Autonomous Vehicle Sensor Calibration in Undefined Environments
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing sensor calibration methods for autonomous vehicles require dedicated infrastructure, such as automated turntables, which are costly and limited in flexibility, making it difficult to adapt to real-world environments with undefined features.
Innovation Solution
Calibrate multiple sensors of an autonomous vehicle in an undefined training area, utilizing various movements like loops and open road navigation to refine sensor alignment and confidence, without relying on predefined maps or infrastructure.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If dedicated infrastructure like automated turntables is used for sensor calibration, then calibration precision can be improved, but device complexity and cost increase significantly
Solution Approach 1:
The autonomous vehicle performs self-calibration using its own sensors and processing systems without external calibration infrastructure. The vehicle autonomously identifies environmental features, collects sensor data, and computes calibration parameters through self-contained algorithms, eliminating the need for automated turntables or other dedicated calibration equipment.
Solution Approach 2:
Environmental features in the undefined training area serve as intermediaries for calibration. Instead of using artificial calibration targets on turntables, the system uses naturally occurring features like建筑物, trees, and road markings as reference points to establish sensor alignment and spatial relationships.
2Measurement precision
If dedicated calibration infrastructure is used, then calibration accuracy is improved, but adaptability to real-world environments deteriorates
Solution Approach 1:
The calibration system is designed to operate universally across diverse real-world environments without requiring specialized infrastructure. The same sensor suite and processing algorithms used for navigation are employed for calibration, allowing the vehicle to calibrate itself in parking lots, open roads, or any area with sufficient environmental features.
Solution Approach 2:
Instead of bringing the vehicle to a controlled calibration facility, the approach inverts the process by taking the calibration capability to the vehicle and enabling it to perform calibration in-situ wherever it operates. The vehicle actively seeks out and uses environmental features rather than passively being positioned on a turntable.
3Adaptability or versatility
If multiple sensors are calibrated in an undefined environment without predefined maps, then adaptability improves, but measurement precision deteriorates
Solution Approach 1:
The system performs preliminary actions to establish reliable measurements in undefined environments. It first identifies and catalogs environmental features, then uses these features as reference points for subsequent calibration calculations. This preliminary feature detection and validation ensures that even without predefined maps, the calibration can achieve sufficient precision.
Solution Approach 2:
The calibration process incorporates feedback loops where the system continuously refines its understanding of environmental features and sensor relationships. By iteratively comparing sensor measurements against identified features and adjusting calibration parameters accordingly, the system improves measurement precision through feedback-driven optimization.
Data Source
AI summary
The subject technology is related to autonomous vehicles (AV) and, in particular, to calibrating multiple sensors of an AV in an undefined training area. An example method includes instructing the AV to pilot itself in an undefined training area subject to at least one constraint, wherein the AV is instructed to pilot itself along a path until the AV has overlapped at least a portion of the path, and at a location at which the AV has overlapped at least the portion of the path, determining that first returns from a previous LIDAR scan overlaps with second returns from a subsequent LIDAR scan taken when the AV has overlapped at least the portion of the path. Initially, the AV does not include a location reference to identify locations of objects in the undefined training area and a plurality of sensors of the AV are uncalibrated with respect to each other.