Such lifetime in LESCS keeps 150%-200% of that in the conventional dynamic clustering. In addition, as the clusters are never reformed, LESCS keeps the stability, same as the conventional static clustering.

As a short summary, a new wireless sensor network architecture LESCS is proposed in the thesis and it is proved that LESCS keeps the stability and prolongs the network lifetime simultaneously.

A wireless application, indoor positioning system is studied in the research as well. Indoor positioning system is essentially a wireless sensor network and a necessary supplement for GPS (global positioning system) that is only available outdoors. There are obstacles in indoor circumstance, concealing objects. In

conventional solutions, more sensors are deployed to eliminate undetectable area resulting from obstacles. However, such solutions increase the cost and make the system more complex.

In the research, a new solution is proposed to detect concealed objects by using reflected signals, without deploying more sensors. An ultrasonic system is designed, including a server, a controller, three ultrasound transmitters and an ultrasound receiver. In the designed system, through a controller, the server synchronizes three transmitters to emit ultrasonic signals to the receiver, which is the detected object in the system. Once receiving the signals, the receiver sends the received signals to the server. If the direct signals are received, the direct distances are calculated by using the propagation times of the direct signals from the transmitters to the receiver, same as the conventional solutions.

If the receiver is concealed and only receives the reflected signals, through geometrical relationship, the direct distances can be converted from the propagation distances of reflected signals as well. With the direct distances from three transmitters to the receiver, the server estimates the position of the receiver.

Moreover, because the reflected signals are easy to be interfered by noises, ultrasonic signals are modulated by BPSK (binary phase shift keying) to make the reflected signals more recognizable.

Experiment is carried out to check the detection accuracy of BPSK. The experimental results show that the detection error of BPSK keeps less than 1cm for detecting both direct and reflected signals, proving that BPSK is reliable and sufficient. Based on the experimental results, simulation is carried out to check the accuracy of the new solution. The simulation results prove that the error of

the new solution keeps no more than 0.25 cm for detecting both unconcealed and concealed objects, sufficient for indoor positioning system. As a short summary, by using reflected ultrasonic signals instead of deploying more ultrasonic transmitters, the proposed new solution can detect concealed object and is proved to be accurate enough for indoor positioning system.

As the final conclusion, the research concentrates on the optimization of wireless sensor network and its application of positioning systems. An optimized structure is designed to realize highly efficient communications for scattered wireless sensors. An optimized architecture is proposed to keep stability and prolong lifetime for wireless sensor network. A new solution to detect concealed objects is introduced for indoor positioning system. These proposals are considered as favorable contributions for constructing the next-generation positioning systems based on scattered wireless sensors.