Active Simultaneous Localization and Mapping Method Based on Model Prediction

In the process of controlling an unmanned vehicle, it is practically important that under conditions of rapidly changing dynamic constraints, control laws be developed that would be optimal with respect to a given quality functional or a multicriteria functional. When static and dynamic constraints do not allow the optimal movement to be chosen to a given quality functional, the authors consider the transition to another quality functional using the predictive integral path model and the method of active simultaneous localization and mapping. In this case, the strategy for choosing the state space is more efficient than the strategy for choosing the control space. The practical question is how to achieve this. The paper presents a method and experiments using an unmanned vehicle platform at a test site in the form of a complex environment, showing the feasibility of the method.