Virtual Object Simulation for Autonomous Vehicle Scenario Testing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Testing autonomous vehicles in realistic scenarios is challenging due to safety concerns and high costs associated with using real people or dummies, and existing simulations often lack the unpredictability of real-world interactions.
Innovation Solution
The method involves using virtual objects and fictitious sensor data to simulate real-world scenarios, allowing autonomous vehicles to maneuver and respond as if interacting with real objects, while logging responses for evaluation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If real people or dummies are used to test autonomous vehicle reactions, then the testing scenarios become more realistic, but safety risks and costs increase significantly
Solution Approach 1:
The patent uses virtual objects that are digital copies or representations of real-world objects (pedestrians, vehicles, obstacles) to create realistic testing scenarios without the physical presence of actual objects. These virtual objects are rendered in the sensor data stream to simulate real object detections, providing reliable and realistic test conditions while eliminating safety risks associated with using real people or dummies.
2Reliability
If real people or dummies are used to test autonomous vehicle reactions, then the testing scenarios become more realistic, but costs and time consumption increase
Solution Approach 1:
Virtual objects are digitally generated and can be instantly deployed in the testing environment without physical setup. This allows multiple realistic scenarios to be tested rapidly by simply changing the virtual object parameters and positions in the sensor data stream, eliminating the time-consuming setup required for physical dummies or staged real-world scenarios.
Solution Approach 2:
The system allows dynamic modification of virtual object parameters (position, speed, type, behavior patterns) within the sensor data stream. This enables rapid scenario changes and repeated testing with different conditions without physical reconfiguration, significantly reducing setup time and allowing comprehensive testing of various edge cases.
3Object-affected harmful factors
If virtual objects are used to simulate real objects, then safety and cost improve, but the unpredictability of real-world interactions may be reduced
Solution Approach 1:
The system introduces controlled randomness and variability in virtual object parameters such as position, velocity, acceleration, and behavior patterns. This allows the simulation to replicate the unpredictability of real-world interactions while maintaining safety, as the virtual objects can be configured to exhibit stochastic behaviors that mimic real-world scenarios.
Solution Approach 2:
Virtual objects are designed with dynamic properties that allow them to change state and behavior during simulation, including unpredictable movements, sudden appearances, and varying response patterns. This dynamic nature ensures that the testing scenarios remain diverse and representative of real-world conditions while using safe virtual representations.
Data Source
AI summary
An autonomous vehicle is tested using virtual objects. The autonomous vehicle is maneuvered, by one or more computing devices, the autonomous vehicle in an autonomous driving mode. Sensor data is received corresponding to objects in the autonomous vehicle's environment, and virtual object data is received corresponding to a virtual object in the autonomous vehicle's environment. The virtual object represents a real object that is not in the vehicle's environment. The autonomous vehicle is maneuvered based on both the sensor data and the virtual object data. Information about the maneuvering of the vehicle based on both the sensor data and the virtual object data may be logged and analyzed.