Probabilistic Drivable Area Detection Using Camera-Radar Fusion
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional vehicle drivable area detection methods are not accurate enough, leading to incomplete recognition of drivable areas and potential collisions with obstacles during autonomous driving, particularly in scenarios with various obstacles.
Innovation Solution
Integrate a camera apparatus and radar to fuse obstacle distribution information, using a neural network to process image data and radar echo signals, representing the drivable area as a probability distribution to enhance recognition accuracy and flexibility in navigation planning.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a single sensor (camera or radar) is used for obstacle detection, then the device complexity is low, but the measurement precision and detection coverage are insufficient due to blind spots and limited ranges
Solution Approach 1:
The patent combines camera and radar sensors into an integrated sensor system that fuses obstacle detection data from both sources. The camera provides visual information for识别 obstacle types and characteristics, while the radar provides distance and velocity data, compensating for each other's limitations and reducing blind spots through multi-sensor fusion.
Solution Approach 2:
The integrated sensor system performs multiple functions simultaneously: the camera captures visual data for obstacle classification and the radar measures distance and speed, creating a universal detection system that handles various obstacle types (pedestrians, vehicles, cyclists) with a single unified architecture.
2Reliability
If conventional single-sensor detection is used, then the system is simple to operate, but the drivable area recognition is not accurate enough leading to potential collisions
Solution Approach 1:
The patent merges camera and radar detection results through data fusion algorithms that combine obstacle position, velocity, and visual classification data. This integration improves reliability by cross-validating detections from both sensors, reducing false positives, and providing redundant detection coverage for critical safety applications.
Solution Approach 2:
The patent introduces a data fusion module as an intermediary that processes and integrates data from camera and radar sensors. This mediator combines the complementary information from both sensors, reconciles their different data formats and coordinate systems, and produces a unified obstacle detection result that is more reliable than either sensor alone.
3Adaptability or versatility
If deterministic drivable area classification is used (drivable or non-drivable), then the navigation planning is simple, but the system lacks flexibility for conditional driving scenarios
Solution Approach 1:
The patent changes the output parameter from deterministic classification (drivable/non-drivable) to probabilistic representation (drivable probability distribution). This allows the system to express uncertainty and conditional drivability, enabling navigation planning to consider factors like obstacle type, velocity, and predicted trajectory, thereby improving adaptability for complex scenarios such as slow-moving obstacles or uncertain pedestrian intentions.
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
The integrated sensor system provides comprehensive obstacle detection, avoiding blind spots and improving the accuracy of drivable area recognition, enabling more flexible and reliable autonomous driving.
Implementation Method 1
obtaining a second probability distribution of the obstacle based on a time of flight and an echo width of a radar echo signal, wherein the echo width is the difference between a second time of flight of the echo signal and a first time of flight of the echo signal, wherein the second time of flight corresponds to a longest echo distance between a radar and the obstacle, and the first time of flight corresponds to a shortest echo distance between the radar and the obstacle
Data Source
Figure 1~2
Figure 3~4
Figure 5
AI summary
This application discloses a vehicle drivable area detection method, an autonomous driving assistance system, and an autonomous driving vehicle. The method includes: processing, by using a neural network, image data obtained by a camera apparatus, to obtain a first probability distribution of an obstacle; obtaining a second probability distribution of the obstacle based on a time of flight and an echo width of a radar echo signal; and obtaining, based on the first probability distribution of the obstacle and the second probability distribution of the obstacle, a drivable area of a vehicle represented by a probability, where the probability is a probability that the vehicle cannot drive through the area. The autonomous driving assistance system includes a camera apparatus, at least one radar, and a processor. The system is configured with the technical solutions that can implement the method. The autonomous driving vehicle includes the foregoing autonomous driving assistance system. In the method in this application, two sensors: a camera apparatus and a radar are integrated, distribution information about the obstacle obtained by the two sensors is fused, and the drivable area of the vehicle obtained after fusion is represented in a form of a probability. Therefore, information about the obstacle around the vehicle can be comprehensively obtained, so that a detection blind spot caused by a blind spot of the camera apparatus or a detection range of the radar is avoided.