Vehicle Position Learning Control for Sensor-Calibrated Navigation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In hybrid navigation systems, the accuracy of vehicle position coordinates acquired through autonomous navigation is compromised during the initial learning stage or when sensors are recalibrated due to variations in sensor attachment or individual differences, and there is a risk of data loss from volatile storage, leading to incomplete or inaccurate position information.
Innovation Solution
A vehicle information processing system that includes a position acquisition device capable of learning and adjusting sensor output values using teacher data from GPS, and an application execution device that determines the suitability of using this output for various applications based on learning progress and vehicle status, allowing for effective use of sensor data during the learning process.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the output value of the sensor is changed based on learned change value, then the position coordinate accuracy is improved, but the system cannot operate reliably during initial learning stage or when learning result is lost
Solution Approach 1:
The system performs preliminary learning of change values during periods when the vehicle is not executing navigation (such as during charging or parking). This allows the learning to be completed in advance, so that when navigation is needed, the change values are already available and the system can immediately provide accurate position coordinates without being affected by the initial learning stage limitations
Solution Approach 2:
The system introduces a determination unit that acts as an intermediary between the learning execution unit and the application execution unit. This determination unit monitors whether change values are properly established and controls the flow of data accordingly, preventing application execution during periods when learning results are incomplete or lost, thus ensuring system reliability
2Measurement precision
If the system waits for learning completion before executing applications, then position accuracy is improved, but the productivity and response time are reduced
Solution Approach 1:
The system dynamically adjusts its operation mode based on the current learning status. When change values are already established, the system executes applications with high accuracy. When learning is incomplete or results are lost, the system temporarily switches to a mode that allows application execution with standard sensor values, thus maintaining productivity while preserving accuracy when possible
Solution Approach 2:
The system performs learning operations in advance during idle periods (when the vehicle is charging or parked), so that change values are ready before navigation applications need to execute. This preliminary preparation eliminates the need to wait for learning completion during actual navigation, thereby maintaining both accuracy and productivity
3Adaptability or versatility
If learning is performed continuously, then the adaptability to sensor variations is improved, but the device complexity and computational load increase
Solution Approach 1:
Instead of continuous learning, the system performs learning periodically or at specific trigger points (such as when the vehicle is charging, parked, or when significant sensor variations are detected). This periodic approach maintains adaptability to sensor variations while significantly reducing the computational load and system complexity compared to continuous learning
Data Source
AI summary
A position acquisition device includes a learning execution unit that learns a change value of an on-board sensor, a changing unit that changes an output value of the on-board sensor based on the change value, a position calculation unit that calculates a traveling position of the vehicle based on the output value of the on-board sensor changed by the changing unit, and an output unit that outputs a progress status of the learning by the learning execution unit and the traveling position of the vehicle calculated by the position calculation unit. An application execution device determines whether the output of the position acquisition device is able to be used for executing the application based on the progress status and the traveling position acquired from the position acquisition device.