Leg tracking method based on the SJPDAF technique

Human Leg Tracking Technology Based on SJPDAF with Combined Leg-Pair Grouping

This technology enhances target assignment accuracy by first calculating probabilities through primary matching of multiple measured legs extracted from distance sensors with target legs using the SJPDAF technique, followed by secondary posterior probability calculation through leg-pair based grouping.

Existing technologies often treat leg measurements as independent targets or simply group the two closest legs, which leads to tracking errors in crowded environments when target legs are swapped or only a single leg is detected.

This technology introduces a new posterior probability calculation algorithm that considers not only individual elements but also combinations of two legs by adding a leg-pair grouping step to the existing matching process. It can be applied to service robots that follow people, such as guide robots and luggage transport robots, ensuring stable tracking of the target person without losing them even in crowded spaces.

‍

‍

Key Features:

• A step of extracting multiple measured legs corresponding to human legs from distance sensor scan data
• A step of calculating a first posterior probability by applying measured legs and target legs to the SJPDAF technique
• A step of generating grouped target legs by pairing multiple target legs
• A configuration that assigns targets by calculating a second posterior probability between grouped target legs and grouped measured legs