Real-time and proactive navigation via spatio-temporal prediction

Abstract

We present a novel approach for real-time and proactive navigation in crowded environments such as event spaces and urban areas where many people are moving to their destinations simultaneously. Our challenge is to develop a real-time navigation system that enables movements of entire groups to be efficiently guided without causing congestion by making near-future predictions of people flow. Our approach tries to detect future congestion by using a spatio-temporal statistical method that predicts people flow. When future congestion is detected, our approach creates an optimal navigation plan based on "what-if" simulations, which accounts for the effect of total people flow change caused by navigation. We experimentally compare the spatio-temporal statistical method with the conventional matrix factorization based approach using a real data set. We also demonstrate the effectiveness of our navigation approach by computer simulation using artificial people-flow data.

Spatio-temporal multidimensional collective data analysis for providing comfortable living anytime and anywhere

“…For this purpose, we first employed a spatio-temporal statistical approach based on the kriging approach, which is commonly utilized in fields like geology and climatology [21] and is equivalent to Gaussian process prediction in machine learning. As for extrapolation, we also compare the kriging approach with the representative time-series analysis, which is the vector auto-regression (VAR) model [10]. In our experiment, the kriging approach slightly outperformed VAR [22].…”

“…As one critical application of spatio-temporal collective data analysis, we describe our novel approach for real-time and proactive navigation in such crowded environments as event spaces and urban areas where many people are simultaneously moving toward their destinations [10,11]. The ideal navigation system requires the following characteristics:…”

“…As one critical application of spatio-temporal collective data analysis, we describe our novel approach for real-time and proactive navigation in such crowded environments as event spaces and urban areas where many people are simultaneously moving toward their destinations [10,11]. The ideal navigation system requires the following characteristics: Navigation should be optimized not for individuals, but for overall human navigation so that new congestion will not be caused by independent individual navigations.Navigation should be performed in real time because countermeasures must be facilitated based on the people-flow changes from moment to moment, depending on the people-flow condition.Navigation should be proactive; it should be achieved before congestion occurs by predicting near-future people-flow conditions.Navigation should cope with the avoidance of congestion, while also taking into account the intentions of individuals as much as possible.…”

From Multidimensional Mixture Data Analysis to Spatio-temporal Multidimensional Collective Data Analysis

“…Looking toward 00, we will continue with research and development on the use of spatio-temporal multidimensional collective data analysis techniques and real-time observation data to predict nearfuture events such as congestion in order to implement proactive navigation to relieve congestion at large-scale event venues [4]. We will also investigate how this research can be applied to stabilize the communication infrastructure.…”

MultiCalib

“…However, to our knowledge, there is little work, if any, on calibration on supply/demand models at national scale, which is essential for commercial logistics. Moreover, many applications require fine-grained traffic information, e.g., real-time navigation on road segment levels [19], which calls for model calibration at a spatiotemporal level of road segments and minutes. A national-scale coverage with fine-grained resolutions poses a new challenge for us, which cannot be addressed by existing urban-scale model calibration.…”