Pedestrian Detection Fusion in Autonomous Emergency Braking
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing autonomous emergency braking systems fail to recognize pedestrians when they are located adjacent to a vehicle due to the radar signal of the vehicle overpowering the pedestrian signal, leading to non-fusion and potential collision recognition issues.
Innovation Solution
An autonomous emergency braking system that includes a camera and a radar sensor, with an electronic control unit generating a pedestrian detection signal when the pedestrian is adjacent to a counterpart vehicle, and fusing this signal with the camera's pedestrian detection signal to recognize the pedestrian, even if the radar signal strength is not higher than the vehicle's signal.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Area of stationary object
If radar sensor is used to detect pedestrian, then detection range is extended, but pedestrian detection fails when pedestrian is adjacent to vehicle due to vehicle signal overpowering
Solution Approach 1:
The patent merges radar sensor detection with camera-based detection to form a fused detection system. When radar detects a pedestrian adjacent to a vehicle, the system combines this with camera image analysis to generate a fused detection result, overcoming the radar's limitation of being overwhelmed by vehicle signals.
Solution Approach 2:
The camera acts as an intermediary device that provides additional verification for pedestrian detection. When radar detects a potential pedestrian, the camera captures images to confirm whether it is indeed a pedestrian, serving as a mediator to resolve the signal interference issue.
2Reliability
If radar signal strength threshold is set high to avoid false detection, then false alarms are reduced, but pedestrian detection sensitivity decreases
Solution Approach 1:
The system combines radar detection results with camera-based detection results through fusion processing. This allows the radar to use higher thresholds for reliability while the camera provides additional detection capability to maintain overall sensitivity, as the fused result considers both sources.
3Device complexity
If fusion processing is performed only on radar signal, then processing complexity is reduced, but pedestrian recognition fails when adjacent to vehicle
Solution Approach 1:
The patent performs fusion processing on both radar detection signals and camera detection signals. This multi-source fusion approach increases processing complexity but significantly improves pedestrian recognition accuracy in challenging scenarios where a pedestrian is adjacent to a vehicle.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
Enables the recognition of pedestrians adjacent to counterpart vehicles by generating a pedestrian detection signal and fusing it with camera data, preventing accidents by accurately identifying potential collision objects.
Implementation Method 1
a radar sensor configured to detect an object located in a preset detection region from the vehicle
Implementation Method 2
a camera installed in the vehicle to acquire an image signal of a proximity of the vehicle
Data Source
AI summary
An autonomous emergency braking system for performing emergency braking when a potential collision object is recognized during traveling of a vehicle, comprises a camera, a radar sensor and an electronic control unit. The camera is installed in the vehicle to acquire an image signal of a proximity of the vehicle. The radar sensor is configured to detect an object located in a preset detection region from the vehicle. The electronic control unit is configured to, when a counterpart vehicle is located adjacent to a pedestrian, generate a pedestrian detection signal that is not included in a detection signal acquired by the radar sensor, and recognize the pedestrian by fusing the generated pedestrian detection signal and a pedestrian detection signal detected by the camera.


