3D Positioning with IMU in various environments

2022 Fiscal Year Final Research Report

• Project/Area Number: 20K11891
• Research Category: Grant-in-Aid for Scientific Research (C)
• Allocation Type: Multi-year Fund
• Section: 一般
• Review Section: Basic Section 61010:Perceptual information processing-related
• Research Institution: Nara Institute of Science and Technology (2021-2022); Kyushu University (2020)
• Principal Investigator: Uchiyama Hideaki 奈良先端科学技術大学院大学, 先端科学技術研究科, 准教授 (90735804)
• Project Period (FY): 2020-04-01 – 2023-03-31
• Keywords: IMU / ディープラーニング / オドメトリ / キャリブレーション

Outline of Final Research Achievements

Our objective is to develop a three-dimensional positioning method for any moving objects by using an Inertial Measurement Unit (IMU) which sensing is robust to the surrounding environment changes. The challenges are the estimation of noise and biases, as well as the incomplete determination of the gravity direction. In this work, we propose a method based on learning using a neural network (NN) to describe both noise and bias estimation and gravity direction estimation. Especially, we use camera-based positioning as ground truth for training the NN. Furthermore, we enhance the robustness against variations in IMU sensors by training the IMU sensor characteristics. During positioning inference, we can achieve three-dimensional positioning in any given environment using only the IMU through the utilization of the trained NN. We demonstrate the practical validation through measurements in real-world environments to assess the generalization performance of the proposed method.

Free Research Field

ナビゲーション

Academic Significance and Societal Importance of the Research Achievements

本研究では，ニューラルネットワークを用いて暗または陽にIMUのセンサ特性をモデル化し，運動学を用いて移動量推定を行う手法を提案した．現在の移動量推定はカメラに基づく技術が主流である．しかし，暗所では測位できず，見えの変化のある動的環境では，測位精度が大きく低下する．そこで，周辺環境の影響を受けないデータ計測を行うIMUを用いて高精度な移動量推定を行う技術を確立した．実験では空中のみならず，水中と空中を行き来する移動においても高精度な推定を行えることを示した．IMUのみを用いて移動量推定を実現することは，計算量及び環境に対するロバスト性の高さの面でナビゲーション技術として貢献した．