Thermal Pose Tracking for Autonomous Vehicle Connector Mating
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing docking operations for vehicles, such as air-to-air refueling and spacecraft docking, rely heavily on human judgment and are challenging to extend to autonomous systems due to complexity and the difficulty in certifying artificial intelligence-based solutions, making them difficult to implement in autonomous vehicles like drones and spacecraft.
Innovation Solution
A thermal imaging sensor and vision processor system that generates thermal image data and estimates the pose of an object by comparing a modeled outline with gradient data, adjusting the pose estimate based on overlap values to facilitate autonomous connector mating.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If human operators guide complex docking operations, then operational reliability is maintained through human judgment, but automation level remains low and operational complexity is high
Solution Approach 1:
The patent replaces human operators with an all-optical vision-based system that uses thermal imaging sensors and computer vision algorithms to automatically guide docking operations. The system processes thermal images to detect connectors, estimate poses, and generate control commands, eliminating the need for human mechanical intervention while maintaining operational reliability through automated decision-making algorithms.
Solution Approach 2:
The patent creates a virtual model of the physical docking environment by generating a virtual copy of the scene from thermal images. This virtual model includes detected connectors, estimated poses, and spatial relationships, allowing the system to plan and execute docking maneuvers in the virtual space before applying commands to the physical system, thereby enhancing reliability through simulation-verified decisions.
2Extent of automation
If artificial intelligence-based solutions are used for autonomous docking, then automation level increases, but certification difficulty increases and reliability becomes harder to verify
Solution Approach 1:
The patent replaces complex AI-based image recognition with a simpler all-optical vision system that uses thermal image processing and geometric algorithms to detect connectors and estimate poses. This substitution reduces certification difficulty by using well-understood optical and geometric principles rather than black-box AI models, while maintaining high automation levels through automated image processing and control command generation.
Solution Approach 2:
The patent introduces an intermediary all-optical vision system that bridges the gap between raw thermal images and autonomous docking control. This intermediary layer uses thermal image processing to extract connector positions and poses, then feeds this information to the control system, providing a transparent and certifiable intermediate representation that simplifies verification of the overall system's reliability.
3Measurement precision
If complex stereoscopic vision systems are used to aid operators, then measurement precision of connector positions improves, but device complexity increases and processing resources are consumed
Solution Approach 1:
The patent changes the operational parameter from visible light to thermal infrared imaging. This parameter change enables the system to detect connectors based on thermal signatures rather than visual appearance, simplifying the vision system by eliminating the need for complex visible light optics while maintaining measurement precision through thermal contrast detection.
Solution Approach 2:
The patent substitutes complex stereoscopic vision hardware with a simpler thermal imaging-based monocular system. By using thermal images and processing them through an all-optical vision algorithm, the system achieves connector detection and pose estimation without requiring multiple cameras or complex stereoscopic processing, thereby reducing device complexity while maintaining precision.
4Productivity
If autonomous vehicle docking is implemented, then productivity increases through automated operations, but measurement precision of connector alignment becomes more critical and harder to achieve
Solution Approach 1:
The patent changes the detection parameter from visible light to thermal infrared, which provides superior contrast for detecting connector components. This parameter change enhances measurement precision by making connector boundaries and features more distinct in thermal images, enabling more accurate alignment measurements that are critical for high-speed autonomous docking operations.
Solution Approach 2:
The patent implements a feedback mechanism where the all-optical vision system continuously processes thermal images to detect connector positions and estimated poses, then feeds this information back to the control system for real-time adjustment of docking maneuvers. This closed-loop feedback ensures high measurement precision is maintained throughout the automated docking process, enabling productivity increases without sacrificing alignment accuracy.
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
Provides an all-optical, passive solution for autonomous connector mating with high confidence, reducing reliance on human operators and enhancing reliability and repeatability of maneuvers without increasing processing resources or sensor costs.
Implementation Method 1
a thermal imaging sensor configured to generate thermal image data depicting at least a portion of an object
Data Source
AI summary
A device includes a thermal imaging sensor configured to generate thermal image data depicting at least a portion of an object. The device also includes a vision processor coupled to the thermal imaging sensor. The vision processor is configured to generate outline image data corresponding to a modeled outline of the object based on a model of the object and a pose estimate of the object. The vision processor is also configured to determine an overlap value indicating an amount of overlap between the modeled outline and gradient data associated with the thermal image data, and to adjust the pose estimate of the object based on the overlap value.