Event-Based Sensor Object Tracking for Driver Assistance
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing driver assistance and semi-automated driving systems face delays in response time due to camera image processing limitations, such as refresh rate constraints and high bandwidth requirements for image data transmission, which can lead to increased accident risks.
Innovation Solution
Employing event-based sensors with light-sensitive pixels that output events based on relative light intensity changes, allowing for faster event availability and reduced semantic information loss, enabling efficient object identification and tracking by assigning new events to detected distributions or objects, and prioritizing unexpected events for rapid response.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of time
If conventional camera image processing is used, then semantic information is maintained, but response time increases due to refresh rate constraints and high bandwidth requirements
Solution Approach 1:
The patent segments the visual information processing into two distinct streams: event-based processing for temporal-critical information (fast response, low bandwidth) and frame-based processing for spatial-critical information (high semantic content). This segmentation allows the system to process different types of visual information through optimized pathways, reducing overall response time while maintaining semantic information through the frame-based stream.
Solution Approach 2:
The patent introduces an intermediary mechanism that correlates events with frames by matching spatial coordinates and temporal sequences. This intermediary layer bridges the gap between the fast event stream and the semantic-rich frame stream, allowing events to be annotated with semantic information from frames without requiring the entire frame processing pipeline for every event, thus reducing response time while preserving semantic context.
2Loss of time
If camera refresh rate is increased to reduce response time, then response time improves, but bandwidth requirements and system complexity increase
Solution Approach 1:
The patent extracts the temporal sensitivity function from the conventional frame-based camera system and implements it separately through event-based sensors. This extraction allows the system to achieve high temporal resolution for motion detection without requiring the entire camera system to operate at high refresh rates, thereby reducing overall system complexity and bandwidth requirements while maintaining fast response capability.
Solution Approach 2:
The patent implements a dynamic processing architecture where the system adaptively routes information based on its type: event-based processing for temporal-critical information and frame-based processing for spatial-critical information. This dynamic routing optimizes resource allocation and reduces system complexity by avoiding the need for a uniformly high refresh rate across all processing pathways.
3Loss of time
If event-based sensors are used, then response time decreases and bandwidth is reduced, but semantic information loss increases
Solution Approach 1:
The patent merges the strengths of both event-based and frame-based systems by processing events and frames through parallel pathways and then correlating them. The event stream provides fast temporal response with low bandwidth, while the frame stream provides semantic information. By merging these streams through coordinate matching and temporal correlation, the system achieves fast response times while recovering semantic information through the frame-based pathway.
Solution Approach 2:
The patent implements a feedback mechanism where frames provide semantic context that feeds back into the event processing pathway. Events are correlated with corresponding frames to inherit semantic information, and this feedback loop allows the system to maintain semantic awareness while primarily relying on the fast, low-bandwidth event stream for temporal-critical decisions.
4Reliability
If all events are processed equally, then comprehensive monitoring is achieved, but false alarms increase due to inability to distinguish expected from unexpected events
Solution Approach 1:
The patent performs preliminary classification of events by correlating them with predicted object trajectories and behaviors before full processing. Events that match predicted patterns are pre-identified as expected, while deviations are flagged as unexpected. This preliminary action allows the system to prioritize processing of unexpected events, reducing false alarms from routine events while maintaining comprehensive monitoring coverage.
Solution Approach 2:
The patent applies different processing quality levels to different events based on their characteristics. Unexpected events or those in critical zones receive higher processing priority and more thorough analysis, while expected routine events receive minimal processing. This local quality differentiation improves monitoring accuracy for critical events while maintaining overall processing efficiency by not uniformly processing all events at maximum detail.
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
This approach significantly reduces response time while maintaining semantic information, allowing for timely adjustments in vehicle behavior and minimizing false alarms in monitoring systems, thereby enhancing safety and accuracy in object tracking and collision avoidance.
Implementation Method 1
The event-based sensor (1) includes light-sensitive pixels and is designed to output, in response to a relative change in light intensity incident upon a pixel by at least one predefined percentage, an event (1b) assigned to this pixel
Data Source
AI summary
A method for identifying and/or tracking objects in a spatial area. The method includes observing the area with the aid of at least one event-based sensor, the event-based sensor including light-sensitive pixels, and a relative change of the light intensity incident upon a pixel by at least a predefined percentage prompting the sensor to output an event assigned to this pixel. The method further includes, in response to the sensor, outputting a new event, ascertaining an assessment for this event which is a measure of the extent to which this event matches an already detected distribution of events, and/or of the extent to which it is plausible that the event stems from an already detected object; and in response to the ascertained assessment meeting a predefined criterion, assigning the new event to the already detected distribution, or the already detected object.