Spatiotemporal Data Indexing Using Spatial and Time Tree Structures
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for processing and querying large volumes of spatiotemporal data, such as RF spectrum measurements, are inefficient, requiring extensive pre-processing and indexing, which leads to delays and high computational costs when retrieving and analyzing data across specific regions and time periods.
Innovation Solution
A method involving a spatial tree index and a time tree index, where each spatial tree node points to a time tree index, allowing for efficient retrieval of data by using pointers to access relevant records, and the generation of aggregated data records to summarize large datasets without needing to retrieve all raw data, using a quad tree structure and data pointers to improve search and analysis efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional pre-processing and indexing methods are used to process large volumes of spatiotemporal data, then data can be organized and stored, but the computational expense and time required to retrieve and analyze data increases significantly
Solution Approach 1:
The patent divides the spatiotemporal data into spatial segments using a quadtree structure, where each node represents a specific spatial region. This segmentation allows the system to quickly identify and access only the relevant spatial segments containing data within the queried region, avoiding the need to process or retrieve data from entire time series for all spatial locations. The temporal data within each spatial node is further organized chronologically, enabling efficient retrieval of both spatial and temporal subsets simultaneously.
2Loss of information
If all raw data records are retrieved and processed to generate summary statistics, then comprehensive analysis can be performed, but the computational cost and processing time increases
Solution Approach 1:
The patent extracts and stores summary statistics (such as mean, maximum, minimum, and standard deviation) directly within each temporal node of the quadtree structure. When a query is executed, the system retrieves only the pre-computed summary statistics for the relevant spatial and temporal subsets, rather than retrieving and processing all raw data records. This extraction approach maintains analysis completeness by providing sufficient statistical information while dramatically reducing computational expense and data transfer requirements.
3Productivity
If detailed indexing structures are created to enable efficient data retrieval, then query performance improves, but the complexity of the database structure increases
Solution Approach 1:
The patent merges the spatial indexing structure (quadtree) with temporal data organization into a unified hierarchical structure. Each node in the quadtree simultaneously represents a spatial region and contains temporal data or references to temporal data. This merging eliminates the need for separate spatial and temporal index structures, reducing overall system complexity while maintaining efficient query performance. The unified structure allows single-pass traversal to retrieve both spatial and temporal subsets, improving productivity without proportionally increasing device complexity.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A method of processing a primary data record (7), a spatial tree index (17) and a plurality of time tree indexes (18) is described. Each primary data record comprises location data (8), time data (9) and measurement data (10). The spatial tree index comprises spatial tree nodes (Sn) linked by pointers (27'), each spatial tree node corresponding to a given spatial region (Rn(k,j)) and further including a pointer to a respective time tree index. Each time tree index corresponds to a respective spatial tree node, and comprises time tree nodes (Tm) linked by pointers. Each time tree node corresponds to a given time period (Pm(p)). The method comprises writing the primary data record to a primary data table (11), determining whether a spatial tree leaf node (SN) exists which corresponds to the spatial region that includes the location data and, on a negative determination, generating a spatial tree leaf node which corresponds to the spatial region that includes the location data, and creating or updating spatial tree nodes connecting the spatial tree leaf node to the spatial tree root node (S1). For the spatial tree nodes which correspond to each region that includes the location data, the respective time tree index is updated by determining whether a time tree leaf node (TM) exists which corresponds to the time period that includes the time data and, on a negative determination, generating the time tree leaf node which corresponds to the time period that includes the time data, and creating or updating time tree nodes connecting the time tree leaf node to the time tree root node (T1). The time tree leaf node is updated to include a pointer (31) to the primary data record.