The intelligent wheelchair behavior control method based on hierarchical POMDP model

In view of the problem that the application of the partially observable Markov decision process (POMDP) model in intelligent wheel chair behavior control is limited due to the difficulty in determining the model parameters, the paper proposes a method for control of intelligent wheelchair behaviors based on a hierarchical POMDP model. The method considers the nondeterministic actions and uncertainty of the state, and decomposes the actions into a set of small ones that can be done more efficiently. The model can account for both state transition and observation uncertainty. This behavior control method based on hierarchical POMDP was implemented onboard an intelligent wheelchair in the context of interactive service and navigation tasks. Finally the course of experiments conducted in an assisted living facility demonstrated that it could autonomously provide service and information for users, and the analyses of the experimental results verified the effectiveness and real-time property of the method.