Hybrid Object Tracking Using Visual and Inertial Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing user identification systems in facilities, such as retail and inventory management systems, face challenges in maintaining accurate user identity association with object representations, particularly in crowded environments or due to camera failures, leading to loss of identity tracking.
Innovation Solution
The system employs a combination of visual tracking module data and device movement data from inertial sensors to associate user identifiers with object representations, using pattern matching and trajectory comparison within threshold values to reassert identity.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If visual tracking only is used to identify users in a facility, then the system is simpler to implement, but user identity association is lost in crowded environments or due to camera failures
Solution Approach 1:
The patent combines visual tracking data from cameras with device movement data from mobile devices to create a hybrid tracking system. This merging of two different tracking approaches allows the system to maintain user identity association reliably in crowded environments or when camera failures occur, while distributing the reliability burden across multiple data sources rather than relying on a single complex system
Solution Approach 2:
The patent introduces device movement data as an intermediary element that bridges the gap when visual tracking fails. By comparing device movement patterns with visual tracking data, the system can reassert user identity associations even when direct visual observation is unavailable or unreliable, acting as a mediator that maintains tracking continuity
2Measurement precision
If continuous visual tracking is performed to maintain user identity, then tracking accuracy is improved, but computational complexity increases
Solution Approach 1:
The patent replaces pure visual-mechanical tracking with a hybrid approach that incorporates device-based motion sensing. By substituting some visual tracking computations with device movement data processing, the system maintains tracking accuracy while reducing the computational burden on the central system, as device sensors perform local motion detection and data preprocessing
3Reliability
If device movement data is integrated with visual tracking data, then user identity association reliability is improved, but data processing complexity increases
Solution Approach 1:
The patent applies partial action by using device movement data selectively rather than continuously. The system compares device movement patterns with visual tracking data only when needed to reassert or verify user identity associations, particularly in challenging conditions. This partial application of data fusion reduces overall processing complexity while maintaining reliability where it matters most
Data Source
AI summary
Objects in a facility, such as users or totes, may be tracked as the object moves within the facility. An object representation of the object may be detected using image data. Apparent motion of the object representation may also be detected based on changes within the image data over time, which can help determine movement data for the object representation. However, the identity of the object representation may not be known. Using data from motion sensors in a device, movement data for the device can be generated. Thereafter, the movement data determined using the image data and the movement data determined using the data from the motion sensors in the device are compared for a possible match. If a match is found, an identifier of the device may then be associated with the object representation in the image data for tracking purposes.