Adaptive Object Detection Device for Vehicle State Prediction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing object detection devices in vehicles cannot adjust detection accuracy based on the vehicle's behavior and predicted route, leading to inadequate information for effective traveling assistance.
Innovation Solution
An object detection device with a host-vehicle state prediction unit, detection unit, and detection characteristic control unit that adjusts detection characteristics such as detection rate and error rate based on the vehicle's position, route, and potential collision scenarios, using data from sensors like cameras and radar to enhance object detection accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If detection accuracy is adjusted based on image pattern and vehicle position, then object acquisition accuracy is improved, but detection coverage and completeness deteriorate
Solution Approach 1:
The detection characteristic control unit dynamically adjusts detection characteristics based on the predicted target state of the host vehicle. The system switches between different detection modes (first detection characteristic for high accuracy, second detection characteristic for broad coverage) according to the vehicle's future position and route predictions, making the detection system adaptive rather than static.
Solution Approach 2:
The system changes detection parameters such as detection range, detection sensitivity, and detection frequency based on the predicted target state. When the vehicle is predicted to turn left, the system increases detection sensitivity in the left direction; when traveling straight, it maintains balanced detection. This parameter adaptation resolves the contradiction by optimizing detection characteristics for each specific scenario.
2Measurement precision
If detection characteristics are optimized for specific positions, then detection accuracy is improved, but adaptability to different vehicle behaviors deteriorates
Solution Approach 1:
The host vehicle state prediction unit predicts multiple possible target states (different positions and routes) based on current vehicle behavior. The detection characteristic control unit then selects appropriate detection characteristics for each predicted state, enabling the system to adapt to turns, straight-line travel, and other maneuvers dynamically rather than being fixed for specific positions only.
Solution Approach 2:
The system performs preliminary prediction of the host vehicle's target state before actual detection optimization. By predicting where the vehicle will be and what route it will take, the system can pre-adjust detection characteristics to match the upcoming scenario, ensuring both accuracy and adaptability to different vehicle behaviors.
3Reliability
If detection rate is increased to prevent collisions, then safety is improved, but erroneous detection rate increases
Solution Approach 1:
The system applies different detection characteristics to different spatial directions based on the predicted target state. When the vehicle is predicted to turn left, high detection sensitivity is applied specifically to the left direction rather than uniformly in all directions. This localized optimization maintains high detection rate where needed while reducing erroneous detections in other areas.
Solution Approach 2:
The detection characteristic control unit adjusts detection parameters dynamically based on the predicted target state. For directions where collision risk is high (based on predicted turns or maneuvers), the system increases detection sensitivity and rate. For directions with low risk, it maintains lower sensitivity, thus preventing both collisions and erroneous detections through context-aware parameter adaptation.
Data Source
AI summary
An object detection device can acquire information of an object in the vicinity of a host-vehicle for appropriate traveling assistance. An object detection device 1 includes a vehicle state detection section 2, an environmental situation acquisition section 3, a road information acquisition section 4, a detection control section 6, and a detection section 7. A host-vehicle state prediction section 61 acquires a target state of a host-vehicle 81. The detection section 7 detects an object. A parameter setting section 63 switches the detection characteristic of the object in the detection section 7 in accordance with the target state.