Demand-Based FOV Allocation for Remote Sensing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Traditional remote sensing systems are inefficient in collecting high-quality 'before' data for regions of interest, as they are typically scheduled in advance and only respond to user requests after an event occurs, limiting the availability of comparative data and leading to resource wastage due to uncertainty and late user requests.
Innovation Solution
A demand-based field of view allocation method for remote sensing systems, where a processor determines the required scenes to collect using a demand map updated in near-real-time, considering market demand, scarcity, and staleness of data, to autonomously configure the sensor field of view and collect data in real-time, reducing uncertainties and optimizing resource allocation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional scheduled remote sensing collections are used, then system simplicity is maintained, but the availability of high-quality 'before' data is limited
Solution Approach 1:
The system performs preliminary actions by proactively collecting and storing sensor data in an inventory system before user requests are made. The demand map pre-identifies regions of interest and schedules collections in advance, ensuring high-quality 'before' data is available when needed for comparison against current data.
Solution Approach 2:
The system transitions from static, pre-scheduled collections to dynamic, demand-driven collections. The demand map is continuously updated based on changing market conditions, user requests, and data scarcity/staleness, allowing the system to adapt its collection schedule in real-time to maximize the availability of useful before data.
2Reliability
If remote sensing collections are scheduled only in response to user requests, then resource allocation is optimized, but the probability of having high-quality 'before' data decreases
Solution Approach 1:
The system performs preliminary data collection based on the demand map before actual user requests are made. By proactively gathering data in regions identified as potentially needed (based on market conditions, news events, and data staleness), the system ensures before data is already in the inventory when users make requests, eliminating collection delays.
Solution Approach 2:
The system uses feedback from the inventory system to continuously update the demand map. When the inventory lacks sufficient before data for certain regions, the demand map automatically increases priority for collecting data in those regions, creating a closed-loop system that learns from data gaps and proactively fills them.
3Productivity
If traditional advance scheduling is used, then system complexity is reduced, but resource wastage increases due to uncertainty and late user requests
Solution Approach 1:
The system replaces static advance scheduling with dynamic demand-based scheduling. The demand map continuously adapts to changing conditions including market demand, current events, and inventory status, allowing resources to be allocated precisely when and where needed rather than following fixed schedules, thereby reducing resource wastage.
Solution Approach 2:
The system implements feedback loops where inventory status and user request patterns continuously inform demand map updates. This feedback mechanism allows the system to learn from past performance and optimize resource allocation over time, reducing wastage by collecting data more strategically based on actual demand patterns rather than predetermined schedules.
Data Source
AI summary
The present disclosure provides a system, method, and apparatus for collecting sensor data by a remote vehicle. The method involves determining, by at least one first processor, at least one scene to collect by using information in a demand map. The method further involves configuring, by at least one processor, a field of view for at least one sensor associated with the remote vehicle to collect at least one scene. Further, the method involves collecting, by at least one sensor, at least one scene.