Tire Pressure Data Cleaning With ML Error Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing tire pressure monitoring systems (TPMS) often generate erroneous tire pressure values that are difficult to distinguish from valid values, leading to inaccurate tire pressure data, which can impact safety, fuel efficiency, and tire wear.
Innovation Solution
A method involving a trained machine learning classifier, such as a tree-based classifier, is used to detect and remove erroneous tire pressure values by analyzing tire pressure data from a fleet of vehicles, utilizing histogram representations and logarithmic comparisons to identify patterns, and then simplifying the data to reduce transmission and processing load.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If TPMS continuously monitors tire pressure using physical sensors or signal processing, then tire pressure detection capability is improved, but erroneous values are generated that reduce measurement precision
Solution Approach 1:
A machine learning classifier is introduced as an intermediary component between the TPMS sensor and the fleet management system. This classifier processes the raw tire pressure data, identifying and filtering out erroneous values before transmission. The mediator resolves the contradiction by maintaining the continuous monitoring capability while eliminating unreliable data points through intelligent classification.
Solution Approach 2:
The system implements a feedback mechanism where the machine learning classifier continuously evaluates tire pressure readings against learned patterns of erroneous values. By comparing current readings with historical data and identifying anomalies, the system provides feedback to filter out incorrect measurements, thereby improving overall data reliability without sacrificing monitoring continuity.
2Loss of information
If all tire pressure values are transmitted to the fleet management system, then data completeness is improved, but bandwidth and processing requirements increase
Solution Approach 1:
The machine learning classifier extracts and removes erroneous tire pressure values from the data stream before transmission to the fleet management system. By taking out only the problematic data points and transmitting only valid readings, the system maintains data completeness for legitimate measurements while significantly reducing the volume of data that consumes bandwidth and processing resources.
Data Source
AI summary
Systems and methods for detecting and removing erroneous tire pressure values are provided. The method involves operating at least one processor to: receive tire pressure data comprising a time series of tire pressure values; simplify the tire pressure data by removing at least some of the tire pressure values that satisfy a predetermined interpolation error threshold; detect, using a trained machine learning classifier, at least one erroneous tire pressure value in the tire pressure data, each erroneous tire pressure value being detected based on the erroneous tire pressure value and a plurality of lagging tire pressure values consecutively trailing the erroneous tire pressure value; clean the tire pressure data by removing the at least one erroneous tire pressure value; and transmit the tire pressure data to a fleet management system, whereby the at least one erroneous tire pressure value is not transmitted to the fleet management system.