HTC VIVE Ultimate Trackerの精度評価:ロボットと人間の動作を用いた多様な環境条件下での検証
Evaluating the precision of the HTC VIVE Ultimate Tracker with robotic and human movements under varied environmental conditions
VRヘッドセット不要で使えるHTC VIVE Ultimate Trackerの追跡精度を、ロボットアームとOptiTrackを基準に照明・速度・環境変化など様々な条件下で評価し、平均4.98mmの精度を明らかにした。
著者: Julian Kulozik, Nathanaël Jarrassé
分類: cs.RO, cs.HC
原文アブストラクト
The HTC VIVE Ultimate Tracker, utilizing inside-out tracking with internal stereo cameras providing 6 DoF tracking without external cameras, offers a cost-efficient and straightforward setup for motion tracking. Initially designed for the gaming and VR industry, we explored its application beyond VR, providing source code for data capturing in both C++ and Python without requiring a VR headset. This study is the first to evaluate the tracker's precision across various experimental scenarios. To assess the robustness of the tracking precision, we employed a robotic arm as a precise and repeatable source of motion. Using the OptiTrack system as a reference, we conducted tests under varying experimental conditions: lighting, movement velocity, environmental changes caused by displacing objects in the scene, and human movement in front of the trackers, as well as varying the displacement size relative to the calibration center. On average, the HTC VIVE Ultimate Tracker achieved a precision of 4.98 mm +/- 4 mm across various conditions. The most critical factors affecting accuracy were lighting conditions, movement velocity, and range of motion relative to the calibration center. For practical evaluation, we captured human movements with 5 trackers in realistic motion capture scenarios. Our findings indicate sufficient precision for capturing human movements, validated through two tasks: a low-dynamic pick-and-place task and high-dynamic fencing movements performed by an elite athlete. Even though its precision is lower than that of conventional fixed-camera-based motion capture systems and its performance is influenced by several factors, the HTC VIVE Ultimate Tracker demonstrates adequate accuracy for a variety of motion tracking applications. Its ability to capture human or object movements outside of VR or MOCAP environments makes it particularly versatile.
関連論文
- EgoWholeView:頭部搭載カメラのみによる身体と両手の動作キャプチャで身体性デモンストレーションを実現モーションキャプチャ
- 生体力学的知識を活用した多視点マーカーレス手の動きのモーションキャプチャモーションキャプチャ
- IMU-HOI: 接触を考慮した慣性融合による人間と物体の相互作用とモーションキャプチャの共生フレームワークモーションキャプチャ
- IMUと多体動力学モデルを密結合した運動・力学推定手法モーションキャプチャ
- 単眼イベントカメラによるモーションキャプチャシステムモーションキャプチャ
- CoCap: 屋外環境における複数人物シーンの協調モーションキャプチャモーションキャプチャ
