In this paper, we present a method for estimation of objects' positions and orientations in an indoor environment. For ubiquitous computing, it is important to realize provision of a location-aware information service for humans. Especially, orientation information can reflect a more detailed context than that obtained merely according to a location: people standing face-to-face or back-to-back have vastly different contexts. We are tackling indoor location and orientation estimation using a new approach called topological estimation, which specifically examines relative relationships of humans rather than their absolute positions. In the initial implementation of topological estimation, we have used infrared tags as sensors to collect real-world related information. This paper describes an ultrasonic and radio device system to collect objects' local relationships.