Vehicle Optical Sensor Motion Correction via Computational Adjustment
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Autonomous and semi-autonomous vehicles face challenges in maintaining accurate optical sensor data due to motion-related misalignments and dynamic changes in sensor positioning, which existing methods address through increased stiffness, adding complexity and weight, or requiring additional sensors.
Innovation Solution
A computer-based adjustment model predicts and corrects for optical sensor motion relative to the vehicle using physics-based models and motion data from existing sensors, allowing for real-time adjustment and modification of the model without additional hardware, thereby maintaining data accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If stiffness is increased to maintain sensor alignment, then measurement precision is improved, but device complexity and weight increase
Solution Approach 1:
The patent replaces the mechanical stiffness-based alignment maintenance system with a computational system. Instead of using rigid mounting structures to prevent sensor motion, the invention uses software algorithms to detect and correct alignment deviations in real-time, substituting mechanical constraints with electronic detection and computational correction.
Solution Approach 2:
The patent introduces intermediate reference features (such as visual markers or fiducial points) that mediate between the sensor and the vehicle body. These intermediaries provide measurable reference points that allow the system to detect and correct alignment changes without requiring the sensor to be rigidly fixed to the vehicle.
2Measurement precision
If additional sensors are added to monitor sensor motion, then measurement precision is improved, but device complexity and cost increase
Solution Approach 1:
The patent makes existing sensors perform multiple functions. The optical sensor itself is used both for its primary detection purpose and for monitoring its own alignment by analyzing changes in its field of view or by using reference features within its scene. This eliminates the need for dedicated alignment monitoring sensors.
Solution Approach 2:
The system uses the optical sensor's own data and existing vehicle sensors to monitor and correct its alignment. The sensor essentially monitors itself by detecting changes in reference features in its field of view, and the vehicle's existing motion sensors provide data that helps characterize the sensor's motion, eliminating the need for external monitoring systems.
3Reliability
If real-time adjustment is implemented, then reliability is improved, but computational requirements and processing time increase
Solution Approach 1:
The patent pre-characterizes the relationship between vehicle motion and sensor alignment by establishing lookup tables or pre-computed correction models during system calibration. When the vehicle operates, the system simply queries these pre-computed models based on current motion sensor readings, avoiding the need for complex real-time calculations and reducing processing time.
Solution Approach 2:
The patent implements a dynamic correction system that continuously adapts to changing conditions. The alignment correction is updated in real-time based on current vehicle motion, sensor data, and environmental conditions, allowing the system to maintain high reliability while using efficient algorithms that adapt to the current operating state rather than using fixed, computationally intensive methods.
Data Source
AI summary
A computer includes a processor and a memory storing instructions executable by the processor to receive optical data from an optical sensor of a vehicle, adjust the optical data using an adjustment model, measure a first value from the optical data over a duration, predict a second value from the optical data over the duration based on the first value, measure the second value from the optical data over the duration, and modify the adjustment model using the predicted second value and the measured second value. The first value and the second value are time-varying and aggregated from the optical data per time step.