A parking recognition method based on spatio-temporal clustering

Abstract

The invention discloses a parking identification method based on space-time clustering, comprising the following steps: collecting GPS track data of individual activities based on a smart phone, and extracting space-time information; searching for the nearest k points of any track point on the time axis, The core points in the trajectory are determined by the distance parameter Eps and the minimum number threshold MinPts; the core points that are continuous in time form the initial cluster, start from the cluster with the highest density, and merge the adjacent clusters adjacent in space and time to obtain the parking; time; The continuous non-core points above constitute the initial trip, and the inspection starts from the trip with the earliest time. If the time interval with the next trip is less than the minimum parking duration threshold, the two are merged, and the false parking is corrected as a trip. The invention can quickly and accurately identify the stops in the GPS track of individual travel, lays a foundation for further identification of travel mode and travel purpose, and provides technical support for long-term, large-scale, passive travel surveys of urban residents.

Description

technical field [0001] The invention belongs to the field of traffic data mining, and in particular relates to cluster analysis of time series data and identification of stops in individual travel trajectories. Background technique [0002] With the rapid popularization and development of smart phones, precise positioning functions and rich sensor modules provide hardware conditions for real-time collection of individual travel trajectories. In the face of a large amount of travel trajectory data, analyzing individual behavior characteristics and identifying activity patterns has become a major problem in data services for urban traffic. Recognition of parking based on individual GPS trajectory data is a prerequisite for judging OD, estimating travel mode and travel purpose. The current related research is mainly based on the static speed characteristics, the direction characteristics of the movement, and combined with the road network to make regular judgments. However, b...

AI Technical Summary

Problems solved by technology

However, the time sequence of points is not considered, and it is easy to mistakenly classify point clusters with similar spatial distances and long time intervals into one category, and it cannot accurately identify stops for overlapping paths, short-term travel, etc.

Specifically, its algorithm has the following disadvantages when processing GPS mobile data of smartphones: when processing a large number of trajectory points of the day, the distance matrix memory occupies a large amount, causing the program to fail to respond or the operation speed to be slow; the parameters Eps and MinPts are highly sensitive, Poor generalization ability; it is difficult to identify those with a large number of trips and a high degree of path overlap, especially for some road sections or intersections with overlapping trajectories due to the accumulation of a large number of trajectory points, so it is misjudged as a stop

However, for complex travel or poor signal quality, the merging threshold may not be suitable; the merging order is random, which affects the recognition results; short-term false stops that are not eliminated for wrong recognition, the result of the number of trips will be high

Images