Automated Driving Sensor Fusion With Common 3D Coordinates
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Combining real-time sensor data from multiple sensors for efficient control of automated driving systems is extremely challenging, as existing technologies face difficulties in accurate sizing and object recognition using RADAR and LIDAR, and distance and velocity determination from camera images.
Innovation Solution
An automated driving system that merges sensor data from diverse sensors like RADAR, LIDAR, and cameras into a common three-dimensional coordinate system, using a processor to generate drive control signals for path-finding and collision avoidance, with dynamic weighting of sensor data based on environmental conditions and sensor reliability.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If sensor data from multiple sensors (RADAR, LIDAR, cameras) is combined for automated driving control, then the characterization capabilities and robustness of the system are improved, but the computational complexity and difficulty of processing heterogenous data increases
Solution Approach 1:
The patent segments the sensor data processing by creating separate processing pipelines for different sensor types (RADAR, LIDAR, cameras) before merging their outputs. Each sensor type is processed independently to extract relevant features, then these features are integrated in a coordinated manner. This segmentation reduces the computational complexity of handling heterogenous data while maintaining the robustness benefits of multi-sensor fusion.
Solution Approach 2:
The patent introduces an intermediary coordinate system transformation layer that converts sensor data from different reference frames into a common vehicle-centered coordinate system. This intermediary transformation enables seamless integration of data from sensors with different modalities and reference frames, reducing the complexity of direct heterogenous data processing while preserving the reliability improvements from multi-sensor fusion.
2Extent of automation
If accurate object tracking information is provided for path-finding and collision avoidance, then the decision-making capability of the system is improved, but the measurement precision requirements for sensor data increase
Solution Approach 1:
The patent merges data from multiple sensor types (RADAR for velocity, LIDAR for distance, cameras for object recognition) to create comprehensive object tracking information. By combining the strengths of different sensors, the system achieves high measurement precision for object tracking, enabling intelligent decision-making for path-finding and collision avoidance while managing the precision requirements through sensor fusion rather than relying on a single high-precision sensor.
3Speed
If real-time sensor data processing is performed for automated driving control, then the responsiveness of the system is improved, but the computational load and energy consumption increase
Solution Approach 1:
The patent performs preliminary processing of sensor data by pre-processing individual sensor inputs (filtering, feature extraction, coordinate transformation) before the main fusion operation. This preliminary action reduces the computational load during real-time decision-making, enabling responsive automated driving control while managing energy consumption through optimized preprocessing steps that can be performed with lower computational resources.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
An automated driving system comprising a first sensor providing first sensor data, a second sensor providing second sensor data, wherein the first and second sensors are different types of sensors and a processor communicatively coupled to the first sensor and the second sensor. The processor is configured to transform the first and second sensor data into a common three-dimensional coordinate system to create a representation of an image for a field of view corresponding to a direction of travel, identify one or more objects in the field of view, and generate one or more drive control signals to control a motion of the vehicle in response to the one or more objects in the field of view.