Autonomous Vehicle Perception Evaluation for Object Detection Tuning

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Autonomous vehicles face challenges in accurately detecting objects due to variations in sensor types and parameters, leading to potential missed detections or incorrect classifications.

Innovation Solution

An apparatus and method that optimize object detection parameters by comparing object labels applied by the autonomous vehicle with those manually applied by reviewers, using a processor to determine correspondence and adjust parameters as needed.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If object detection parameters are adjusted to increase detection sensitivity, then the likelihood of detecting objects improves, but the likelihood of false detections increases

Engineering Contradiction:

Improveobject detection accuracyVSAvoidfalse detection rate

Core Design Contradiction:

Measurement precisionVSReliability

Solution Approach 1:

The patent combines multiple sensor types (cameras, lasers, radar) to capture data from different modalities. By merging the detection results from these diverse sensors, the system achieves more reliable object detection with reduced false positives, as each sensor type compensates for the weaknesses of others

Inventive Principle:

Principle #5Merging (Combining)

Solution Approach 2:

The system implements feedback mechanisms where detection results are continuously evaluated and used to adjust detection parameters. Manual reviewer feedback on detected objects is incorporated to refine parameter settings, creating a closed-loop system that improves accuracy while maintaining reliability through iterative optimization

Inventive Principle:

Principle #23Feedback

2Measurement precision

If manual review of detected objects is performed to verify accuracy, then detection precision improves, but processing time increases

Engineering Contradiction:

Improveobject label accuracyVSAvoidparameter optimization time

Core Design Contradiction:

Measurement precisionVSLoss of time

Solution Approach 1:

Instead of manually reviewing all detected objects, the system applies partial manual review only to cases where automated detection confidence is below a threshold or where objects are particularly critical. This selective approach maintains high accuracy while minimizing time loss

Inventive Principle:

Principle #16Partial or excessive action

Solution Approach 2:

The system performs preliminary automated detection and filtering before manual review, pre-processing the data to identify only those cases requiring human verification. This preliminary action reduces the volume of work for manual reviewers and accelerates the overall process

Inventive Principle:

Principle #10Preliminary action

Data Source

PatentUS12228928B1System and method for evaluating the perception system of an autonomous vehicle

Publication Date: 2025.02.18 WAYMO LLC

AI summary

A method and apparatus are provided for optimizing one or more object detection parameters used by an autonomous vehicle to detect objects in images. The autonomous vehicle may capture the images using one or more sensors. The autonomous vehicle may then determine object labels and their corresponding object label parameters for the detected objects. The captured images and the object label parameters may be communicated to an object identification server. The object identification server may request that one or more reviewers identify objects in the captured images. The object identification server may then compare the identification of objects by reviewers with the identification of objects by the autonomous vehicle. Depending on the results of the comparison, the object identification server may recommend or perform the optimization of one or more of the object detection parameters.