Inertial Measurement Unit
Inertial Measurement Units (IMUs) are small, self-contained sensors that measure acceleration and angular velocity, providing crucial data for motion tracking and localization across diverse applications. Current research focuses on improving IMU data processing through advanced algorithms like Kalman filters, recurrent neural networks (RNNs), transformers, and novel approaches leveraging data augmentation and equivariant neural networks to enhance accuracy and robustness, particularly in challenging environments or with limited data. These advancements are driving significant improvements in areas such as human activity recognition, visual-inertial odometry, and pose estimation, with implications for fields ranging from healthcare and fitness to robotics and autonomous navigation. The development of efficient calibration techniques and the exploration of sensor fusion with other modalities, like cameras and UWB, are also key areas of ongoing investigation.
Papers
An Experimental Comparison of Floating Base Estimators for Humanoid Robots with Flat Feet
Prashanth Ramadoss, Stefano Dafarra, Silvio Traversaro, Daniele Pucci
Automated Mobility Context Detection with Inertial Signals
Antonio Bevilacqua, Lisa Alcock, Brian Caulfield, Eran Gazit, Clint Hansen, Neil Ireson, Georgiana Ifrim
Indoor SLAM Using a Foot-mounted IMU and the local Magnetic Field
Mostafa Osman, Frida Viset, Manon Kok
Transformer Inertial Poser: Real-time Human Motion Reconstruction from Sparse IMUs with Simultaneous Terrain Generation
Yifeng Jiang, Yuting Ye, Deepak Gopinath, Jungdam Won, Alexander W. Winkler, C. Karen Liu
A GNSS Aided Initial Alignment Method for MEMS-IMU Based on Backtracking Algorithm and Backward Filtering
Xiaokang Yang, Gongmin Yan, Hao Yang, Sihai Li
Aggressive Racecar Drifting Control Using Onboard Cameras and Inertial Measurement Unit
Shuaibing Lin, JiaLiang Qu, Zishuo Li, Xiaoqiang Ren, Yilin Mo