Development of Low-Cost Tracking System for Infantry Training
Agency / Branch:
DOD / NAVY
PhaseSpace and the University of Southern California (USC) propose a hybrid tracking solution that will combine the PhaseSpace Impulse LED tracking technology with computer vision image processing and inertial tracking. The hybrid tracker will track IR LEDs on a soldier''s head, fingers, weapon, and boots, to sub-millimeter precision at 960 Hz. It will also be capable of capturing each soldier''s silhouette at 12 Hz. Inertial sensors will be used to track a soldier''s movements in areas that are not in the field-of-view of the camera system, such as passageways between rooms. The inertial sensors will also be used when the LEDs are occluded by other people or objects in the room. All data from the sensors located on the soldier will be transmitted via wireless communication using a protocol that will allow eight users with multiple sensors in the same vicinity. Accurate weapon and helmet pose estimation is crucial for virtual and augmented reality training exercises. To meet the requirements of the low-cost tracking solicitation, PhaseSpace proposes to mount cameras to the weapon and helmet of each soldier and use identifiers (IR LEDs or passive fiducials) in the training facility to estimate the pose of the tracked objects. This outward looking approach can be implemented on a low-cost per soldier basis. The following hybrid tracking component areas were designed and preliminarily tested in Phase I of this STTR: LED tracking, skeleton fitting to LED data points, finger tracking, silhouette tracking, inertial tracking, weapon and helmet pose estimation using on-board cameras, eye tracking, facial tracking, and combined PhaseSpace Vision system and LED (Impulse) system calibration and synchronization. In Phase II we propose to further develop these components and combine them into one integrated tracking system over the course of 18 months.
Small Business Information at Submission:
Research Institution Information:
1937 A Oak Park Blvd. Pleasant Hill, CA 94523-
Number of Employees:
University of Southern California
MMRD Lab, Attn: Mark Bolas
5318 McConnell Ave.
Los Angeles, CA 94043-