Thermal Sensor Object Confirmation for Autonomous Navigation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current vehicle navigation systems, particularly for autonomous vehicles, face challenges in accurately distinguishing between live objects and non-live entities, such as cardboard cutouts or sculptures, due to limitations in existing sensor technologies like lidar and visible light cameras, which can be fooled by appearances and patterns.
Innovation Solution
Integration of a thermal sensor or imager to confirm object classifications by providing unique thermal signatures for objects like people, animals, and vehicles, allowing the system to differentiate between live and non-live objects, thereby enhancing navigation accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If visible light cameras and lidar are used for object detection, then the system can identify objects based on color boundaries and patterns, but the system cannot reliably distinguish between live objects and non-live entities such as cardboard cutouts or sculptures
Solution Approach 1:
The patent introduces a thermal sensor as an intermediary device that detects thermal radiation from objects. This thermal information serves as a mediator between the existing visual sensors and the object classification system, providing additional evidence about whether detected objects are live (humans, animals) or non-live (cardboard cutouts, sculptures). The thermal sensor data is integrated with lidar and camera data to improve classification reliability.
Solution Approach 2:
The patent changes the detection parameter from optical properties (color, patterns) to thermal properties (thermal radiation, temperature). By measuring thermal radiation in the infrared spectrum, the system gains a new parameter that fundamentally differs between live objects (which generate heat metabolically) and non-live objects (which do not), thereby resolving the classification ambiguity.
2Reliability
If thermal sensors are integrated into the navigation system, then the system can confirm object classifications and differentiate live from non-live objects, but the device complexity and computational requirements increase
Solution Approach 1:
The patent applies partial action by using the thermal sensor selectively rather than continuously processing all thermal data. The system queries the thermal sensor about specific detected objects to confirm or refute classifications, rather than attempting to identify all objects in the scene. This reduces computational burden while maintaining high reliability for critical detections.
Solution Approach 2:
The thermal sensor is designed to serve multiple functions: confirming object presence, differentiating live from non-live objects, and providing thermal signature data for classification. This multi-functionality justifies the added device complexity by consolidating multiple detection capabilities into a single sensor type.
3Measurement precision
If the thermal sensor queries all detected objects, then the system can maximize detection accuracy, but the compute and data demands on the thermal imaging camera system become excessive
Solution Approach 1:
The patent applies local quality by directing thermal sensor queries selectively to specific regions of interest rather than uniformly across the entire field of view. The system focuses computational resources on areas where objects have been detected by other sensors, concentrating energy expenditure where it is most needed for accurate classification.
Solution Approach 2:
The system performs preliminary detection using lidar and visible light cameras before querying the thermal sensor. This preliminary action filters the scene to identify only those regions containing potential objects, thereby reducing the number of thermal queries needed and lowering overall computational demands while maintaining detection 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
The thermal sensor improves object detection accuracy by confirming the presence of live pedestrians and differentiating them from inanimate objects, crucial for safe navigation and path planning, especially in emergency braking scenarios.
Implementation Method 1
a thermal sensor or imager to confirm object classifications by providing unique thermal signatures for objects like people, animals, and vehicles
Data Source
AI summary
A thermal imager is used to confirm object classifications as described. In one example, a scene modeling system has a sensor system to generate an image of a scene. A thermal camera generates a thermal image of the scene within a field of regard, and a modeling processor coupled to the sensor system and to the thermal camera correlates a position of a selected object in the scene to the field of regard of the thermal camera and queries the thermal camera to confirm a classification of the selected object. The thermal camera is configured to receive the position of the selected object and to confirm the classification.