Improving environment detection by behavior association for context-adaptive navigation

Navigation and positioning systems depend on both the operating environment and the behavior of the host vehicle or user. The environment determines the type and quality of radio signals available for positioning, and the behavior can contribute additional information to the navigation solution. In order to operate across different contexts, a context-adaptive navigation solution is required to detect the operating contexts and adopt different positioning techniques accordingly. This paper focuses on determining both environments and behaviors from smartphone sensors, serving for a context-adaptive navigation system. Behavioral contexts cover both human activities and vehicle motions. The performance of behavior recognition in this paper is improved by feature selection and a connectivity-dependent filter. Environmental contexts are detected from global navigation satellite system (GNSS) measurements. They are detected by using a probabilistic support vector machine, followed by a hidden Markov model for time-domain filtering. The paper further investigates how behaviors can assist within the processes of environment detection. Finally, the proposed context-determination algorithms are tested in a series of multicontext scenarios, showing that the proposed context association mechanism can effectively improve the accuracy of environment detection to more than 95% for pedestrian and more than 90% for vehicle.