System analysing pedestrian movement in city environment

Motivation and summary

Advanced system for pedestrian movement analysis become necessary during development Spirit Guide - interactive tourist support system. The Spirit Guide is a system providing voice guide service for tourists. The provided content depends on location and according to user's movements audio stream switches to one or another.

One of the problems to be solved is positioning with high accuracy. We have number of positioning system, and every system has its advantages and disadvantages. Using GSM etworks we can determinate position with accuracy about 200–400 meters in big city. Another issue is pattern of position switching. After jump from one base station to another coordinates change to new point in a couple of hundreds meters away what is no good for analysing algorithm.

GPS (or A-GPS) accuracy is affected by a number of factors, including satellite positions, noise in the radio signal, atmospheric conditions, and other barriers to the signal such as buildings. GPS can give accuracy up to 10 meters with good conditions. But other problems of GPS positioning are time of cold start which can be up to 20 minutes and big power consumption.

Also GPS can lose signal or go to "bad" state, especially at small streets. In this research we are going to investigate possible solutions of problems described above using combination of GSM, GPS, user interaction, city map and introducing a mathematical model based on behavioural patterns.

Project goals

1. Make an overview of existing positioning solutions and technologies
2. Determine and analyze problems of those technologies
3. Propose functional specification for mathematic model which would collect data from different positioning technologies, combine them, analyze and predict possible user's movements
4. Implement prototype of the system based on mathematic model described in 3.

Timeline and deliverables

1. The short-term (9–12 month) deliverables are following:
2. An article describing current state of positioning technologies and our approach to solving this problems
3. An article describing mathematic model which will be able to solve these problems
4. A prototype of the system which uses combination of existing positioning technologies and implements mathematical model described in article 2

Project team and contacts

Alexander Buntakov, student at SUAI, project leader

Alexander Sidorenko, PhD student at SUAI