The dissertation is devoted to increase of reliability indoor navigation systems, due to the use of wireless data transmission systems and methods of verification of embedded systems.
The modified integrated method of determining the current position on the map of the buildings is based on the analysis of the signal from iBeacon and accelerometer, modified by the Kalman method, which allows to quickly adjust the value of the current position, and allows to reduce the error.
The route correction method was modified, which, unlike to the existing ones, uses a Kalman filter and a jump point method, which allows checking the error variation location on the map with the error level 4.7%.
One of the most important tasks to be solved in the creation of indoor navigation systems is to ensure their quality. The developed method allows to increase the accuracy, but one of the effective tools which helps to improve quality is verification. Therefore, there was set a task to develop own verification method.
For verification there was build a neural-fuzzy system, which is represented as a multilayered neural network, each layer responsible for a specific step of the logical inference algorithm. The usage of this model makes it possible to extract the knowledge base as a whole, which makes the implementation more flexible.
The system verification criteria were distinguished: Euclidean space metric, Mankovsky space metric, backup reliability factor, Kalman filter error, angles determination, accelerometer acceleration, coordinate acceleration errors, ling errors.
The method of verification of indoor navigation systems is developed on the basis of Kohonen neural networks and fuzzy logic, which, unlike the existing ones, allows to choose a set of decisive rules, rebuild the structure at retraining, use linguistic rules.
Information technology is developed that could receive input data by using: BLE 4.0, map representation, video, inertial navigation and voice messaging. It is then possible to determine the current position and verify the current state of the indoor system
The method of the reliability assessment of indoor navigation systems is modified by the usage of the integrated method of determining the current position and the method of neuro-fuzzy verification, which allows to evaluate the probability of trouble-free operation of indoor navigation systems.
The main scientific result of the research aimed to solve the scientifically practical problem of improving the reliability of indoor navigation systems. Developed navigator integrated into the mobile application of the Smart-Campus system, which provides support for students, staff and visitors to the University. The introduction of the voice navigator makes it easier for people with visual impairments to adapt to social benefits.
