Cognitive Mapping for Vehicle Environment Perception
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing vehicle systems face challenges in accurately and timely determining objects in their environment, which is crucial for safe and comfortable operation, especially in autonomous or semi-autonomous modes.
Innovation Solution
A method using a convolutional neural network (CNN) to process images and create cognitive maps, combining ground truth data with LIDAR data to accurately depict the vehicle's environment, including objects like other vehicles and pedestrians, enabling the vehicle to predict trajectories and operate safely.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional sensor systems are used to detect objects in the vehicle environment, then the system structure is relatively simple, but the accuracy and timeliness of object detection is insufficient
Solution Approach 1:
The patent combines multiple data sources including LIDAR data, ground truth data, and image data into a unified cognitive map representation. This integration allows the system to leverage the strengths of each data source (LIDAR for depth accuracy, ground truth for validation, images for semantic information) to achieve superior object detection accuracy while maintaining a coherent system architecture.
Solution Approach 2:
The patent transforms 2D image data into a 3D cognitive map representation by integrating LIDAR depth information. This dimensional transformation enables the system to perceive spatial relationships and object distances more accurately, converting planar image features into three-dimensional spatial understanding for better trajectory prediction.
2Reliability
If multiple data sources are integrated to improve environmental perception, then the accuracy of cognitive maps improves, but the computational complexity increases
Solution Approach 1:
The system performs preliminary processing of LIDAR data and ground truth data before integrating them with image data. By pre-processing and organizing spatial information from multiple sources, the system reduces the computational burden during real-time operation, as the data is already structured and validated before being combined in the cognitive map generation process.
Solution Approach 2:
The patent introduces a cognitive map as an intermediary representation that mediates between raw sensor data and trajectory prediction. This intermediate structure organizes and standardizes information from multiple data sources, making the integration process more manageable and reducing overall computational complexity by providing a unified framework for data fusion.
3Loss of time
If real-time processing of environmental data is performed, then timely trajectory prediction is achieved, but the processing speed may be compromised due to complex computations
Solution Approach 1:
The system performs preliminary organization and validation of data from multiple sources before real-time processing. By pre-structuring LIDAR point clouds, ground truth annotations, and image features, the system reduces the computational workload during critical real-time operation, enabling faster processing speeds when actual trajectory prediction is needed.
Solution Approach 2:
The patent segments the environmental perception process into distinct modules: LIDAR processing, ground truth integration, image processing, and trajectory prediction. This segmentation allows each module to be optimized independently and processed in parallel where possible, improving overall processing speed while maintaining real-time performance for the complete system.
Data Source
AI summary
A system, comprising a processor, and a memory, the memory including instructions to be executed by the processor to acquire the images of the vehicle environment, determine a cognitive map, which includes a top-down view of the vehicle environment, based on the image, and operate the vehicle based on the cognitive map.