In this chapter we present the ROS implementation of our Bayesian User Model, BUM. BUM is a distributed user modelling technique that can be easily implemented in several system topologies. It is able to infer the characteristics of multiple users from heterogeneous data gathered by multiple devices, such as social robots, ambient sensors and surveillance cameras. This chapter presents the BUM process and its implementation, emphasizing the essential and advanced ROS concepts used and extended to achieve the modularity and flexibility needed. Instructions on how to achieve our experimental set-ups are also presented, including a discussion on the role of ROS in the experimental success of the system, and illustrations of the results that can be achieved with our technique. This chapter serves as a thorough description and tutorial for the usage of our package, which can now be useful to the scientific community in user modelling and user-adaptive systems.
CITATION STYLE
Martins, G. S., Santos, L., & Dias, J. (2019). bum_ros: Distributed User Modelling for Social Robots Using ROS. In Studies in Computational Intelligence (Vol. 778, pp. 531–567). Springer Verlag. https://doi.org/10.1007/978-3-319-91590-6_15
Mendeley helps you to discover research relevant for your work.