Object Detection Certainty Index for Autonomous Vehicles
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional object detection techniques fail to accurately recognize the accuracy of detection points, leading to potential erroneous detections and unintended vehicle control.
Innovation Solution
An object detection device that acquires peripheral information, vehicle information, and map information to output traveling status information and certainty indices for each detection point, allowing for accurate recognition of detection point accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional object detection techniques use sensor signals to generate detection points, then detection coverage is achieved, but accuracy of detection points cannot be recognized leading to erroneous detections
Solution Approach 1:
The patent segments the detection result into multiple detection points, each independently evaluated for accuracy. By dividing the overall detection output into discrete points with individual certainty indices, the system can assess and communicate the accuracy of each specific detection point rather than providing a single aggregate result.
Solution Approach 2:
The patent implements feedback by calculating and providing certainty indices for each detection point based on multiple sensor inputs. This feedback mechanism allows the system to communicate back to the control unit how reliable each detection point is, enabling more informed decision-making about vehicle control actions.
2Reliability
If multiple sensors are used to detect objects, then detection coverage and potential accuracy improve, but the system cannot determine the reliability of individual detection points
Solution Approach 1:
The patent merges information from multiple sensors (millimeter-wave radar, ultrasonic sensor, imaging sensor) to evaluate each detection point's reliability. By combining data from these different sensor types and calculating a unified certainty index for each detection point, the system leverages the strengths of multiple sensors without requiring complex individual evaluations of each sensor's contribution.
Solution Approach 2:
The patent changes the parameter representation by introducing certainty indices as a new parameter that quantifies detection point reliability. This transformation converts complex multi-sensor data into a straightforward numerical indicator that directly represents reliability, making it easy to interpret and use for control decisions.
3Productivity
If detection results are used for vehicle control without accuracy information, then responsive control is achieved, but unintended control actions may occur due to erroneous detections
Solution Approach 1:
The patent performs preliminary evaluation of detection point accuracy by calculating certainty indices before vehicle control actions are executed. This preliminary assessment allows the system to identify and potentially filter out unreliable detection points ahead of time, preventing erroneous control actions while maintaining responsiveness to valid detections.
Solution Approach 2:
The patent introduces certainty indices as a new parameter that transforms the control decision-making process. By adding this reliability parameter to the control input, the system can weigh detection points differently based on their accuracy, enabling more nuanced and reliable control responses without sacrificing speed.
Data Source
AI summary
An object detection device includes: a peripheral information acquiring unit to acquire peripheral information including detection information for each of detection points detected by sensing a periphery of a current position; a vehicle information acquiring unit to acquire vehicle information including a current position and a state of a vehicle present at the current position;a map information acquiring unit to acquire map information including a peripheral area of a current position; a traveling status recognizing unit to output traveling status information indicating a traveling state of a vehicle present at a current position and a status around the vehicle by using the peripheral information, the vehicle information, and the map information; and a peripheral information updating unit to output, for each of the detection points, information indicating certainty relating to a detection point by using the traveling status information.