Edge Object Tracking via Region of Interest Extraction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current object tracking systems for vehicles face inefficiencies due to the need for raw sensor data processing across multiple vehicles and edge servers, leading to high computational resource usage and bandwidth requirements, especially when tracking objects that move out of the field of view.
Innovation Solution
Implementing an object tracking system that selects specific edge servers and connected vehicles to process only a region of interest, reducing computational load and bandwidth by transmitting object attributes rather than raw sensor data, and seamlessly handing over tracking as objects move between coverage areas.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If raw sensor data is processed across multiple vehicles and edge servers, then object detection and tracking can be performed cooperatively, but computational resource usage and bandwidth requirements increase significantly
Solution Approach 1:
The patent applies local quality by having each vehicle process only its own sensor data locally to extract object features and attributes, rather than sharing raw sensor data. Each vehicle performs localized feature extraction and sends only the essential object attributes to the mobility operator, reducing computational burden while maintaining cooperative tracking capability
Solution Approach 2:
The patent extracts only the necessary object attributes (features, location, velocity) from sensor data at the vehicle level before transmission. This extraction principle removes unnecessary raw sensor data from the communication and processing pipeline, keeping only the essential information needed for cooperative tracking
2Loss of information
If raw sensor data is transmitted between vehicles and edge servers, then complete object information can be shared, but bandwidth consumption increases
Solution Approach 1:
The system extracts and transmits only the essential object attributes (features, location, velocity) rather than raw sensor data. This selective extraction maintains the necessary information for tracking while dramatically reducing the data volume transmitted over the network
Solution Approach 2:
The patent creates simplified copies of object information in the form of structured attributes (features, location, velocity) that represent the essential characteristics of detected objects. These attribute copies are transmitted instead of original sensor data, reducing bandwidth requirements while preserving tracking capability
3Area of stationary object
If all connected vehicles process sensor data for object tracking, then tracking coverage is comprehensive, but device complexity and computational load increase
Solution Approach 1:
The patent segments the object tracking task by dividing responsibilities between vehicles and the mobility operator. Each vehicle independently performs local feature extraction and sends results to the mobility operator, which performs centralized tracking. This segmentation reduces the computational complexity at each vehicle while maintaining comprehensive coverage
Solution Approach 2:
Each vehicle performs partial action by processing only its own sensor data to extract object features, rather than processing data from all vehicles. This partial processing approach reduces individual vehicle computational load while the mobility operator aggregates results to achieve comprehensive tracking coverage
Data Source
AI summary
A method includes receiving features associated with an object to be tracked, receiving an estimated location of the object, determining a region of interest with respect to a first connected vehicle that includes the estimated location of the object, transmitting the features and the region of interest to the first connected vehicle, receiving object data associated with the object from the first connected vehicle, the object data comprising a location of the object, and updating an object track associated with the object based on the object data, the object track comprising the location of the object at a plurality of time steps.