Sensor Confidence Calculation for Indoor Navigation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing position information providing systems, such as those using autonomous navigation, struggle to accurately calculate confidence levels due to reliance on cumulative travel distance and direction change, which do not always reflect the reliability of positioning results, especially in environments where noise is low and sensor data aligns with expected values.
Innovation Solution
A system that includes a sensor data acquiring unit, a feature amount calculating unit, and a confidence level calculating unit, which uses statistical and form-based features of sensor data from multiple sources to accurately determine confidence levels, allowing for more precise guidance information generation and presentation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If autonomous navigation is used to calculate position and azimuth by accumulating sensor measurements from a reference position, then positioning can be achieved in environments where GPS fails (indoors, canyons), but measurement errors accumulate over distance and time leading to reduced accuracy
Solution Approach 1:
The patent implements feedback by continuously monitoring the confidence level of sensor measurements and using this information to adjust the positioning strategy. The system calculates confidence levels based on statistical properties of sensor data and feedback from multiple sensors (acceleration, gyro, magnetic), then uses this feedback to determine whether to trust autonomous navigation results or switch to alternative positioning methods when available.
Solution Approach 2:
The patent changes parameters by introducing confidence level as a dynamic parameter that varies with measurement conditions. Instead of using fixed accumulation methods, the system adjusts the weighting and selection of positioning results based on calculated confidence levels derived from sensor data statistics, allowing adaptive precision control throughout the navigation process.
2Device complexity
If confidence level is calculated using cumulative travel distance and cumulative amount of direction change, then a simple metric is obtained, but this does not accurately reflect positioning reliability when noise is low and sensor data aligns with expected values
Solution Approach 1:
The patent transforms the confidence level calculation from simple cumulative metrics to a multi-parameter statistical analysis. It introduces parameters such as variance, standard deviation, and correlation coefficients derived from multiple sensor measurements, changing the nature of the calculation from purely arithmetic to statistically-based while maintaining computational efficiency through feature amount extraction.
Solution Approach 2:
The patent creates a composite confidence metric by combining information from multiple sources: statistical properties of sensor data, feature amounts extracted from sensor patterns, and correlation analysis between different sensors. This composite approach integrates diverse data types to form a more reliable and accurate confidence level assessment.
Data Source
AI summary
A position information providing apparatus includes a sensor data acquiring unit acquiring sensor data, a feature amount calculating unit calculating a feature amount from the sensor data, and a confidence level calculating unit calculating a confidence level using the feature amount, wherein the sensor data is more than one, and the feature amount includes a statistical amount of more than one piece of the sensor data and an amount showing a form of the sensor data.