Tracking Technology Explained

The Alius Motion Capture System employs advanced inertial measurement and forward kinematics technology to achieve high-precision, low-latency full-body motion capture. This document will introduce the system’s tracking principles and technical details in depth.

Inertial Measurement Principles

IMU Sensor Working Principle

The Inertial Measurement Unit (IMU) is the core sensor of the Alius Motion Capture System. Each tracker integrates a 6-axis IMU, including:

  • 3-axis accelerometer: Measures linear acceleration, used to detect gravity and motion acceleration
  • 3-axis gyroscope: Measures angular velocity, used to track rotational motion

Through sensor fusion algorithms, the system combines accelerometer and gyroscope data to calculate the precise orientation of the tracker (roll, pitch, yaw angles).

Sensor Fusion Algorithm

Alius uses an adaptive Kalman filter for sensor data fusion:

  1. Prediction step: Predict the next moment’s orientation based on gyroscope data
  2. Update step: Use accelerometer data to correct the prediction, especially the gravity direction
  3. Adaptive adjustment: Dynamically adjust filter parameters based on motion intensity

This fusion method effectively overcomes the limitations of single sensors:

  • The gravity vector from the accelerometer can be used for long-term orientation correction
  • The gyroscope provides high-frequency, low-latency rotation measurements
  • The combination achieves accurate and stable orientation tracking

Forward Kinematics Solving

Kinematic Chain Modeling

The Alius system models the human body as a series of kinematic chains connected by joints:

  • Torso chain: Pelvis → Spine → Neck → Head
  • Left arm chain: Pelvis → Left shoulder → Left elbow → Left hand
  • Right arm chain: Pelvis → Right shoulder → Right elbow → Right hand
  • Left leg chain: Pelvis → Left hip → Left knee → Left foot
  • Right leg chain: Pelvis → Right hip → Right knee → Right foot

Each joint has specific degrees of freedom (DOF):

  • Ball joint (e.g., shoulder, hip): 3 DOF
  • Hinge joint (e.g., elbow, knee): 1 DOF
  • Compound joint (e.g., spine): 2-3 DOF

Forward Kinematics Calculation

Forward Kinematics (FK) is the process of calculating the position of end effectors based on joint angles.

In the Alius system, the FK calculation process is as follows:

  1. Each tracker measures the orientation in its local coordinate system
  2. The system knows the human bone lengths (estimated from user height and proportions or manually measured)
  3. Starting from the root joint (usually the pelvis), calculate the world coordinate position of each joint level by level
  4. Use quaternion rotation representation to avoid gimbal lock problems

Mathematical expression:

P_end = P_root + R_root × (L_1 + R_1 × (L_2 + R_2 × L_3 + ...))

Where:

  • P = position
  • R = rotation matrix (converted from quaternion)
  • L = bone length vector

6-DOF Tracking Implementation

Orientation Tracking (3-DOF)

Each tracker provides complete 3-DOF orientation information:

  • Roll angle: Rotation around X-axis
  • Pitch angle: Rotation around Y-axis
  • Yaw angle: Rotation around Z-axis

Orientation is output in quaternion format, avoiding the gimbal lock problem of Euler angles and ensuring smooth and continuous animation.

Position Tracking (3-DOF)

Although IMU cannot directly measure position, Alius achieves position tracking through the following methods:

  1. Forward kinematics propagation: Starting from the known root joint position (such as head position tracked by VR headset), calculate all joint positions through the bone chain
  2. Zero Velocity Update (ZUPT): When the tracker detects a stationary state, reset the velocity to zero to reduce integration drift
  3. Foot contact detection: Recognize foot-ground contact states through acceleration patterns, providing position constraints

Latency Optimization Technology

Hardware-level Optimization

  • High sampling rate: IMU data sampling rate up to 1000 Hz
  • Fast preprocessing: Sensor data undergoes preliminary processing inside the tracker
  • Low-latency wireless communication: Optimized WiFi protocol to reduce transmission latency

Software-level Optimization

  • Prediction algorithm: Use extrapolation to predict the current frame’s orientation, compensating for transmission delay
  • Time synchronization: Timestamps of all trackers are synchronized to ensure data consistency
  • Buffer management: Dynamically adjust receive buffer to balance latency and stability

End-to-End Latency Analysis

Typical end-to-end latency < 20ms consists of:

  • IMU sampling and preprocessing: ~2ms
  • Wireless transmission: ~5ms
  • Reception and parsing: ~3ms
  • Kinematics solving: ~5ms
  • Application rendering: ~5ms

Drift Compensation

Problem and Causes

The main challenge of pure inertial tracking is drift, mainly caused by:

  • Gyroscope integration error accumulates over time
  • Accelerometer cannot distinguish between gravity and motion acceleration
  • Sensor noise and bias

Compensation Strategies

Alius employs multiple drift compensation mechanisms:

  1. Gravity alignment: Use the gravity direction measured by the accelerometer to periodically correct pitch and roll angles
  2. Magnetic interference suppression: Do not use magnetometer (susceptible to environmental interference), instead rely on motion constraints
  3. Contact point constraints: When body parts contact stable surfaces like ground or chair, enforce position constraints
  4. User calibration: Periodically prompt users to return to known poses (such as standing) for recalibration

Multi-Configuration Adaptation

5-Point Configuration

Suitable for basic VR applications. Trackers placed on:

  • Head (usually replaced by VR headset)
  • Left hand
  • Right hand
  • Left foot
  • Right foot

Pelvis and knee positions are estimated through forward kinematics.

16-Point Configuration

Full-body tracking, including:

  • Head
  • Pelvis
  • Left/right shoulder blades
  • Left/right upper arms
  • Left/right forearms
  • Left/right hands
  • Left/right thighs
  • Left/right calves
  • Left/right feet

Achieves film-level precision motion capture.

Technical Advantages

Compared to competitors, Alius’s tracking technology offers:

  • No external reference dependency: No cameras or base stations required
  • Strong anti-interference capability: Not affected by lighting or magnetic fields
  • Large range of movement: Not limited by venue size
  • Low cost: Compared to optical motion capture systems, cost reduced by over 80%

Future Improvement Directions

We recognize the following areas for improvement and plan to optimize in future versions:

  • Better position tracking: Explore visual-inertial fusion (VIO) technology
  • Longer calibration intervals: Improve algorithms to reduce user intervention
  • Multi-person simultaneous tracking: Support multi-user scenarios (such as group VR)

We welcome user feedback to help improve the Alius Motion Capture System.