Multi-network path generation for autonomous parking
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing autonomous and assisted driving systems face inefficiencies in parking due to excessive resource utilization and inability to function without GPS or parking lines, leading to repeated path generation and excessive control adjustments.
Innovation Solution
A multi-network-based system utilizing on-vehicle sensors for semantic scene segmentation, object detection, and vehicle dynamics to generate paths independently, combining GPS maps and deep learning for accurate parking line detection and path planning, even in GPS-denied areas.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If surveillance camera information and computer vision are used to generate paths, then path planning capability is improved, but resource utilization (computing, processor, bandwidth, memory) increases excessively
Solution Approach 1:
The system segments the path planning task into multiple stages: surveillance camera-based initial path generation, followed by on-vehicle sensor-based refinement. This segmentation allows the computationally intensive computer vision task to be performed only once for initial path generation, while subsequent adjustments use lighter on-vehicle sensors, thereby reducing overall resource utilization while maintaining path planning capability.
Solution Approach 2:
The system performs preliminary path generation using surveillance camera information before the vehicle arrives at the parking zone. This preliminary action establishes an initial path that reduces the need for repeated computationally intensive path generation once the vehicle is in the parking zone, thereby reducing computing and processor utilization during the actual parking maneuver.
2Adaptability or versatility
If surveillance camera information is used for path generation, then path planning is enabled, but repeated path generation occurs causing excessive resource utilization
Solution Approach 1:
The system generates an initial path using surveillance camera information before the vehicle enters the parking zone. This preliminary path generation eliminates the need for repeated computer vision-based path generation during the parking maneuver, as on-vehicle sensors merely refine the pre-established path, thereby reducing the time loss from repeated path generation.
3Measurement precision
If on-vehicle sensors and multi-network system are used, then parking assistance accuracy is improved, but device complexity increases
Solution Approach 1:
The system merges surveillance camera information with on-vehicle sensor data in a unified processing framework. By combining these data sources and their respective path generation capabilities, the system achieves higher parking assistance accuracy through multi-source information fusion, while the integrated architecture manages device complexity through coordinated operation of multiple networks.
Solution Approach 2:
The system implements a multi-network architecture where the same on-vehicle sensor network serves multiple functions: initial environment perception, real-time obstacle detection, path refinement, and parking execution monitoring. This multi-functionality reduces the need for separate dedicated systems, thereby managing device complexity while maintaining high parking assistance accuracy.
Data Source
AI summary
Systems and methods of deep neural network based parking assistance is provided. A system can receive data sensed by one or more sensors mounted on a vehicle located at a parking zone. The system generates, from a first neural network, a digital map based on the data sensed by the one or more sensors. The system generates, from a second neural network, a first path based on the three-dimensional dynamic map. The system receives vehicle dynamics information from a second one or more sensors located on the vehicle. The system generates, with a third neural network, a second path to park the vehicle based on the first path, vehicle dynamics information and at least one historical path stored in vehicle memory. The system provides commands to control the vehicle to follow the second path to park the vehicle in the parking zone.