Recently, autonomous vehicle technologies have been widely developed. Moreover, these technologies are applied to autonomous carts for security guards, guidance robots, delivery carts, etc. A variety of human detection technologies are important for these robotic carts. There are many studies on avoiding collisions with humans and other obstacles. However, technologies to distinguish between humans and other obstacles are especially important for security guard robots. In this paper, we propose a standing human detection method using 2D-LiDARs to reduce costs. We show the details of our methods and results of evaluation.

In recent years, different robotic carts have been developed, and are suitable for use as transport, guide, and patrol robots on campuses or in buildings, or for event venues such as trade shows, with digital signage displays. These small robots move autonomously using a location information system, obstacle detection system, and maps. We focus on a patrol system for finding suspicious but harmless people such as aged wanderers. These suspicious people are found about once a year or a few times during events. Therefore, autonomous robots are useful for deterrence and to find the suspicious people. We assume that the patrol robots have global navigation satellite systems (GNSS) for their location information system, laser imaging detection and ranging (LiDAR) for obstacle detection, an infrared sensor for human detection, and a Bluetooth device to detect authorized humans. Moreover, we assume that event venues are outdoors and that there are many temporary booths. Therefore, simultaneous localization and mapping (SLAM) technologies are required. In this paper, we describe a localization method in an event space for SLAM. Moreover, we show the evaluation result of our proposed method.

This paper considers a problem of user and access point associations in heterogeneous wireless networks consisting of multiple radio access technologies. The current association policies are mostly based on a selfish user behavior optimizing their own decision criteria. Such simple strategies, however, might fail especially in a congested place with a large number of users. From this perspective, this study propounds a strategy for optimizing the problem that aims at sustainably and fairly maximizing the minimal utility of users. The minimal utilities can be obtained from bare essentials of the quality of experience and defined by considering different requirements of each user. In order to analyze characteristics of the proposed selection strategy, we conduct simulations comparing it with the other selection method under each evaluation criterion. Numerical results suggest that the proposed selection i) enhances the total minimal utility in sustainable and fair manners, ii) indicates a distinctive trend in load balancing and iii) requires a comparable computational time and handover frequency in contrast with the other selection.

Localization protocol is important for estimating node positions in a wireless multi-hop network. Routing protocol is also important for controlling paths. In previous research, localization and routing protocols have been discussed and evaluated separately. In this paper, we propose an integrated protocol for the optimized link state routing (OLSR) and OLSR based localization (ROULA). Our protocol enables simultaneous localization and routing. ROULA's localization is performed using OLSR overhead such as hello packets and routing tables. The routing overheads and the processing procedures can be efficiently integrated. We demonstrate that the integrated protocol for ROULA and OLSR enables simultaneous localization and routing.

Modeling of ad hoc sensor networks becomes difficult when uncertain features of the network increase. Deterministic modeling is difficult and some stochastic arguments should be introduced instead. In this paper, we introduce the concept of random networks. One remarkable feature of random graphs is that a phase transition occurs as the probability of edge connection increases. At the critical probability, fragmented pieces of edges suddenly start to be mutually connected, forming one large component. This graph-theoretical change parallels phase transitions in states of mutter, e.g. the jump from water to ice. In ad hoc sensor networks, the nodes are connected by wireless links. In order to meet this requirement, we propose a model using percolation, a kind of random graph where the edges are formed only between the nearby nodes. We also present some numerical examples to simulate the jump effects of the phase transition.