Speed Correction Model Using Aggregated Telemetry Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Accurate speed estimation in electronic maps is challenging due to biases in telemetry data from varying road conditions, traffic, and time, leading to inaccurate arrival time predictions.
Innovation Solution
A speed correction model is trained on aggregated telemetry data to correct estimated speeds by using a server computer that receives data from multiple devices, anonymizing and segmenting it to eliminate individual device information, and applying a similarity score-based correction mapping for geographic regions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If speed estimation is performed using telemetry data from mobile devices, then speed data can be obtained for electronic map applications, but the speed estimates contain biases due to varying road conditions, traffic, and time
Solution Approach 1:
The patent transforms speed estimation from a direct calculation to a correction process by changing the parameters used. Instead of relying solely on raw telemetry data, the system uses correction mappings that adjust estimated speeds based on geographic region, time of day, and day of week parameters. This parameter transformation resolves the contradiction by accounting for varying conditions that cause estimation biases.
Solution Approach 2:
The system implements feedback by using actual observed speeds from telemetry data to train correction models. These correction models then feed back into the speed estimation process, continuously improving accuracy. The feedback loop addresses the reliability issue by learning from historical data to compensate for systematic biases in speed estimates.
2Adaptability or versatility
If correction models are trained on aggregated telemetry data from multiple geographic regions, then the model can generalize across different areas, but regional differences in driving patterns and conditions may reduce correction accuracy
Solution Approach 1:
The patent segments the geographic space into discrete regions (e.g., hexagons or grid cells) and trains separate correction mappings for each region. This segmentation allows the system to capture regional differences in driving patterns while maintaining adaptability across multiple areas. Each region's correction model is trained on locally-relevant telemetry data, preserving measurement precision while achieving broad geographic coverage.
Solution Approach 2:
The system applies local quality by creating region-specific correction mappings that are tailored to local driving conditions, road types, and traffic patterns. Instead of using a single global correction model, each geographic region receives a customized correction mapping trained on local telemetry data. This ensures high accuracy for each local area while maintaining overall system versatility through the collection of regional models.
3Quantity of substance
If telemetry data is collected and processed to determine device speeds, then speed information can be derived from GPS and WiFi data, but the large amounts of data make accurate speed estimation challenging
Solution Approach 1:
The patent extracts only the essential speed information from the large volume of telemetry data by using correction mappings that operate on aggregated statistics rather than individual data points. The system extracts key patterns from the data (such as average speeds by region and time) and uses these extracted features for correction, rather than processing every raw telemetry record. This reduces computational complexity while maintaining estimation accuracy.
Solution Approach 2:
The system merges multiple data sources (GPS location, WiFi data, device speed sensors) and multiple telemetry records into aggregated speed statistics. By combining information across many data points and sources, the system derives more reliable speed estimates. The correction mappings then merge regional, temporal, and contextual factors to produce accurate corrected speeds from the aggregated information.
Data Source
AI summary
A method for correcting speed estimates for route planning using a machine-learned speed correction model trained on aggregated road data. Location and movement data collected from a plurality of mobile computing devices is aggregated on a server computer and used to train a speed correction model to correct estimated speeds corresponding to roads in one or more geographic regions. Speeds estimates for a road segment in a geographic region are corrected using a speed correction model trained on road data describing road segments in the same geographic region. In some embodiments, road data corresponding to one or more geographic regions is assigned to groups in training the speed correction model. The road data may be anonymized or segmented such that an originating device or route is unidentifiable. More fine-grained speed correction models may also be trained for different or additional factors than geographic region, such as day and/or time.