Synthetic Sensor Data Generation for Autonomous Driver Testing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for testing autonomous systems, such as advanced driver-assistance systems (ADAS) and autonomous driving systems, face challenges in accurately simulating sensor data due to variations in sensor position and physical characteristics, leading to high costs and inefficiencies in data capture and processing.
Innovation Solution
A system and method utilizing machine learning models to compute depth maps from real signals captured by multiple sensors, applying point of view and physical characteristic transformations to create synthetic data that simulates signals from a target sensor, reducing distortion and enhancing accuracy for testing autonomous systems.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If real sensor data is captured from multiple positions and configurations to test autonomous systems, then testing accuracy and reliability improve, but cost and time consumption increase significantly
Solution Approach 1:
The patent creates synthetic sensor data by copying and transforming data from a limited number of real sensor captures. Virtual sensor data is generated through mathematical transformations that simulate what sensors would capture at different positions, orientations, and environmental conditions, eliminating the need to physically recapture data for every test scenario
Solution Approach 2:
The system transforms sensor data by changing parameters such as sensor position, orientation, environmental conditions (weather, lighting), and sensor characteristics. These parameter transformations allow one set of real captures to generate multiple virtual datasets with different conditions, maintaining testing reliability without additional physical captures
2Adaptability or versatility
If multiple sensors are deployed at different positions on the vehicle to capture comprehensive data, then sensor position variations are accounted for, but device complexity and manufacturing cost increase
Solution Approach 1:
Instead of physically installing multiple sensors at different vehicle positions, the system creates virtual copies of sensor data through mathematical transformations. A single physical sensor capture is transformed to simulate readings from multiple sensor positions, reducing hardware complexity while maintaining adaptability to different sensor configurations
Solution Approach 2:
The synthetic data generation system serves multiple functions: it simulates different sensor positions, orientations, environmental conditions, and sensor types from a single data capture. This multi-functional approach replaces the need for multiple specialized sensors with a universal data transformation framework
3Quantity of substance
If extensive real-world data capture is performed to cover all possible driving conditions, then data comprehensiveness improves, but cost and resource requirements increase
Solution Approach 1:
The system generates large volumes of synthetic training data by copying and transforming limited real sensor captures. Virtual data is created for numerous driving scenarios, weather conditions, and environmental variations without requiring physical deployment in each scenario, dramatically improving data capture efficiency while maintaining data comprehensiveness
Data Source
AI summary
A system for creating synthetic data for testing an autonomous system, comprising at least one hardware processor adapted to execute a code for: using a machine learning model to compute a plurality of depth maps based on a plurality of real signals captured simultaneously from a common physical scene, each of the plurality of real signals are captured by one of a plurality of sensors, each of the plurality of computed depth maps qualifies one of the plurality of real signals; applying a point of view transformation to the plurality of real signals and the plurality of depth maps, to produce synthetic data simulating a possible signal captured from the common physical scene by a target sensor in an identified position relative to the plurality of sensors; and providing the synthetic data to at least one testing engine to test an autonomous system comprising the target sensor.