Model-based Multiple Object Tracking Using Capacitive Proximity Sensors
Published:
Abstract
In this work, an object tracking method on capacitive proximity sensor is presented. Arranged as a vector and installed in the sidewalls of a work table with a robot, these sensors are able to detect proximity events in the near field of the workspace, including human approaching the table or interacting with the robot. Considering the low spatial resolution of the sensors, a preprocessing method was implemented and the data for tracking was prepared. The tracking method is based on a Kalman-Filter, where the data association and occlusion problem are discussed.
This work has been presented in IROS 2019 with a post: 
If you are interested in this project, please check our paper.