Driving Action Evaluation Logic Selection for Computational Load Reduction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing driving action evaluation systems face high computational loads when processing various driving scenes, particularly in vehicles with advanced drive assist or automatic driving systems, as they need to execute evaluations in parallel for multiple scenarios, leading to inefficiencies.
Innovation Solution
A driving action evaluating device that specifies the driving scene using onboard camera images and position information, then derives evaluations by executing evaluation logic corresponding to that scene among a set of predefined logics, reducing the computational load by only using relevant evaluation logic for the specified scene.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If evaluation logics for multiple driving scenes are executed in parallel, then comprehensive driving action evaluation is achieved, but computational load increases
Solution Approach 1:
The patent segments the evaluation process by dividing multiple driving scenes into distinct categories, each with its own dedicated evaluation logic. The specifying section identifies which driving scene is currently active, and only the corresponding evaluation logic is executed. This segmentation allows comprehensive evaluation coverage while avoiding the computational overhead of running all evaluation logics simultaneously.
Solution Approach 2:
Instead of executing all evaluation logics for all possible driving scenes, the system performs partial action by selecting and executing only the evaluation logic relevant to the currently identified driving scene. This reduces computational load significantly while still achieving accurate evaluation for the active scene, applying the principle of doing just enough work needed for the current context.
2Adaptability or versatility
If multiple evaluation logics are executed simultaneously, then all driving scenes are covered, but processing efficiency decreases
Solution Approach 1:
The system dynamically adapts its processing behavior based on the identified driving scene. The specifying section continuously monitors driving conditions and dynamically selects which evaluation logic to execute next. This dynamic approach ensures that the system maintains high processing efficiency by avoiding static execution of all evaluation logics, while still adapting to cover all necessary driving scenes as they occur.
3Measurement precision
If comprehensive evaluation for all driving scenes is performed, then evaluation accuracy is maintained, but computational resources are overutilized
Solution Approach 1:
The patent extracts and isolates only the necessary evaluation logic corresponding to the currently active driving scene, separating it from other evaluation logics that are not needed at this moment. This extraction process maintains evaluation accuracy for the active scene while eliminating unnecessary computational resource consumption associated with evaluating other unrelated driving scenes, directly addressing the contradiction between accuracy and resource utilization.
Data Source
AI summary
A driving action evaluating device that specifies a driving scene by using at least one of an image captured by an onboard camera and position information of an own vehicle; and derives an evaluation of driving actions by executing an evaluation logic, which corresponds to only a driving scene specified by the specifying section, among a plurality of evaluation logics that correspond to a plurality of driving scenes, respectively.