Sensor Confidence Selection for Accurate Self-Position Estimation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing position estimation devices face accuracy degradation due to varying sensor positions and time stamp variations, especially during constant speed linear motion, which affects self-position estimation accuracy.
Innovation Solution
A position estimation device that estimates state amounts for sensors, calculates confidence degrees, and selects target sensors based on these confidence levels to improve self-position estimation accuracy by using observation values and state amounts from high-confidence sensors.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the position estimation device uses state amount representing sensor state (bias, gravitational acceleration, speed) to estimate self-position, then the estimation can be performed using sensor observation values, but the accuracy of state amount estimation decreases during constant speed linear motion, causing degradation of self-position estimation accuracy
Solution Approach 1:
The system dynamically adjusts the confidence degree calculation based on the trajectory type. During constant speed linear motion, the confidence degree for acceleration sensor-based state amounts is reduced, while during non-linear motion, the confidence degree is increased. This dynamic adjustment resolves the contradiction by adapting the reliability assessment to the actual motion conditions.
Solution Approach 2:
The system changes the parameter of confidence degree based on motion trajectory characteristics. By detecting whether the motion is constant speed linear motion or other trajectories, the system adjusts the confidence degree parameter for different state amounts, thereby maintaining accurate self-position estimation across different motion scenarios.
2Adaptability or versatility
If the position estimation device uses state amount with varying accuracy depending on trajectory, then the system can handle different motion patterns, but the overall estimation accuracy becomes unstable due to low confidence during certain trajectories
Solution Approach 1:
The system employs dynamic confidence degree adjustment that responds to trajectory changes. The confidence degree is not fixed but varies dynamically based on the detected motion pattern, ensuring stable estimation accuracy across different trajectories while maintaining adaptability to new motion patterns.
Solution Approach 2:
The system uses feedback from trajectory detection to adjust confidence degrees. By continuously monitoring the motion trajectory and feeding this information back to the confidence degree calculation, the system maintains stable estimation accuracy adaptively across different motion patterns.
Data Source
AI summary
A position estimation device according to an embodiment of the present disclosure is provided for estimating a self-position of a moving object device provided with one or more sensors for observing information related to movement. The position estimation device includes one or more hardware processors configured to: estimate a state amount representing a state of each of the one or more sensors based on an observation value of the corresponding sensor; calculate a confidence degree representing a degree of confidence of the state amount of each of the one or more sensors; select one or more target sensors from among the one or more sensors based on the confidence degree of the state amount of each of the one or more sensors; and estimate the self-position based on the observation value and the state amount of each of the selected one or more target sensors.