Autonomous Vehicle Component Validation Using Adverse Event Simulation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Simulations for testing autonomous vehicle components often require thousands or millions of scenarios to achieve a reasonable confidence interval, making them computationally demanding and time-consuming, which can decrease safety and prevent other components from being tested effectively.
Innovation Solution
Modifying the probability distribution of scenario parameters to increase the likelihood of adverse events in simulations, allowing for a higher confidence interval without increasing the number of scenarios, and using a correction factor to ensure accurate metric determination.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If thousands or millions of simulations are run to achieve a reasonable confidence interval, then measurement precision is improved, but loss of time and use of energy increase significantly
Solution Approach 1:
The patent modifies the probability distribution parameters of scenario selection to overweight adverse event scenarios. By changing the parameter weights in the simulation scenario selection, the system achieves higher measurement precision for safety-critical components without requiring proportionally more simulation time, as the modified distribution concentrates computational effort on the most relevant test cases.
Solution Approach 2:
The patent applies partial action by selectively simulating only the most critical adverse event scenarios rather than exhaustively simulating all possible scenarios. The correction factor compensates for the non-uniform sampling, allowing the system to achieve adequate measurement precision through a subset of carefully selected scenarios rather than requiring complete scenario coverage.
2Measurement precision
If thousands or millions of simulations are run to achieve a reasonable confidence interval, then measurement precision is improved, but use of energy increases significantly
Solution Approach 1:
The patent modifies the probability distribution parameters to concentrate simulation resources on adverse event scenarios. This parameter change in the scenario selection process reduces the total number of simulations needed to achieve a given confidence interval, thereby reducing computational energy consumption while maintaining measurement precision for safety-critical assessments.
Solution Approach 2:
The patent uses partial action by focusing computational energy on simulating only the most critical adverse event scenarios rather than uniformly sampling all scenarios. The correction factor enables accurate metric determination from this focused subset, reducing overall energy consumption while achieving the required confidence interval.
3Productivity
If the probability distribution is modified to increase adverse event likelihood, then productivity is improved, but measurement precision may be compromised without correction
Solution Approach 1:
The patent implements feedback by calculating a correction factor based on the modified probability distribution. This correction factor is derived from the ratio of the original to modified probability distributions and is applied to the simulation results to compensate for the biased sampling. The feedback mechanism ensures that productivity gains from modified scenario selection do not compromise measurement precision.
Solution Approach 2:
The patent applies preliminary anti-action by pre-calculating the correction factor that counteracts the bias introduced by modifying the probability distribution. Before final metric determination, the system applies this correction to neutralize the effect of overweighting adverse events in scenario selection, thereby maintaining measurement precision while achieving faster validation through modified distribution sampling.
Data Source
AI summary
Validating a component of an autonomous vehicle may comprise determining, via simulation, a likelihood that operation of the component will result in an adverse event. Such simulations may be based on log data developed from real world driving events to, for example, accurately model a likelihood that a scenario will occur during real-world driving. Because adverse events may be exceedingly rare, the techniques may include modifying a probability distribution associated the likelihood that a scenario is simulated, determining a metric associated with an adverse event (e.g., a likelihood that operating the vehicle or updating a component thereof will result in an adverse event), and applying a correction to the metric based on the modification to the probability distribution.