Synthetic Sensor Data Generation Using Depth Maps for ADAS 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 that utilize 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 simulating signals from a target sensor, thereby reducing distortion and enhancing accuracy for testing and validation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If real sensor data is captured from multiple positions to test autonomous systems, then testing accuracy is improved, but data capture cost and complexity increase
Solution Approach 1:
The patent creates synthetic sensor data by copying and transforming data from a single real sensor position. Machine learning models generate virtual sensor readings that simulate what multiple sensors would capture, eliminating the need for physically deploying sensors at multiple locations while maintaining testing realism
Solution Approach 2:
The patent introduces machine learning models as intermediaries between the single real sensor and the autonomous system under test. These models act as virtual sensors that translate real sensor data into synthetic readings representing different sensor positions and characteristics, simplifying the data capture process
2Measurement precision
If sensors are positioned differently to capture varied data, then system validation accuracy is improved, but data processing complexity increases
Solution Approach 1:
The patent changes the parameters of existing sensor data through machine learning transformations. By adjusting virtual sensor position, orientation, and physical characteristics in the synthetic data generation process, the system achieves multi-position validation accuracy without actually processing multiple physical sensor streams
Solution Approach 2:
The patent separates the complexity of multi-position data capture into distinct processing stages: (1) capture single real sensor data, (2) apply machine learning transformations to generate synthetic multi-position data, (3) use transformed data for validation. This segmentation reduces overall processing complexity while maintaining validation thoroughness
3Productivity
If synthetic data is generated without accurate depth maps, then data processing speed is improved, but signal accuracy deteriorates
Solution Approach 1:
The patent performs preliminary depth map computation using machine learning models before generating synthetic sensor data. By pre-computing accurate depth information from real sensor inputs, the system ensures high signal accuracy in the synthetic data while maintaining efficient processing speeds during the actual synthetic data generation phase
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.