中央研究院 資訊科學研究所

The building/environment data and information system (BeDIS) described in this report is a part of infrastructure needed to support location-specific, active emergency preparedness and responses within large buildings. BeDIPS (Building/environment Data and information based Indoor Positioning System) is one of its components. BeDIPS can provide people in large buildings with sufficiently accurate location data. It is scalable, disaster resilient and is easy to configure, deploy and maintain. BeDIPS works without Internet and serves both smart phones and most legacy Bluetooth devices. The other component of BeDIS is BeDi mist, a virtual repository of data and information on the building, interior layouts and facilities. The mist uses micro data servers and smart gateways to deliver fine-scale, location-specific decision support data on a timely basis to hundreds and thousands of active devices and mobile applications.

The purpose of the present article is to examine the essence of what has commonly been described as a ‘‘projective line’’, but which is here named a ‘‘meridian’’. This shall be done in several papers: this first paper devoted to the meridian itself, the second to the character and form of the family of projective isomorphisms of one meridian onto another and the third to some connections between meridians and higher dimensional projective space. In this first paper we shall view the meridian from various points of view: (1) as a set acted upon by a family of involutions; (2) as a set acted upon by a 3-transitive group of permutations; (3) as a set with a quinary operator; (4) as an equivalence class of quadruples, relating to the cross ratio. In the final section of this first paper we shall investigate how the existence of a certain single-valued exponential on a meridian is characteristic of the meridian corresponding to the field of real numbers.

This paper examines the algebraic and geometric structures inherent in the family of homographies of a one dimensional projective space (or meridian).

A libra L is essentially a group without any particular element of L distinguished as an identity. The paper sets forth the fundamental facts about libras, characterizes which libras arise from the sets of projective isomorphisms from one meridian onto another, and shows how the structure of the meridian (projective line) may be constructed from such libras.

The paper also examines the geometric structure of a circular hyperboloid T in real 3-dimensional projective space, and identifies and describes the defining permutations (in the sense of Klein’s Erlanger Programm).

Kelly McKennon (November 2017)

　　This paper describes the development of an indoor robotic surveillance system based on reactive wall following and human detection built on intelligent differential-drive mobile robot navigation. For this development of a surveillance mobile robot system that is assistive to human surveillance, we employ a commercially available wheeled mobile robot equipped with infrared sensors, sonar sensors and human detection sensors for perceiving the local indoor environment and detecting human presence. The system has two components: reactive wall following and human detection. The wall following is performed by a two-input (distances) single-output (heading angle) interval type-2 fuzzy logic system for heading direction control of a constant speed unicycle with online decision rules specifically designed for the task of stable right wall following. Proper reaction to the presence of an unexpected human is enabled in the wall-following course by human detection sensors. The design and implementation of an autonomous mobile robot human-aware surveillance system using reactive right wall following and human detection for two indoor surveillance tasks is demonstrated to work well in our lab testing.

Background: Extracting sequential patterns, in which an event α happens before another event β with high probability, from a time-stamped event database is an important task in data science. In cohort studies, which are frequently conducted on data in such databases, it is essential to find a matching between the chosen cases and the control candidates to identify the causality when the control candidates that cases can match are different. The hidden causality is investigated by observing several statistical indicators in the returned matched results. Methods: We present an efficient exact matching method that matches each case with exactly K controls such that a control candidate can be matched with at most one case, where K is an input constant. Results: Our experiment results show that the method is fast and memory efficient, even though the database contains a large number of events. Moreover, the quality of matchings found is high, i.e., a control candidate is matched with all matchable cases with roughly equal probability when multiple matchings are performed. We posit that to obtain stable statistical results, K should be set to be around, but not larger than, the maximum matching ratio, which is the largest integer L. Then each case can be matched with exactly K controls rather than a fixed constant as proposed in previous studies. In addition, to ensure that the results are stable, several such matchings should be performed to find the average, instead of performing the matching just once.