Adaptive Item Counting Algorithm for Weight Sensors
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing position tracking systems face challenges in scaling to larger spaces due to synchronization issues and limited computing power, which affects their ability to provide real-time, accurate tracking of people and objects.
Innovation Solution
The system employs a distributed architecture with camera clients processing frames close to cameras, using edge computing to improve synchronization, and assigns unique addresses to weight sensors for accurate item tracking, along with an adaptive item counting algorithm to account for sensitivity changes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a distributed architecture with edge computing is used to process frames close to cameras, then synchronization accuracy and latency are improved, but device complexity increases
Solution Approach 1:
The system divides the centralized processing architecture into distributed edge computing nodes (camera clients) that process video frames locally near the cameras. Each camera client independently performs item counting and tracking operations, reducing the need for constant centralized coordination and improving synchronization accuracy while distributing computational load.
Solution Approach 2:
The patent transitions from a two-dimensional centralized processing model to a multi-dimensional distributed architecture where processing occurs at multiple levels: edge devices (camera clients), intermediate servers, and centralized systems. This dimensional expansion allows simultaneous local processing for synchronization and centralized coordination for overall system management.
2Area of stationary object
If more weight sensors are deployed to track items in larger spaces, then tracking coverage is improved, but synchronization issues and computing power limitations worsen
Solution Approach 1:
The system segments the large tracking space into multiple zones, each monitored by dedicated camera clients and weight sensor groups. This segmentation allows independent processing of sensor data in each zone, reducing the synchronization burden on the centralized system while expanding overall tracking coverage to larger areas.
Solution Approach 2:
The patent introduces intermediate servers as mediators between the distributed camera clients/weight sensors and the centralized system. These intermediaries aggregate and pre-process data from multiple sensors, reducing the communication overhead and synchronization complexity for the centralized system while enabling expanded sensor deployment coverage.
3Measurement precision
If conventional item counting methods are used without sensitivity analysis, then processing speed is maintained, but measurement accuracy deteriorates due to sensor sensitivity changes
Solution Approach 1:
The system performs sensitivity analysis and calibration operations in advance during system initialization or idle periods. By pre-computing sensitivity correction factors and storing them for later use, the actual item counting process can apply these corrections without real-time computational overhead, maintaining both accuracy and processing speed.
Solution Approach 2:
The patent dynamically adjusts counting parameters based on sensor sensitivity characteristics. Instead of using fixed threshold values, the system modifies counting thresholds and parameters according to the measured sensitivity of each weight sensor, compensating for manufacturing variations and environmental factors while maintaining efficient processing.
Data Source
AI summary
A system includes a weight sensor and a weight server. The weight sensor generates a signal indicative of a weight of at least one of an item. The weight server detects an event corresponding to a weight change on the weight sensor when a quantity of the item is removed from the weight sensor. The weight server determines the item quantity by calculating a result of dividing the weight change over a unit weight of the item. If the result is within a threshold range from an integer that is closest to the result, the weight server determines that a quantity of the item with the amount of the integer is removed from the weight sensor. If the result is not within the threshold range from the integer, the weight server uses weight change patterns of historically observed signals to determine item quantity that was removed from the weight sensor.