Robust positioning and optimization techniques in mobile robotic networks

Abstract

The increasing public demand for positioning and navigation services has generated a need for higher performance global navigation satellite systems (GNSSs). New applications of global positioning system (GPS) are currently expanding far beyond its traditional focus and new challenges to the current GPS system are presently occurring due to these expansions. To overcome such challenges, GPS receivers are expected to be improved using higher performance tools such as robust acquisition, tracking, and positioning techniques. Large signal attenuation, multipath, and radio frequency interference are some examples of impediments that the new applications have inherited. The emergence of new navigation applications, increased the need to develop path planning methods. Achieving nearly fully autonomous unmanned navigation requires the integration of many technologies such as path planning, communication, and various navigation techniques. Furthermore, in scenarios involving multiple resources and multiple objectives, the resource management task becomes complicated for human and needs to be automated. This thesis, firstly derives the probability of having false peak detection in GPS signal acquisition which can occur under weak signal conditions, namely in the threshold region. This is based on novel hypothesis test to analyze GPS detection performance in weak signal conditions. The probability of outlier for different detection techniques is also obtained using Monte Carlo simulation. When the probability of outlier is larger than zero, two novel robust positioning techniques are proposed. The first technique is based on outlier detection and the second technique is based on using a genetic algorithm (GA). Using the proposed techniques, the probability of finding a reliable solution is considerably increased. The second part of the thesis investigates novel optimization techniques for mobile robotic networks. Three multi-objective optimization techniques are developed to optimize path and coverage for unmanned vehicles. The first two techniques are based on GA and dynamic programming, while the last technique is based on ant colony optimization. In addition, a cooperative motion planning technique based on using chaos synchronization is proposed to decrease the redundant search attempts and maximize the autonomous coverage. Furthermore, a resource management optimization method using fuzzy analytic hierarchy process is developed to ease the decision making process.